European Research Consortium for Informatics and Mathematics Number 41 April 2000 www.ercim.org

Special Theme: Web Te c h n o l o g i e s

NEXT ISSUE: July 2000 Special Theme: Robotics CONTENTS

KEYNOTE 28 The Social Web Cockpit: A Tool to Support Knowledge Sharing Communities

3 Tim Berners Lee: by Wolfgang Gräther and Wolfgang Prinz The future Web and Europe – the Web and 30 Eliminating the I/O bottleneck from World Wide future Europe Web Proxies by Evangelos P. Markatos

31 WEBPATH – a 3-D Browsing History JOINT ERCIM ACTIONS Visualisation by Emmanuel Frécon

4 ERCIM launches the DELOS Network of 32 The sView System Excellence for Digital Libraries by Carol Peters by Fredrik Espinoza and Markus Bylund

5 Cor Baayen Award 2000 – Call for Candidates 34 Building a multinational R&D proposal through the Web by Andrew Hallan, Alexander M. 5 Joint European-US Strategic Workshops Samsonov and Wolfgang Ziegler by Jean-Eric Pin 35 The Collaborative Web 6 ANFAS: Data Fusion for Flood Analysis and by László Kovács and András Micsik Decision Support by Simon Lambert 37 Separation of Concerns in Multimedia Services SPECIAL THEME: WEB TECHNOLOGIES by Olle Olsson 38 Increasing Security in Virtual Enterprise 7 Web Technologies – Introduction Communication for Mobile Environment by Robert Caillau by István Mezgár, Tamás Szabó

8 Components for Data Intensive XML and Zsolt Kerecsen Applications by Peter Fankhauser, Gerald Huck, 40 Creating Hypertextual Dossiers in XML for the and Ingo Macherius Italian Parliament

9 EXMO: Computer mediated Exchange of by Andrea Marchetti and Patrizia Andronico Structured Knowledge by Jérôme Euzenat 41 Providing the Context for Business Decisions 10 Escrire: Embedded Structured Content by Brian Matthews Representation in Repositories 42 Virtual Physics and Astronomy by Jérôme Euzenat by Detlef Skaley and Udo Zlender

11 Workflow Demonstrator by Damian Mac Randal 44 CommonGIS – Common Access to 12 Promoting XML in Italy by Laura Abba Geographically Referenced Data by Hans Voss, Natalia Andrienko and Gennady Andrienko 13 The XMLC Browser by Fabio Vitali

14 Using XML for Data Management by Victoria RESEARCH AND DEVELOPMENT Marshall, Brian Matthews and Kevin O’Neill

15 Corporate Knowledge Management through 46 StraDiWare: Linking Business and IT Models Intranet/Internet by Rose Dieng using Graph Theory by Tony Conway

16 Multilingual to access Social Science 47 Computer Modelling in Geosciences Databy Brian Matthews and Michael Wilson by Stanislav Saic and Josef JeÏek

17 ICS-VRP: a Tool for Parsing and Validating RDF 48 Efficient Use of Networks through equal Metadata & Schemas Bandwidths by Annette Kik and Henk Nieland by Karsten Tolle and Vassilis Christophides 49 Interactions in Virtual Environments 18 XHTML 1.0 for Device-Independent Web Access by Lakshmi Sastry and David Boyd by Steven Pemberton 50 Alternative Processing of Verbal Data 19 : a Testbed for Web Protocols and by Milan Mare‰ Formats by Irène Vatton 51 Towards the Distributed Virtual Education and 20 XML Accessibility Guidelines Training Environment by Igor Nikitin, Valery by Daniel Dardailler Burkin and Stanislav Klimenko

21 The W3C in 7 Points 52 Wavelets improve Information on Earthquakes 23 Web Images and Blind Users by Henk Nieland by Zdenûk Mikovec, Martin Klíma, Du‰an Pavlica and Pavel Slavík EV E N T S

24 Adaptive Browsing: Supporting All Users on the 53 SOFSEM’99 Web by Constantine Stephanidis by Jan Pavelka 25 Vector by David Duce 54 Announcements 27 More Quality for Services in Internet by Juha Koivisto 55 IN BRIEF

Cover image: WEBPATH, a 3D visualisation of a browsing session. ERCIM News No. 41, April 2000 KEYNOTE

The future Web and Europe – the Web and future Europe

The World Wide Web Consortium (W3C) has gained a lot from ERCIM. The strong overlap between the membership of ERCIM and the W3C offices has made it very easy to set up the European offices; it has helped W3C to be connected into Europe. Concerning Europe, a frequently asked question is: Why did the Web take off more quickly in the US, when it was invented in Switzerland? There are many reasons. The Internet was far more deployed in the US than in Europe. When I invented the Web in 1989, the Internet had just got through to CERN. Many US universities were connected on the Internet, which enhanced a rapid spread of the Web across the USA. The American entrepreneurial spirit may also have something to do with it. There are many start-ups in America. The European ethos to doing things is discussing them first and then doing them. The huge difference in diversity between the US and Europe is a third factor. Europe does not have the same model language block. There is an incentive to put up a website in America, because so many people understand English.

Diversity is very valuable, and Europe should use this. We get a much richer set of ideas with Tim Berners-Lee is boundaries which allow people to think independently. At the same time, you need balance director of the W3C between diversity and homogeneity. ERCIM is part of that: it builds bridges between different consortium, and a cultures and allows technical ideas to move freely within the academic community and across Principal Research different borders. Scientist at MIT. He invented the Web in 1989, There are several directions in which Europe could now go. American industry seem to move and recently wrote the ahead faster, but Europe could still very well be ahead of the next wave, because it has a lot of book ‘Weaving the Web’. the know-how. There could be an excellent opportunity for Europe to address issues such as intellectual property rights, negotiation, and privacy. The Web is not yet done. The Web as you see it now, is just an information dissemination system for most people. I’d always wanted it to be a collaborative system, in which creating links, creating hypertext, is just as simple and intuitive as browsing currently is.

ERCIM provides a way of moving information and facilities between different people in different countries. It coordinates and facilitates, and the Web is a tool for doing that. Europe should think carefully about how to use the Web, in order to build a new society on top of that. Europe is going through a large change, finding out how to construct a society of many nations. The Web can give Europe more of a varied set of connections, which will make it very resilient and able to remain stable but at the same time to move very fast. This interesting general question I leave with you to think it over.

ERCIM News No. 41, April 2000 3 JOINT ERCIM ACTIONS

ERCIM launches the DELOS Network of Excellence for Digital Libraries

by Carol Peters

23 participants attended the kick-off meeting of the extending the activities of the DELOS Working Group new Network of Excellence for Digital Libraries, on Digital Libraries, and intends to promote European organised by ERCIM in Sophia Antipolis, 16-17 research in the digital library domain by providing a February 2000. The Network aims at continuing and set of ‘services’ for the digital library community.

The aim of the DELOS Network of The activities of the Network will be evaluation infrastructures: (i) a digital Excellence for Digital Libraries (DLs) is organised under Forums and will be open library test suite and (ii) a cross-language to provide an open context in which an to the European DL research world and system evaluation infrastructure. international research agenda for future the relevant application communities research activities in the digital libraries (electronic publishing, libraries, cultural Standardization Forum. W o r k i n g domain can be developed and heritage, archives, etc.). Groups will be set up to (i) examine continuously updated. The Network will specific clusters of relevant standards; (ii) constitute a reference point for all DL Digital Library Research Forum. This effect the transfer of research results into projects funded by the 5th FP-IST, activity will be aimed at stimulating the implementation; (iii) disseminate the stimulating the exchange of experiences advancement of DL-related research and results through workshop proceedings, and know-how in this multidisciplinary encouraging scientific collaboration reports, Web-based observatories, etc. domain, and will also establish close between European teams working in this contacts with relevant application domain through the organisation of: (i) Training and Technology Transfer communities. It will make testbeds thematic workshops, (ii) the European Fo r u m . This forum is aimed at providing available, facilitate their interoperability, Conferences on Digital Libraries (ECDL), training possibilities and demonstrating and provide mechanisms for the (iii) an international research exchange the effectiveness and benefits of DL evaluation of models, techniques, and programme. technologies within European industrial approaches, and the exchange of open- environments. Summer schools will be source software components. It will also Digital Library Evaluation Forum. DL organised on key issues in the DL domain. contribute to the definition of relevant evaluation implies defining new metrics standards. Appropriate take-up models for components and new combinations of International Cooperation Forum. will be studied and tested to facilitate the components in order to measure Collaborations will be established with exploitation of new DL technologies in performance aspects related to DL those countries for which a scientific the industrial/commercial environment. interoperability, metadata and resource cooperation agreement is defined under A dense network of links with the discovery, multilingual access, etc. the 5FP. In particular with the US international DL research community will DELOS will promote the building and (through NSF), Russia (through the be created. operation of publicly available reusable Russian Foundation for Basic Research), Mediterranean countries, and Japan.

The Network of Excellence has 28 member organisations: the academic community is represented by leading European research teams working on key topics in the digital library area from eight ERCIM institutions (CNR-IEI, CWI, ETHZ, FORTH, GMD, INRIA, SICS, MTA SZTAKI) and from nine other leading European academic organizations (Universities of Athens, Dortmund, and Mannheim, Norwegian University of Science and Technology, Technical University of Crete, Bavarian Research Centre for Knowledge-Based Systems, Milan Polytechnic, Austrian Academy of Sciences, Social Sciences Information Participants in the kick-off meeting. Centre - Bonn. The application community

4 ERCIM News No. 41, April 2000 JOINT ERCIM ACTIONS

is strongly represented by members from Stephen M. Griffin, Chris Rusbridge, CALL FOR CANDIDATES libraries, (National Library of Portugal, Makoto Nagao, Yuri Hohlov, Rudi Italian National Institute for Union Schmiede, representing US, UK, Cor Baayen Award 2000 Catalog), electronic publishing (Springer Japanese, Russian, and German national and Elsevier), education (Arts and DL programmes, respectively. The Board The Cor Baayen Award for the most Humanities Data Service - UK), archives will meet annually but will be kept in promising researcher in computer (Italian Ministry of Cultural Heritage), and continual contact with the work of the science and applied mathematics was broadcasting (BBC – UK). There are also Network through regular reports from the created in 1995 to honour the first two IT industrial members: Eurospider scientific coordinator: Costantino Thanos, ERCIM President. Any young Information Technology, Switzerland, and IEI-CNR, Pisa. researcher from one of the ‘ERCIM Intracom, Greece). As one of the guiding countries’ (Czech Republic, principles of the Network will be that of Links: Denmark, Finland, France, Germany, “openness”, delegates from other research Greece, Hungary, Italy, Norway, DELOS website: groups or interested communities will be http://www.iei.pi.cnr.it/DELOS/ Slovakia, Sweden, Switzerland, The invited to participate in activities of the Netherlands or the UK) can be Network as appropriate. Please contact: nominated. The award consists of a cheque for 5000 EURO. The selected Costantino Thanos – IEI-CNR An international advisory board has been Director – DELOS Network of Excellence fellow will be invited to the ERCIM set up to approve and evaluate the Tel: +39 050 315 2910 meetings in autumn. programme of activities planned by E-mail: [email protected] Network. Members of the Board are: Rules for Nomination • nominations for each country are made by the corresponding ERCIM Executive Committee member Joint European-US Strategic Workshops (national contact); those who wish to propose a candidate must contact by Jean-Eric Pin the ERCIM Executive Committee member for their country (see The European Commission and the US National Science Foundation will h t t p : / / w w w . e r c i m . o r g / c o n t a c t s / be organising a series of strategic research workshops to identify key execom/execom.). research challenges in Information Technologies. • nominees must have carried out their work in one of the fourteen ‘ERCIM The strategic workshops will be organised • long-term/high risk nature of the countries’ jointly under the auspices of the European research involved, justifying risk • on the date of the nomination, Commission’s IST-FET programme and sharing at an international level candidates must have been awarded the US National Science Foundation • high potential payoffs both in the EU their PhD (or equivalent) not older CISE division. Focus is on key emerging and the US that make up for the long- than 2 years areas of mutual interest on which a world- term/high risk nature of research • Each institute can nominate up to two class series of workshops would be • existence of sufficient scientific and persons from its country relevant. One outcome of each workshop technological bases in both the US and • a person can only be nominated once will be recommendations for future joint the EU to entail balanced research • the selection of the Cor Baayen European-US activities in the domains efforts. Fellow is the responsibility of the covered by the workshop. ERCIM Executive Committee. Scientists interested in this co-operation ERCIM has been chosen to lead the can submit their suggestions by a short Deadline European part of this activity and to solicit online questionnaire available at the Nominations are to be received by the proposals from the European research ERCIM website. national contacts by 30 June 2000. community. ERCIM is interested in bottom-up ideas for high-level workshops Links: Links: from the European scientific community. EU-NSF Joint Strategic Workshops Cor Baayen Award: Initiative: http://www.ercim.org/EU-NSF/ http://www.ercim.org/activity/ This activity is supervised by a Strategic cor-baayen.html Workshop Review Committee (SWRC) Please contact: composed of high-level independent Please contact: Jean-Eric Pin – ERCIM Office scientists from Europe and the US. The Tel: +33 4 9238 5010 Lubos Brim – CRCIM committee will select the themes by E-mail: [email protected] Tel: +420 5 4151 2323 criteria such as: E-mail: [email protected]

ERCIM News No. 41, April 2000 5 JOINT ERCIM ACTIONS

ANFAS: Data Fusion for Flood Analysis and Decision Support

by Simon Lambert

Several ERCIM partners are working together in a new develop a decision support system for flood prevention project under the European Information Society and protection, integrating the most advanced Technologies programme. The ANFAS project aims to techniques in data processing and management.

Large-scale flooding is a serious problem in many parts of the world. In Asia, countries such as China and Bangladesh regularly suffer, while in Europe many countries are at risk. There is danger of loss of life and serious damage to the economies of the areas affected.

For this reason, the ANFAS project has been set up to include partners from both the European Union, other European countries, and China. The partners bring together expertise in soil sciences, remote sensing, sceintific computing, computer vision, internet technology, geographic information systems, knowledge-based systems and other areas. The ERCIM partners are: • ERCIM itself in the project management role Floods regularly cause life loss and damage like here in France in 1999. The • FORTH, with expertise in Geographic aim of the ANFAS project is to develop a simulation and prevention tool to Information Systems limit flood damage. • SRCIM: Slovak Academy of Sciences, Institute of Informatics, one of the links with the end-users and also with of the ANFAS project is to develop a flood. If simulation is performed, time expertise in geophysical flow modelling simulation and prevention tool to help sequence maps tracing the propagation and high-performance computing decision-makers in taking decisions that of the simulated flood will be provided • CLRC, bringing expertise in will limit flood damage. • direct impact assessment, evaluation of knowledge-based assistance and the total affected area cooperative capabilities for information The functionality of the ANFAS system • regional and statistical analyses over systems will be: the affected area • INRIA, with expertise in geophysical • to perform flood simulation, simulating • simple socio-economic assessment of flow modelling and high-performance water flow propagation based on damage directly caused by flooding. computing. scenarios. The user will be able to interact with the scene that models the The final system will be an immensely The project team also includes three other real site in order to add/remove dikes valuable tool in planning national and European partners (Matra System & or other human constructions regional responses to the threat of Information, the French geological survey • to assess flood damage. Flood flooding, and will be adaptable to other and ore exploration - BRGM, Reading assessment can be done either using the areas. University) and three Chinese institutes simulation results or using the remote (Institute of Automation (IOA), Institute sensed images of a real flood event Links: of Atmospheric Physics (IAP), Institute • at a prospective level, to analyse Anfas website: of Remote Sensing Applications (IRSA), floodplain morphology including http://www.ercim.org/ANFAS/ of the Chinese Academy of Sciences). changes in riverbed and subsurface properties due to repeated floods at a Please contact: When there is a risk of flooding, decision given site. Bruno Le Dantec – ERCIM Office makers have to decide what are the most Tel: +33 4 9238 5010 appropriate actions to take: evacuation of The output of the system will be: E-mail: [email protected] the population, reinforcement of dikes, • visualisation: consisting of two- intentional breaking of dikes, etc. The aim dimensional maps of the extent of the

6 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

Web Technologies – Introduction by Robert Caillau

The current crop of articles in this issue’s special theme is all about the Web. Or is it? I read the articles with interest, sometimes a frown, sometimes a smile and sometimes a sigh.

The most interesting observation for me was to see how much SVG is on the horizon, it is non-proprietary and we may yet see XML has become a tool of choice for experimenting. This is it replace PostScript. So there is another smile. good news. When we were working on spreading the Web, HTML was clearly a second priority, and it remained in a pitiful RDF, XML and SVG all at some level are re-inventions of state from 1990 to 1993. It proved impossible to get people at wheels that have been turning in many places for a long time. large to understand that SGML with its DTDs was a better It is necessary to re-invent once in a while, because only that technology, so we had to wait until there had been enough way can we integrate, improve, get the youngsters enthusiastic, corruptions and style problems. That made people look for any and become non-proprietary. Smile! (SMIL?) good proposal, and XML was one. So that was a smile. I discovered a number of cats. Cats are independent agents over But the other thing I observed is that Java is the unchallenged which you have no control. If they like you, they may spend choice for doing the actual horsework: getting some semantics some time sitting on your lap and purr, giving you an impression into it all. Somehow semantics remains an untrappable ghost. of cosy well-being. There are several articles on systems that When I studied programming languages, in the days when are supposed to do things for you. But I see cats (mind you, I’m computers would still sit up, I always got confused as to where owned by a cat myself). others defined the boundary between syntax and semantics. Notation does not solve the problem. Only execution does. The Finally, there are some very interesting reports on techniques prime example of how real programmers define real semantics for improving efficiency and for applying the Web in social is C, that ugliest of all ugly systems. But as with many interaction. The range of this collection is wide, if you are an monopolies, one of C’s problems is that it works. Java should eclectic reader as I am, you will find a lot of stimulating stuff give us portable semantics. But does it? It’s all in the intractable here! classes. Yet, there is a lot of good stuff, and certainly programming feels orders of magnitude better in Java than in Robert Cailliau – Web Communications & Public Education, C (there are of course still those curly brackets). I always believed Education & Technology Transfer Division C E R N, European Organization for Nuclear Research that ultimately the Web needed a programming language in the Turing machine sense. This was sometimes frustrating to communicate, and many, including friends, went off into the blue yonder of declarative stuff, often so esoteric that no sane person would spend her time on it. But, be it implicitly, we now have a programming language, unfortunately a proprietary one. So that was a sigh.

When I first heard of the Dublin Core metadata, I got enthusiastic and asked someone to prepare a talk on the state of metadata. An interesting discussion ensued, though I was quite annoyed at how people thought it could all be very simple. So at one subsequent meeting I showed them the lid of a pot of mayonnaise as sold by the local supermarket. It had on it the date of fabrication, the sell-by date and the best-before date. How could the Dublin Core people think they could do it with a single date, if just the local grocery store had used three for decades? So that was a frown. But I am greatly relieved to see how much influence RDF has now obtained.

Another idea (there I go again...) that should have been with the Web from the start was . Before GIF, we (I) used PostScript for Web drawings. This was certainly 2D vector graphics, but PostScript was to some a word you could not say without conjuring up the devil, lightning and sulphurous smells. Then, to my great disappointment, GIF took over, but look,

ERCIM News No. 41, April 2000 7 SPECIAL THEME: WEB TECHNOLOGIES

Components for Data Intensive XML Applications

by Peter Fankhauser, Gerald Huck and Ingo Macherius

XML-based applications blur the lines between Internet, and data exchange scenarios. At GMD Institute for publishing and database technology. While classic Integrated Publication and Information Systems (IPSI), SGML was focused on presentation oriented publishing a component-based framework has been designed and applications, typical XML deployments in e-commerce implemented to allow for the rapid development of data bias toward message oriented middleware, transactions intensive, XML based applications.

The eXtensible Markup Language (XML) is the next generation data format for structured information interchange on the World Wide Web. Under the auspices of the World Wide Web Consortium (W3C), XML has grown into a family of standards integrating key technologies from three GMD-IPSI’s previously independent domains: middleware documents, databases, and the Internet. architecture This powerful mix is a strategic component for native, of the rapidly growing ‘dot com’ industry. data intensive A Native Approach to XML Processing XML Many of today’s XML technologies applications. originate from document processing, where scalability was not considered PDOM (Persistent Document Object proposal, and augments it with extensions important. In e-commerce, however, Model) is an object manager, which to cross-document joins and restructuring scalability is essential. The use of XML transparently maps standardized W3C- of results. Its robust and efficient mix of with database technologies overcomes this DOM API method calls into operations algebraic and physical query optimization problem, but raises others, such as data on binary files, enabling processing of techniques yields superior performance. model and query paradigm mismatches. XML documents far beyond main The processor can also be used on top of XML has a semistructured nature, which memory limits. Unlike traditional any W3C compliant DOM implementation, is incompatible with the flattened structure databases, the DOM does not require a including the PDOM. of relational tables, and is not easily schema and thus allows skipping the decomposable into objects. Queries do not design intensive setup phase typical of Component 3: capture the expressiveness of XML. traditional DBMS. The XML-Broker Data Server Both the XQL processor and the PDOM We designed middleware components, The implementation is scalable and were integrated into an XML data server, which overcome the mismatches without achieves a throughput of several MB of making their functionality accessible sacrificing the simplicity of XML. They XML data per second on a standard PC. through the HTTP protocol. Our provide two important DBMS PDOM is lightweight, in that the code extensive experience showed the capabilities: declarative queries and size is relatively small as well as importance of diversity in interfaces. Thus transaction safe persistence. Thus the programming effort is less compared to we support the low-level DOM API, components can cope with the weakly DBMS based solutions. string-based queries, URL with embedded structured, high volume XML data queries, and means to post-process query typically generated by wrappers from Component 2: results with XSLT (XSL Transformations). legacy data sources such as HTML pages. The XQL Query Processor Complicated information processing tasks The Extensible Query Language (XQL) on XML data can be concatenated into Component 1: The Persistent DOM is a declarative, path-oriented query pipelines, encoded in standard URL The (DOM) is language for XML. It includes most syntax – ready to be bookmarked and a platform- and language-neutral interface operations familiar to SQL, eg selection, reused. for XML standardized by the W3C and restructuring, joins, and views, and widely used throughout the industry. It handles the semi-structured nature of Applications: provides a standard set of objects for XML. Introduced first at W3C’s From Molecules to Markets representing XML documents, a standard conference on XML query languages in The first deployment of our components model of how these objects can be 1998, XQL has since been implemented was the nationally funded RELIMO combined, and a standard interface for by several large IT-vendors. IPSI’s query project, whose goal is the integration of accessing and manipulating them. The processor implements the complete XQL data sources of interest to drug designers.

8 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

This still evolving technology was chosen the GMD’s XML Competence Center. Links: by the IST funded OPELIX and eBroker This center has already become a focal projects as a platform for B2B e-commerce point for cooperation between industry, XML Competence Center: http://xml.darmstadt.gmd.de/ applications. The software package is also research partners and standardization XQL and PDOM download: being distributed on the Web. So far, more bodies such as W3C. Our state-of-the-art http://xml.darmstadt.gmd.de/xql/ than one thousand have been downloaded. know-how and proven research This popularity made its commercialisation background in database technology, Please contact: viable, and resulted in a product called the document management, publishing, Peter Fankhauser – GMD-IPSI Infonyte XQL Suite. information retrieval, and graphical user Tel: +49 6151 869 939 interfaces make GMD-IPSI a natural E-mail: [email protected] Research meets Business partner for advanced XML information The components discussed in this article management issues, and their application are part of an ambitious project at IPSI, in the evolving digital economy.

EXMO: Computer mediated Exchange of Structured Knowledge by Jérôme Euzenat

The World Wide Web has generalised document transformation of structured data. However, if this exchange on a world-wide basis. XML promises to do provides interoperability, it does not help for the same thing for the exchange, manipulation and intelligibility.

Exmo is an INRIA action which studies for the preservation of intelligibility the concurrent engineering framework), the communication of formalised through transformations. it is necessary to apply treatments which knowledge among people mediated by do not challenge the current stage of computers. In the communication process, A transformation is a computational way development. It is thus relevant to warrant the computer can add value to its medium of generating representations from other the preservation of content. and memory functions by performing representations (not necessarily in the advanced operations such as formatting, same language). Transformations are used One outcome of this work is the capability filtering, categorising, consistency for putting representations together or to decide, given an organisation checking, or generalising. The goal of generating specific representations for a composing several transformations, and Exmo is the development of theoretical particular need. The goal of the the properties of these transformations, and software tools for helping the transformation topic is the elaboration of the properties satisfied by the system as a organisation, manipulation, presentation a ‘general theory of transformations’ whole. This should be a precious tool for and combination of structured knowledge based on the properties that the future information system architects. chunks though allowing communication transformations satisfy (rather than on the between people. The knowledge is representations or the transformation More generally, given a set of represented in a formally defined themselves). Such properties include transformation types and a set of language (which can range from XML - content or structure preservation, source transformation composition operators, structured, general but without semantics traceability, or conversely, confidentiality. one goal is to establish if a property can - to knowledge representation languages For instance, it can be useful to define a be decided for an instance of this. Exmo - structured, semantically characterised, transformation which delivers a does not aim at helping the but with a narrow scope of expression). documentation to a customer but which implementation of transformations or at hides some information from the initial inferring properties from implementations. In order to warrant the suitability and source (eg, if a design document includes However, some (generally simple) intelligibility of knowledge for the users, cost study, it is not advised to transformation languages implement only this requires to develop an abstract communicate this information to sub- one specific interaction type (eg filter) and understanding of representations and the contractors). In the context of thus our work is connected to transformations applied to them. The collaborative work, it is useful to establish transformation languages in that way. research of the Exmo action is focussed the properties of such a filter in order to in two directions. The transformation know where to implement it in the The communication topic aims at topic accounts for the modification of organisation. Conversely, when the contributing to the proper understanding knowledge during the communication elaboration of a representation is a of the knowledge by people. Formal while the communication topic accounts collaborative and continuous process (in semantics developed for knowledge

ERCIM News No. 41, April 2000 9 SPECIAL THEME: WEB TECHNOLOGIES

representation systems is suited to the use result. Hence, when a transformation One step further on this line leads to the of a system by one user because it is operates a translation between formal consideration, together with structure and expected that (s)he interprets in a coherent languages, good understanding cannot be semantics, of interpretation policies. way the terms (identifiers) used. However, ensured by meaning preservation. A when several users communicate, this semiotic treatment must be applied for These concepts are applied to technical understanding becomes problematic. ensuring the meaning reconstruction. This memories and representation of web Work developed in the field of consensual treatment comes as a complement to the document content. However, it will spread ontology construction could help to solve sheer semantic treatment used in the to all the aspects of information systems the problem of term interpretation. knowledge representation area. in the future (especially in the context of However, other problems are involved. the generalisation of computer supported For instance, a user can express In order to contribute to the solution of collaborative work and the development knowledge under the form of class this last problem, we investigate the of the so-called ‘semantic web’). hierarchies and first order clauses and then embedding in XML of the semantics of communicate it by using an the language used. Thus, the form can be Links: interoperability language. But if this last preserved as much as possible (form is http://www.inrialpes.fr/exmo/ language expresses the whole knowledge relevant to human understanding) and with clauses (though preserving the interoperability is ensured through the Please contact: semantics of the assertions), the initial availability of the semantics (correct user will hardly recognise (and hardly computer treatment mainly depends on Jérôme Euzenat – INRIA Tel: +33 4 7661 5366 understand) the semantically equivalent se m a n t i c s ) . E-mail: [email protected]

Escrire: Embedded Structured Content Representation in Repositories

by Jérôme Euzenat, Rose Dieng and Amedeo Napoli

Content representation seems unavoidable in some candidates as content representation languages, their areas of the future web. Although there are many respective merits in this context are not yet known.

An Intranet or, more generally, the use of ESCRIRE is a coordinated action of three queries. Each team specifies the the Internet technology, is an opportunity INRIA teams (Acacia, Exmo, Orpailleur). integration of the formalism in XML and for companies, to publish and share One of its first objectives is to propose an develops a query evaluation strategy knowledge often difficult to reach in implementation model for this. But there depending on the considered formalism. documentary form. The numerical and are various knowledge representation The formalisms together with full-text digitised documents can be made available formalisms and their respective qualities search will be evaluated along a in a standard and transparent way to all (compared to each other) are not exactly predefined protocol for assessing the users concerned. The ambition, in the known. The main goal of ESCRIRE thus qualitative criteria (eg query expressivity, long term, is to produce knowledge servers consists of comparing three types of legibility) and quantitative ones (eg allowing the search and the handling of knowledge representation formalisms precision, recall). This evaluation will the corporate resources. However, the (conceptual graphs, object-based provide a precise analysis of the strengths limits of this approach appear quickly: the knowledge representations and description and weaknesses of each knowledge organisation and maintenance of the sites logics) from the standpoint of the representation formalism of content appears an expensive task and full-text representation and the handling of representation. search is not very effective. document content. That will enable to highlight the desired properties and to Links: The representation of the content of evaluate qualitatively and quantitatively http://www.inrialpes.fr/exmo/cooperation documents is becoming a necessity. the performances of the implementation /escrire/ Content representation allows to formalisms. Beyond the better knowledge manipulate content, to make search by of the techniques implied by the various Please contact: analogy, by specialisation, similarity, etc. projects, this work will advance the state Jérôme Euzenat – INRIA XML enables to insert content of the art. Tel: +33 4 7661 5366 representation (through RDF) within the E-mail: [email protected] documents (in XHTML or other XML A set of documents and a set of queries format) and knowledge representation have been selected in a coordinated way. formalisms are good candidates for The efficiency of the respective formalisms representing content. will be assessed with regard to these

10 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

Workflow Demonstrator by Damian Mac Randal

As part of the World Wide Web Consortium’s Leverage system was developed to explore XML/RDF and their Action project (W3C-LA), several small demonstrators interaction with other client side protocols. The were developed to promote European adoption of the demonstrator focuses on delivery of the workflow Web and W3C protocols. At RAL, a pilot workflow steps from multiple servers to remote actors.

Business processes are now recognized technologies, by building on the W3C can be sensitive, restricted to specific as being key to thinking about developments in HTML, DOM, XML actors identified by the process. There is organizations, their effectiveness, and and RDF (Metadata) in particular, also a significant amount of content about their ability to cope with change. IT together with the portability features the activity, such as the DueDate, support for business process is provided offered by Java and the servlet API. containing process, status, etc. This by workflow management technology: Workflow is seen as a challenging metadata is useful to the actor typically, a workflow system will provide application area that will explore many independently of the activity itself. Since support for each of the activities in a of the emerging Web technologies, such most actors will have many activities process by offering the necessary work as push technology, security, client-side allocated to them, some sort of Intray is (and information needed to complete it) processing, etc. required, listing outstanding activities and to those staff with the correct attributes allowing selection of the next. Access to to undertake that work. A server retains The W3C Workflow demonstrator metadata, such as DueDate, is clearly the information on the process focuses on the delivery of activity essential to prioritize selections. Equally specifications and the state of each information to the appropriate actors as clearly, the Intray has to be kept up-to- process which is currently active. Major required by the process, in a timely, secure date. issues in workflow management include and useful manner while minimizing the portability, coping with the network and server loads. The demonstrator uses a 3-tier geographical distribution of users, architecture: relationships with organizational Activity information consists of • a database holding the process and Intranets, and consistency of access. instructions as to what has to be done, and activity data, state information and actor ‘parameters ‘, that is data items required information. All workflow management issues are to carry out the activity and the data • Java servlets (on a Web server) addressable by Web server / browser generated by the activity. Some of this providing the process ‘engine ‘

Screenshot of the demonstrator Intray and Activity pages.

ERCIM News No. 41, April 2000 11 SPECIAL THEME: WEB TECHNOLOGIES

managing the delivery of activities to Web server. By stopping the mouse appropriate connects the ‘Go ‘ button to actors and the process flow pointer over a list item, the full metadata the required application. When the actor • scripts on the actor’s Web browser for that activity can be displayed in a is finished, clicking the ‘done ‘ button providing local processing of the Intray popup window. Finally, so-called push causes the results form to popup. Clicking and Activity pages. technology (actually a scheduled ‘send’ on this sends the results back in check/refresh) is used to keep the list up- XML to the workflow engine and closes Intrays to-date. the popup and activity window. The An Intray is a dedicated page (or desktop) workflow engine saves the results, marks displaying an up-to-date list of the Activity Pages the activity as completed, identifies the activities allocated to the actor. The list Clicking on an activity in the intray next activity and ‘pushes’ it to the provided by the Web server uses RDF to indicates to the workflow engine that the appropriate Intrays. record the metadata about each of the actor intends to work on that activity. A activities. The Intray page, or more new Activity page is created using an Please contact: precisely, scripts in the page then use this XML description of the activity sent by Damian Mac Randal – CLRC RDF to display a scrollable list of the the Web server. (The Activities metadata Tel: +44 1235 44 5403 activities, highlighting overdue items. is also sent as RDF embedded in the E-mail: [email protected] Further scripts allow the actor to sort and XML). The Activity page displays the filter the list without going back to the Activity instructions and data, and if

Promoting XML in Italy

by Laura Abba

During 1999, the Institute for Telematic Applications groups. One of the objectives of XML-Italy is to of the Italian National Research Council (IAT-CNR) promote R&D collaborative activities between Italian promoted the creation of XML-Italy, an open working research and industry. group similar in constitution to the Internet technical

XML-Italy’s primary goal is to coordinate the meetings and through the ‘group In addition to XML-Italy, IAT-CNR is existing national expertise in order to communication’ system via electronic also currently responsible for several spread XML know-how in an organised mail at the address: [email protected]. More research programmes on technologies for manner. Technicians, researchers, than 230 public and private Italian the processing, exchange and protection operators and vendors interested in XML organisations subscribe to this list and the of information. Languages and models have joined the group, and it has become average traffic is about 4 e-mails per day. for the representation of the information a point of reference for the Italian Technical problems – such as XML content of data files are being studied for scientific and industrial community. The parsers, CR/LF treaty in XML, use by the next generation browsers and group facilitates the approach to the new applicative co-operation using XML web servers, in order to facilitate access technologies, and thus contributes to a instead of CORBA, parameter transfer in to distributed data and surfing on the net. more efficient and safer management of , WAP and XML files, MS-WORD The considerable changes that have been information on the net. document representation in XML, etc - made in the architectures of information are debated. Requests for assistance and systems, following the rapid evolution of The Italian XML Group has four main information on markup languages, the de facto standards, are also being goals: particularly XML, for publishing, data investigated with particular attention. The • to disseminate know-how on a national exchange, and the visualisation of technologies used for this line of research basis through the organisation of complex and structured documents on the are WEB, JAVA, and XML. technical transfer activities net are frequently received. • to facilitate the exchange of ideas via Links: e-mail and meetings Many of the most relevant research activities Activities of XML-Italy: http://www.xml.it/ • to create an XML information bank or and applications of XML technologies in reference point both for developers and Italy were presented last June in Bologna, Please contact: for users at the first XML-Italy w o r k s h o p • to demonstrate the new technologies (h t t p : / / w w w . x m l . i t / b o l o g n a . h t m l ) . More Laura Abba – IAT-CNR Tel: +39 050 315 2633 through the development of prototypes than 200 researchers attended this event. E-mail: [email protected] and simple applications. The next XML-Italy meeting will be held in Pisa, at the Research Area of the National Participation in the activities of the group Research Council, 9-10 May 2000. is open. The work is carried out during

12 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

The XMLC Browser by Fabio Vitali

XMLC is a modular browser for XML documents. It is the New Jersey Institute of Technology (USA) in 1996, the result of a project born initially to display generic has now become an important enterprise, involving extensions of HTML on Web browsers, and later faculty and graduate students of the University of became a browser for arbitrary XML documents. This Bologna with additional support from the activity, which began as a simple student project at DaimlerChrysler research centre in Berlin.

The World Wide Web has developed in an idiosyncratic, difficult-to-predict fashion. For instance, the Web community has valued the development of standards and protocols over that of functionalities. This has lead to the creation of some dozens of different languages and protocols that are necessary to master the task of creating satisfactory websites.

In our opinion this richness of languages is clear evidence that while, on the one hand, there exists the possibility of implementing a large number of interesting The XMLC functionalities, on the other hand, environment unfortunately the Web does not enforce or showing a even encourage their use. This is left up to complex the willingness and awareness of individual XML authors of Web pages and sites. document.

The XML family represents a considerable and provide the rendering behaviour for We have applied the displet approach to advancement over previous languages and that element. Support for the most hypertext functionalities, implementing standards. The possibility given by XML common element types is provided (for full support for XPath, XPointer and to define a syntax (ie, a Document Type instance, text elements and paragraphs) XLink, the tripartite hypertextual Definition, or DTD) tailored for one’s own but, at any time, it is possible to add new architecture of XML, and also in software document classes, and to use standard modules enabling specialised rendering engineering applications, providing XML tools to create, verify and exchange semantics for specific needs. displets for Z specifications, UML data is a real bonus. Additionally, XSL schema, finite state machines in the adds much to XML in terms of reach and Our implementation of displets is based StateMate form, and Petri nets. flexibility, including a mapping language on the XMLC prototype. XMLC can be that can be used to transform one XML considered as a very general architecture In our opinion, XMLC is the most document into another. to provide very sophisticated customizable and expandable architecture functionalities to documents created in the for displaying XML documents. Full Our long-term aim is to create an XML format. However, the architecture information on the XMLC browser and environment that, while relying on several of XMLC can be fruitfully used for more references to additional literature can be existing Web languages and protocols, than visualisation, as it provides an found on our website. can provide fundamental and substantial extremely general way to associate hypertext functionalities in a streamlined behaviours with XML elements, and thus Links: and easy way. Currently we are to produce active documents that perform XMLC website: concentrating on browsing and displaying computations, enact goals, produce results. http://www.cs.unibo.it/projects/displets/ hypertext data. In fact, the XMLC prototype uses XSL styles to associate XML elements of the Please contact: In order to provide flexible support for source document with Java classes that Fabio Vitali – University of Bologna the special rendering needs that authors can be displayed within a generic applet, Tel: +39 051 2094 872 often have, we propose the use of the XML Manager. Furthermore, these E-mail: [email protected] ‘displets’. Displets are software modules classes can be made to respond to events, (currently Java classes) that are associated and to be activated in any manner decided with each element of an XML document by their implementers.

ERCIM News No. 41, April 2000 13 SPECIAL THEME: WEB TECHNOLOGIES

Using XML for Data Management

by Victoria Marshall, Brian Matthews and Kevin O’Neill

Recognising the limitations of HTML for the define their own formats for structured data whilst representation, discovery and exchange of structured allowing easy and open integration with other XML data, the World Wide Web Consortium (W3C) has formats. XML has been widely taken up as a potential developed an alternative, XML, together with a set of enabler of a wide variety of applications. related standards. This allows user communities to

XML is an open and simple format for any other XML format. Thus graphics can Different User Views representing structured data tailored to be described in a flexible yet compact XML is a purely structural format; there the Web, and as each user community can format compared with, say, binary is no semantics provided for defining its provide its own format, there is no loss formats such as GIF or JPEG. appearance. This separation of structure in semantics. Many user communities and presentation is an important feature have been developing XML formats for Common Interfaces to Data Sources of XML: the data may not be displayed the exchange of information including XML can define schema formats for at all, but passed straight to some MathML for mathematics, Chemical classes of documents. These XML program, or target database, so Markup Language (CML) for chemistry, schemas can be used to define common presentation information may be superfluous. Also, by providing presentational information by means of a stylesheet, different views can be presented on the same data, for example, data in trees, tables, graphs, or charts. These views can be tailored to the user requirements. Using the new W3C recommendation XSL, a high degree of interactivity with the data can be provided on the client, without requiring further interaction with the server.

ISE has been experimenting with providing different views on the laboratory personnel database (see figure 1). This allows the user to explore the structure of the organisation without downloading further date from the server. The same data set also can be used to generate alternative views by simply Figure 1: A view on a personnel Figure 2: Alternative view on the changing the colours and interface, or by database. personnel database. combining the data with SVG to, say, provide an annotated map of the laboratory offices (see figure 2). Bioinformatic Sequence Markup interfaces to live data sources, Language (BSML) for the human genome independently of the concrete Links: project, Extensible Scientific Interchange representation of the repository. This http://www.itd.clrc.ac.uk/ Language (XSIL), DDI for Social Science raises issues in using the hierarchical Data, and Schematic Vector Graphics structure of XML, essentially derived Please contact: (SVG) for diagrams. from the structured document community, to represent relational database semantics. Victoria Marshall – CLRC Tel: +44 1235 44 6799 The latter has been developed by ISE at XML is a relatively poor language for E-mail: [email protected] RAL working with W3C (see article on expressing concepts such as primary keys page 24). SVG uses graphical primitives or constraints. Nevertheless, the common such as boxes, lines, and circles which interface layer it offers means that it can can be integrated with other XML be a powerful mechanism for exchanging formats. An SVG graphic fragment can data between heterogeneous data sources. be embedded within an HTML document, and can itself have labels which are of

14 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

Corporate Knowledge Management through Intranet and Internet by Rose Dieng

The ACACIA research team at INRIA-Sophia-Antipolis, management ie, for building, managing and aims at offering methodological and software support distributing a corporate memory. in form of models, methods and tools for knowledge

The ACACIA team studies the problems It enables to exploit projection on deductions. This engine was applied to raised by the dissemination of knowledge conceptual graphs in order to retrieve the make deductions using a XML-expressed through a knowledge server via the adequate XML documents semantically knowledge base in road accident analysis. company’s Intranet or via the Web: we annotated by RDF metadata thus consider the Web as a privileged way for translated into conceptual graphs. This Moreover we are studying an exploitation supporting management of knowledge work is implemented in JAVA in the of RDF and RDF Schema for taking into distributed either inside a company or OSIRIX tool. We are studying account the notion of viewpoint. The between several companies (for example exploitation of complex inference purpose is to enable to represent an in the framework of a project memory or mechanisms on conceptual graphs in ontology with multiple viewpoints, of technological monitoring). We aim at order to enable different kinds of corresponding for example to multiple building knowledge servers enabling approximate search of information. professions in a company. search for information in a heterogeneous corporate memory, this search being We also developed an inference engine Applications ‘intelligently’ guided by ontologies and able to perform elementary, event-guided, We developed RESEDA, a system aimed ontological annotations. We focus on the processings, on XML documents: at supporting road accident analysis. case of a corporate memory materialised consistency management, elementary Installed on the intranet of INRETS, it in the form of XML documents. The ontologies can then be expressed either in RDF Schema, or in Sowa’s Conceptual Graphs or in CML (CommonKADS Modelling Language).

Research on Knowledge Servers We developed WebCOKACE, a knowledge server for distributing CommonKADS expertise models on the Web. It enables to explore expertise models and to perform hypertextual navigation in expertise models and towards associated electronic documents thanks to a Web browser .

Research on XML Galaxy Considering that information search in the memory can be guided by knowledge models and semantic metadata, Acacia studies information search in XML documents semantically annotated by RDF metadata or by conceptual graphs. We built a translator transforming an ontology expressed in CommonKADS, into an RDF schema that may serve as an annotation model of the corporate XML documents. We built an ontological filtering engine to answer simple requests of the user. We proposed a model of interpretation of RDF and RDF Schema in terms of conceptual graphs. We developed a translator of RDF WebCokace, a CommonKADS Expertise Model Web Server: example of Schema/RDF towards conceptual graphs. search.

ERCIM News No. 41, April 2000 15 SPECIAL THEME: WEB TECHNOLOGIES

enables the investigators of this institute Cooperative projects teams of INRIA (See articles on pages 9 to fill the accident dossiers in a The ACACIA project takes part in two and 10). This project aims at comparing computational form, with an intelligent IST projects that started officially in three types of knowledge representation support. The check-lists constituting an February 2000: formalisms (conceptual graphs, object- accident dossier are reconstituted from a • COMMA (Corporate Knowledge based knowledge representation and generic XML document and from data Management through Agents) aims at description logics) from the standpoint of stored in a database. According to the data building a multi-agent system, made of the representation and the handling of entered about the current accident, several agents cooperating for document content. RESEDA exploits a knowledge base information search in a corporate (written in XML format) in order to memory materialised by XML Last, the AC A C I A project takes part in provide the user with suggestions useful documents, these agents having the Working Group on Agent-Based for his task of accident analysis. These capabilities to adapt to the user. This Mediation Among Information Systems, suggestions stem from a knowledge base project involves ATOS, CSELT, Data, and Knowledge Management, in that was built thanks to acquisition and Deutsche Telekom, INRIA, LIRMM the framework of the AgentLink SIG on modelling of knowledge of experts of and University of Parma Intelligent Information Agents. INRETS. RESEDA also exploits a base • C-WEB (Community Webs), that aims of generic scenarios (written in XML at designing a generic platform to Links: format) in order to suggest the most support community-Webs. This project ACACIA website: http://www.inria.fr/ plausible generic scenarios that may be involves EDW-International, ERCIM, Equipes/ACACIA-eng.html applicable for the reconstitution of the ICS-FORTH, and INRIA. WebCokace: http://inria.fr/acacia/Cokace current accident analysed by the user. The scenarios are visualized thanks to XSLT The ACACIA team also takes part in the Please contact: stylesheets. RESEDA illustrates an cooperative project ESCRIRE Olivier Corby and Rose Dieng – INRIA application combining knowledge (Embedded Structured Content Tel: +33 4 9238 7871/7810 engineering and XML-based technologies. Representation In REpositories) with E-mail: [email protected], EXMO and ORPAILLEUR, two other [email protected]

Multilingual Metadata to access Social Science Data

by Brian Matthews and Michael Wilson

Most countries have a national archive for social government. These archives are usually a few hundred science data. This is data about social attitudes, and gigabytes in size, and accessible partially via the financial and environmental details, which describe Internet and the Web. However, it is becoming the current and past state of each nation. This increasingly important to allow access across national information derives from government statistics, and and linguistic boundaries so that decision makers through academic studies, often commissioned by have a comparative picture of European society.

The European funded LIMBER project, LIMBER will achieve this by providing terms in a multilingual search, relevant which began in January 2000, brings a uniform metadata description. Metadata data across the archives can be discovered. together CLRC-RAL, Intrasoft, the UK allows the explicit specification of the national archive at Essex University, the semantics of data, relationships between Indexing Norwegian national archive at NSD, and data, and its quality (recency, accuracy Creating metadata usually takes three other national archives within etc). We describe LIMBER’s key features. considerable effort, and whilst its benefits Europe The vision behind LIMBER is the are evident to users, data contributors interoperability of data, for example, from Multilingual Thesaurus have little incentive to spend this effort. WHO Health archives, with social science Using a controlled vocabulary to index To ameliorate this, an automatic tool will datasets on behaviour, and genetic metadata increases the relevance to index the metadata, by scanning for datasets from the Human Genome retrieval and when this is structured in a relevant terms and converting these to the Project. These can be integrated together thesaurus it further helps to refine controlled vocabulary, in the language of to show the localisation of potential searches. LIMBER will extend existing the metadata. genetically abnormal populations. This thesauri by using an XML representation would then be presented through a and adding ‘equivalent terms’ in other Multilingual User Interface multilingual interface so that it can used languages. This provides the terms to The multilingual thesaurus will allow for policy making and planning. catalogue datasets, and, using equivalent users to perform searches using their own

16 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

language. To refine their search, the Description Framework (RDF). This science. RDF also offers a uniform thesaurus will be presented in the user’s defines a model, syntax, and a vehicle to manage multilingual thesaurus natural language. The relevance of data representation of schema to capture as full ontologies. Further, common RDF returned, can be shown by displaying semantics in terms defined by format allows the seemless integration terms from the controlled vocabulary in authoritative bodies. RDF is defined in with metadata from related fields, such the user’s own language. XML, so as browsers are supporting as geographical, environmental, and XML, they are also supporting RDF, health data. Metadata using RDF providing economic and easy to use tools. Metadata is a very active area with many Links: proposed schemes. The problem with LIMBER will develop a metadata model http://www.linglink.lu/hlt/projects/limber/ standardising metadata definitions, is to (in XML and RDF) for social science achieve the correct layering, so that datasets to allow their integration within Please contact: domain specific elements are defined on and across archives, building on the a generic base technology. Recognising existing work of the Data Documentation Michael Wilson – CLRC Tel: +44 1235 44 6619 this, the World Wide Web Consortium Initiative (DDI) from University of E-mail: [email protected] (W3C) has defined a recommendation for Michigan, the most advanced metadata web-based metadata, the Resource structure currently proposed for social

ICS-VRP: a Tool for Parsing and Validating RDF Metadata & Schemas by Karsten Tolle and Vassilis Christophides

The Web provides a simple and universal content of the various Web resources (eg data, infrastructure to exchange various kinds of documents, images, etc.) and for different purposes. information. In order to share, interpret, and The emergence of the Resource Description manipulate information worldwide, the role of Framework (RDF) is expected to enable metadata metadata is widely recognized. Indeed, metadata allow interoperability across different communities or us to easily locate information available in the Web, applications by supporting common conventions by providing descriptions about the structure and the about metadata syntax, structure, and semantics.

More precisely, it provides a) a Standard descriptions in a both human readable and only the well-formedness of RDF resource Representation Language for Web machine understandable form. Many descriptions according the W3C RDF metadata; and b) a Schema Definition information providers like ABC News, M&S specifications. For this reason, we Language (RDFS) to interpret (meta)data CNN and Time Inc., Web portals like have developed the ICS-FORTH using specific class and property Open Directory as well as Web browsers Validating RDF Parser (VRP) allowing hierarchies (ie vocabularies). Moreover, like Netscape, and search engines like the validation of RDF resource descriptions RDF/RDFS offer a syntax for representing Altavista, Yahoo and Webcrawler already against the associated RDFS schemas, as metadata and schemas in XML, enabling support the RDF proposal. Unfortunately, well as of the schemas themselves. the creation and exchange of RDF existing RDF parsers (eg, SiRPAC) check

ICS-FORTH Validating RDF Parser.

ERCIM News No. 41, April 2000 17 SPECIAL THEME: WEB TECHNOLOGIES

RDF is based on a directed graph model The most distinctive feature of VRP is its import the external statements we need that alludes to the semantics of resource ability to verify the constraints specified to validate our RDF descriptions. Note description. The basic idea is that a in the RDF Schema specification. This that the RDF and RDFS namespaces are Resource (identified by a URI) can be allows the validation of both the RDF necessary for every RDF description and described through a collection of descriptions against one or more RDFS therefore their statements are by default Statements forming a so-called RDF schemas, and the schemas themselves. included into VRP. Description. A specific resource together The VRP validation module relies on (a) with a named property and its value is an a complete and sound algorithm to Currently VRP provides a command line RDF statement. RDFS schemas are then translate descriptions from an RDF/XML interface with various options to generate used to declare vocabularies, ie form (using both the Basic and Compact a textual representation of the internal collections of classes and properties, that serialization syntax) into the RDF core model (either graph or triple based). In the can be used in resource descriptions for model (ie triples) (b) an implementation future, we plan to implement a graphical a specific purpose or domain. VRP is a of this model in Java to efficiently verify user interface to visualize the analyzed tool to analyze, validate and process RDF the RDFS constraints. RDF statements as well as to interact with descriptions based on standard compiler VRP during parsing/ validation. Finally, generator tools for Java, namely To favour metadata reusability, RDF there is an ongoing effort to develop VRP CUP/JFlex (similar to YACC/LEX). As supports a) the sharability of RDFS APIs in order to facilitate the integration a result, users do not need to install schemas using the XML namespace of the program into other systems (eg on- additional programs (eg, XML Parsers) mechanism (ie provide only incremental line loaders to DBMS). in order to run VRP while they can easily modifications to a base schema in order update or extend the VRP BNF grammar to create a new variant); and b) the Links: in case of changes in the RDF/RDFS creation of RDF (meta)data using ICS-VRP website: specifications. VRP is a 100% Java(tm) multiple schemas at the same time (ie http://www.ics-forth.gr/proj/isst/RDF/ development understanding embedded merging different types of metadata). This RDF in HTML or XML and providing implies to maintain for real scale Please contact: full Unicode support. The quick LALR applications, several interconnected RDF Vassilis Christophides – FORTH-ICS grammar parser (ie CUP) as well as the schemas than can be potentially used to Tel: +30 81 391628 stream-based parsing support (ie JFlex) describe Web resources. To meat these E-mail:[email protected] ensure good performance during the requirements VRP supports validation Karsten Tolle – Universität Hannover processing of large volumes of RDF across several namespaces: we can E-mail: [email protected] descriptions. connect to remote namespaces in order to

XHTML 1.0 for Device-Independent Web Access

by Steven Pemberton

The World Wide Web Consortium (W3C) has that a specification is stable, contributes to Web released the XHTML 1.0 specification as a W3C interoperability, and has been reviewed by the W3C Recommendation. This new specification represents membership, who favour its adoption by the industry. cross-industry and expert community agreement on XHTML 1.0 was developed in a W3C working group the importance of XHTML 1.0 as a bridge to the Web chaired by CWI researcher Steven Pemberton. of the future. A W3C Recommendation indicates

HTML currently serves as the lingua In designing XHTML 1.0, the challenge browsers and that may be processed by franca for millions of people publishing was how to design the next generation XML-enabled software as well. Authors hypertext on the Web. While that is the language for Web documents without writing XHTML use the well-known case today, the future of the Web is making obsolete what is already on the elements of HTML 4 (to mark up written in W3C’s Extensible Markup Web. The answer was to take HTML 4.0, paragraphs, links, tables, lists, etc.), but Language (XML). XML is bringing the and rewrite it as an XML application. In with XML syntax, which promotes Web forward as an environment that simple terms that means making the markup conformance. better meets the needs of all its HTML a well-formed XML document participants, allowing content creators to where start and end tags are always there The benefits of XML syntax include make structured data that can be easily and match precisely. Empty elements extensibility and modularity. With processed and transformed to meet the have to use the correct syntax. XHTML HTML, authors had a fixed set of varied needs of users and their devices. 1.0 allows authors to create Web elements to use, with no variation. With documents that work with current HTML XHTML 1.0, authors can mix and match

18 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

known HTML 4 elements with elements rendered using independent style sheets World Wide Web. In 1988 CWI was from other XML languages, including such as CSS style sheets. XHTML 1.1, Europe’s first non-military Internet site, those developed by W3C for multimedia already under development, coupled with and became a Dutch and European (Synchronized Multimedia Integration device-specific style sheets and Composite backbone site. Since then it was involved Language – SMIL), mathematical Capability/Preference Profiles (CC/PP) - with the development of the Web expressions (MathML), two dimensional a protocol which allows a user to describe programming language Python, HTML, vector graphics ( both user preferences and device CSS (Cascading Style Sheets), and SMIL, – SVG), and metadata (Resource capabilities - will bring mobile and other the Web’s multimedia language, as well Description Framework – RDF). devices to the Web as full participants. as spinning off companies such as NLNet that set up the initial Dutch Internet In addition to its extensibility, moving The XHTML 1.0 Recommendation was infrastructure, and General Design, the from HTML to XML via XHTML 1.0 written by members of the HTML first Dutch Web Design company. lays the foundation for making Web working group, which includes key content available to millions more users. industry players such as Ask Jeeves, Links: People browsing the Web with cell CNET, Gateway 2000, CWI, GMD, http://www.w3c.org/ phones or other mobile devices want Web Hewlett-Packard, HTML Writers Guild, content tailored to their needs. People IBM, JetForm, , MITRE, Please contact: with disabilities need ways to transform Philips Electronics, Phone.com, Quark, content into accessible formats. Stack Overflow, Sun Microsystems, and Steven Pemberton – CWI Tel: +31 624 671 668 WebTV Networks. E-mail: [email protected] XML documents can already be transformed using Extensible Stylesheet CWI always has been at the forefront of Language Transformations (XSLT), and developments on the Internet and the

Amaya: a Testbed for Web Protocols and Formats by Irène Vatton

Amaya is the World Wide Web Consortium’s ( W3C) of being a testbed for experimenting with, testing and client which acts as both a browser and an authoring demonstrating new specifications and extensions of tool. It has been designed with the primary purpose Web protocols and formats.

Amaya is the World Wide Web Integration is the word that best MathML[3], and SVG[4] elements, but Consortium’s ( W3C) client. It provides characterizes Amaya. Since the also provides editing functions for creating a complete Web browsing and authoring beginning, Amaya adopted the spirit of them. Figure 1 shows an Amaya window environment and has a WYSIWYG style the Web as a distributed collaborative where the three precedent elements are of interface, similar to that of the more medium and it integrated seamlessly the combined. XML namespaces are used to popular commercial browsers. With the two functions: authoring and browsing. distinguish between each set of elements. extremely fast moving nature of Web At any time, a user is able to edit any technology, Amaya plays a central role document that s/he is browsing, without Styling at the Consortium. Easily extended to having to resort to another tool. As Amaya has extensive support for the integrate new ideas into its design, Amaya Amaya is able to work on several open W3C’s Cascading Style Sheets (CSS[5]) provides users with many specialized documents, it’s very easy to copy/paste and offers a simple-to-use interface features including multiple views, where information from one page to another enabling users to write style sheets for the internal structural model of the page or to link pages by a simple click. Web documents without having to know document can be displayed alongside the Moreover, the user is able to publish the details of the CSS syntax. You can browser’s view of how it should be documents on Web servers using the create colored text, change the font, set presented on the screen. Amaya has a standard HTTP PUT method. the background color and control other counterpart called Jigsaw[1] which plays presentational effects. CSS style sheets a similar role on the server side. Nested Structures in Amaya may also be used to alter the look of With the introduction of XML, it’s now graphics or mathematical expressions. An important benefit of using Amaya as possible to create very rich documents on Amaya provides several features for an editor is that it implements the W3C the Web, mixing text, graphics and testing, reusing, and publishing CSS style specifications very carefully, thus mathematical expressions. Amaya allows sheets. You can also temporarily apply or guaranteeing the interoperability and users not only to browse these rich remove a CSS file to/from a document in legacy of the markup it produces. documents including HTML[2], order to test its effect.

ERCIM News No. 41, April 2000 19 SPECIAL THEME: WEB TECHNOLOGIES

Web Accessibility imagemap areas are already built in - it is Conclusion Amaya can be used to produce content impossible to add an image until it has Amaya has been an useful tool for early that conforms to W3C’s Web Content alternative text. Another feature is an showcasing of W3C’s technology, such Accessibility Guidelines[6]. Some alternate view that lets the user see how as MathML. Moreover, it has allowed us features, such as the requirement for an a document would be displayed with a to give useful feedback to the W3C’s ALT attribute for both images and text-only browser. working groups concerning the feasibility of the implementations. Our plans are to integrate more XML features into Amaya and to internationalize the code. We are also starting a new experiment using RDF-based annotations in a Web collaborative environment.

Amaya is an Open Source software, which runs on both Windows and Unix platforms. Source code and pre-compiled binaries can be downloaded from the W3C site.

Links:

http://www.w3.org/Amaya http://www.w3.org/pub/WWW/Jigsaw/ [2] http://www.w3.org/TR/html401/ [3] http://www.w3.org/TR/REC-MathML/ [4] http://www.w3.org/Graphics/ [5] http://www.w3.org/TR/REC-CSS2/ [6] http://www.w3.org/WAI/

Please contact:

Irène Vatton – INRIA and W3C Tel: +33 4 7661 5361 E-mail: [email protected] Nested XML structures in Amaya.

XML Accessibility Guidelines

by Daniel Dardailler

Compared to the HTML or MathML language, XML is support. In this context, guidelines are needed that one level up: it is a meta syntax used to describe these explain XML formats and tools designers how to languages, as well as new ones and it provides no include basic accessibility features - such as the ones guarantee of device independence or textual alternate present in HTML - in all their new development.

XML (Extensible Markup Language) is a For instance, in HTML, authors can write In SGML and XML, authors can define meta-syntax, used to create new languages. documents like: their own set of elements, and end up It can be seen as a simplification of with documents like: SGML (Standard Generalized Markup XML and Accessibility Language), designed to promote a wider

Daniel New England Restaurant acceptance in Web markets, but serving Dardailler
Clam Chowder the same functionality of extensibility and

Background

A large creamy bowl new language design. HTML (Hypertext of clam showder, with Markup Language) is one particular and they can only use elements (TITLE, bread crumbs on top application of SGML, which covers one H1, etc) defined by the HTML
set of needs (‘simple’ hypertext documents) specification (which defines about a and one set of element and attributes. hundred), and their attributes.

20 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

which may fit more closely the needs of alternative user agents can do something • describe how discrete, sequential, their information system. intelligent with just the document bits. structured, and search navigation mechanisms should work. Within W3C, the HTML language is now This semantic knowledge can be provided migrating from SGML to XML – this is through human readable documentation An additional advice we give to DTD called XHTML – including a of course, but having machine readable designers is that in their specification itself modularization of HTML to suit the needs assertions of some semantics that can then (the documentation) they always of a larger community (mobile users, Web be used to present the document in various emphasize the accessibility features of TV, etc). XML is therefore not to be seen media is paramount for seamless access their new language and try to include as a replacement of HTML, but as a new (ie, you don’t need a programmer, you accessibility as part of any conformance building layer on top of which HTML is just need a program). statement that they introduce (be it for the to be placed, next to other languages document themselves, or for designed by W3C, such as MathML (for Guidelines for Designers of User readers/editors of the language). See the representing mathematical formula), Interface-oriented XML Tagset Scalable Vector Graphics (SVG) SMIL (for synchronizing multi media), This section provides a list of proposed specification for an example of both SVG (for scalable graphics), etc., and abstract guidelines. Some examples of practices (http://www.w3.org/Graphics/ other new languages designed by other checkpoints are provided, and detailed SVG/). organizations (such a OpenEBook, XML- checkpoints and techniques that DTD EDI, etc). designers can follow to achieve Conclusion accessibility when designing new XML Our primary message is simple: be device Problem statement DTDs still have to be defined by WAI independent, and export your semantics The Web Accessibility Initiative (WAI) and W3C. as much as you can. We believe abstract has done extensive work in the HTML guidelines and verifiable checkpoints/ area, resulting in lots of new 1. Ensure that authors can associate a text techniques (using implementation functionalities being added to the version description with any non-text content mechanisms associated with abstract 4.0 of the language. One area of concern (graphics, sound, multimedia, scripts, guidelines) are the best way to address with the advent of XML is that the etc.): this problem and we are in process of freedom of design it brings will result is • make sure this is done in the most defining them in the framework of the a loss of accessibility features, present natural way possible WAI at W3C. today because of HTML pervasive • point at XHTML modules for presence and widely available OBJECT, MAP element, SMIL switch, Links: specification. For instance, one can design etc. WAI: http://www.w3.org/WAI a set of XML tags that would prevent the WAI and XML: creation of accessible document, such as: 2. Create semantically-rich languages. Do http://www.w3.org/WAI/PF/xmlgl not define presentation elements or

New England Restaurant attributes: Please contact: Clam Chowder • Enable the use of style sheets (linked Daniel Dardailler – W3C or internal). Tel: +33 4 9238 7983 • use the standard linking mechanism E-mail: [email protected] • define element types that allow with no way to include an alternate textual classification and grouping description of the photo, for instance. • define element types that identify important text content. But let’s start by defining what we mean by accessible DTD and documents. An 3. Export semantics: XML DTD is accessible if it enables and • provide accessible documentation promotes the creation of accessible • provide schemas and a mean to access documents. A document is accessible if it it can be equally understood by its • reuse accessible DTDs, schemas, etc. targeted audience regardless of the device (, xml:lang). used to access it. For User-centric XML based languages, the message is simple: 4. design an accessible user interface: be device independent and export your provide default style sheets for multiple semantics as much as you can. While the output modes: priority is stronger on the first aspect • describe navigable structures (multi-modality), both aspects are • use CSS or XSLT to describe a basic important, as without the knowledge of outline view the meaning of the XML elements and • document navigable structures attributes, there is little chance that

ERCIM News No. 41, April 2000 21 SPECIAL THEME: WEB TECHNOLOGIES

The W3C in seven points

The W3C was created to lead the Web to its full potential by developing common protocols that promote its evolution and ensure its interoperability. It is an inter-national industry consortium jointly run by the MIT Laboratory for Computer Science (MIT LCS) in the USA, INRIA and Keio University in Japan. To date, over 390 organizations are members of the Consortium. For more information, see http://www.w3c.org/

1. Universal Access interchangeable. They also expect to be able to W3C defines the Web as the universe of network- view Web content with their preferred software accessible information (available through your (graphical desktop browser, speech synthesizer, computer, phone, television, or networked braille display, car phone...). W3C, a vendor- refrigerator...). Today this universe benefits society neutral organization, promotes interoperability by by enabling new forms of human communication designing and promoting open (non-proprietary) and opportunities to share knowledge. One of W3C’s computer languages and protocols that avoid the primary goals is to make these benefits available to market fragmentation of the past. This is achieved all people, whatever their hardware, software, through industry consensus and encouraging an network infrastructure, native language, culture, open forum for discussion. geographical location, or physical or mental ability. W3C’s Internationalization Activity, Mobile 5. Evolvability Activity, TVWeb Activity, Voice Browser Activity, W3C aims for technical excellence but is well and Web Accessibility Initiative all illustrate our aware that what we know and need today may be commitment to universal access. insufficient to solve tomorrow’s problems. We therefore strive to build a Web that can easily 2. Semantic Web evolve into an even better Web, without disrupting People currently share their knowledge on the Web what already works. The principles of simplicity, in language intended for other people. On the modularity, compatibility, and extensibility guide Semantic Web ("semantic" means "having to do all of our designs. with meaning"), we will be able to express ourselves in terms that our computers can interpret and 6. Decentralization exchange. By doing so, we will enable them to solve Decentralization is a principle of modern problems that we find tedious, to help us find quickly distributed systems, including societies. In a what we’re looking for: medical information, a centralized system, every message or action has movie review, a book purchase order, etc. The W3C to pass through a central authority, causing languages RDF, XML, and digital signatures are the bottlenecks when the traffic increases. In design, building blocks of the Semantic Web. we therefore limit the number of central Web facilities to reduce the vulnerability of the Web as 3. Trust a whole. Tolerance of errors is the necessary The Web is a collaborative medium, not read-only companion of distributed systems, and the life and like a magazine. In fact, the first Web browser was breath of the Internet, not just the Web. also an editor, though most people today think of browsing as primarily viewing, not interacting. To 7. Cooler Multimedia! promote a more collaborative environment, we must Who wouldn’t like more interactivity and richer build a ‘Web of Trust’ that offers confidentiality, media on the Web, including resizable images, instills confidence, and makes it possible for people quality sound, video, 3D effects, and animation? to take responsibility for (or be accountable for) W3C’s consensus process does not limit content what they publish on the Web. These goals drive provider creativity or mean boring browsing. much of W3C’s work around digital signatures, Through its membership, W3C listens to end-users annotation mechanisms, group authoring, versioning, and works toward providing a solid framework for etc. the development of the Cooler Web through languages such as the Scalable Vector Graphics 4. Interoperability (SVG) language and the Synchronized Multimedia Twenty years ago, people bought software that only Integration Language (SMIL). worked with other software from the same vendor. Today, people have more freedom to choose, and they rightly expect software components to be

22 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

Web Images and Blind Users by Zdenûk Mikovec, Martin Klíma, Du‰an Pavlica and Pavel Slavík

The Blind Information System (BIS) is a system for description methodology based on XML which interpretation of graphical information for blind users characterizes pictures and charts by structure. It also designed and developed at the Czech Technical looks at objects and the relations among themansd University. The project is developing a picture then makes this information accessible to blind users.

User-friendly communication with Explorer, etc. Using a browser module implemented and tested (for testing computers is a very important issue. that allows the user to browse the textual purposes the system is implemented as a Development of user interfaces is at the description of a picture it is possible to stand-alone Java application). The centre of attention of the software industry obtain information about picture both the response from our blind volunteers has in the few last years and due to structural one and the list of objects in the been very positive. The approach used technological development new methods picture. allowed them to get relatively easy of communication are brought into orientation in pictures. It is important to everyday use. The past 15 years were Besides the ‘pure’ pictorial information note that the use of the system is limited characterised by intensive development the picture also contains information of to pictures that can be characterised by of graphical user interfaces that increased a semantic nature. This type of structure, objects and relations among the user-friendliness and thus the information covers relations like ‘a person them. These can only be described by the efficiency of communication with talks to another person’, ‘a girl waves to formalism implemented. computers. This allowed a wider audience a boy’, etc. Our system allows the creation to use computers. Nevertheless there is of both types of descriptions and to Future work will be devoted to the one user group that has been seriously browse them. In both cases a hierarchical development of the plug-in version for affected by the introduction of graphical description of a picture is created. The Internet browsers (eg Netscape) and the user interfaces – blind users.

The amount of graphically presented information is steadily increasing and very often the textual information can not be understood without perception of the attached image. It is obvious that blind users have to get access to graphical information to be able to understand the information presented. The solution is to create an alternative textual description the blind users can work with. The easiest way is to create the list of objects that the Picture perception process. picture consists of which the user can search or perform other operations on.. appropriate formalism by which it is extension of the approach described into This approach does not provide the user possible to create such a description is a 3D. This means that scenes described by with any structural information about grammar that can describe both structures means of VRML could be both displayed relations among objects in the picture (describing structural and semantic view in a normal way and investigated by investigated. of the picture). The description grammar means of textual queries concerning the was implemented in XML. The use of scenes internal structure. In such a way The aim of our BIS (Blind Information XML has several advantages. Because the blind users will get access to System) was to create a web tool that XML is a standard the system is easily information on the web currently not would allow the user to browse graphical portable into different environments. available to them. information. The basic idea is to create Another advantage is that XML can be an alternative picture description in processed in web environments (XML Links: textual form that would be appended to was specially designed for use on Internet, http://sgi.felk.cvut.cz/cgi-bin/toASCII/ optional section of some image formats and all newly developed web applications Research/bis/ like GIFor PNG. are able to work with XML documents). In the figure it is possible to see our Please contact: This approach would then allow the approach to the picture perception Pavel SlavÌk – Czech Technical University picture to be displayed by means of process. The system that allows the Tel: +420 2 2435 7480 standard tools that are an integral part of creation of picture descriptions and E-mail: [email protected] common browsers like Netscape, Internet picture browsing has been successfully

ERCIM News No. 41, April 2000 23 SPECIAL THEME: WEB TECHNOLOGIES

Adaptive Browsing: Supporting All Users on the Web

by Constantine Stephanidis

The AVANTI Web browser aims to facilitate access to (including people with special needs), in different Web-based information sources, applications and contexts of use. services, by potentially all categories of end users

The design space for the AVANTI Web the practical application of the concepts in the AVANTI browser is based on user browser was large and necessitated a and principles of ‘User Interfaces for All’ (dis)abilities, skills and knowledge, move away from traditional user interface during the user interface development life requirements and preferences, as well as development approaches, which cannot cycle. Unified User Interfaces address the the characteristics of the context of use. cater for the foreseen levels of diversity individual requirements of potentially all Adaptations in the browser take place and variability in the target user users in different contexts of use, by during the initiation of interaction population and contexts of use. The novel employing self-adaptation, to cater for (adaptability), as well as at run-time Unified User Interface Development their diverse, and often dynamically (adaptivity). method was adopted in order to support changing, characteristics. Self-adaptation Two instances of the interface that demonstrate adaptation based on the characteristics of the user and the usage context are presented in Figure 1(a) and Figure 1(b). Specifically, Figure 1(a), presents a typical instance intended for use by a user familiar with Web browsing. Note the additional functionality that is available to the user (eg, a pane where the user can access an overview of the document itself, or of the links contained therein, an edit field for entering the URLs of local or remote HTML documents), as well as the facilitation of the link selection (links are presented as buttons, arguably increasing their affordance).

The second instance, Figure 1(b), presents the interface in ‘kiosk mode’, ie, making use of the kiosk-based information system metaphor. Note that the typical windowing paradigm is abandoned and replaced by a simpler, content oriented interface (for example, scrollbars have been replaced by ‘scroll buttons’, which perform the same function as scrollbars, but have different presentation and behavioral attributes – eg, they ‘hide’ themselves if they are not selectable). Furthermore, the interface ‘toolbar’ contains ‘content-sensitive’ buttons, ie, buttons which differ based on the actual information content of the pages being viewed.

Figure 1(c), presents disability-oriented adaptations in the browser’s interface. The interface is presented with automatic scanning activated. The additional toolbar that was automatically added in the user Figure 1: Instances of the browser interface. interface is a ‘window manipulation’

24 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

toolbar, containing three sets of controls of interaction by providing real-time guidelines and recommendations for User enabling the user to perform typical support and guidance to the user as he/she Agents and Web Content Accessibility. actions on the browser’s window (eg, interacts with the system. The system The AVANTI consortium, partly funded resizing and moving). The three sets of monitors the interaction session, identifies by the ACTS Programme of the European controls, as well as the ‘rotation’ sequence possible interaction gaps and triggers the Commission, comprised: ALCATEL between the sets are also depicted on activation of appropriate dialog patterns Siette (Italy) - Prime contractor; CNR- Figure 1(c). These sets occupy the same or interface adaptations. The different IROE (Italy); ICS-FORTH (Greece); space on the toolbar, to better utilize categories of run-time adaptations include GMD (Germany); University of Sienna screen real estate, and to speed up the provision of awareness prompting, (Italy); MA Systems (UK); MATHEMA interaction; the user can switch between adaptive guidance and help facilities, and (Italy); VTT (Finland); ECG (Italy); them by selecting the first of the controls. the activation of alternative interaction University of Linz (Austria); TELECOM In the same figure one can also see the patterns for each browser task. ITALIA (Italy); EUROGICIEL (France). on-screen, ‘virtual’ keyboard activated for text input (interaction with the The AVANTI Web browser prototype, Please contact: keyboard is also scanning-based), as well developed by ICS-FORTH, is currently Constantine Stephanidis - ICS-FORTH as the scanning highlighter over the the only system known to provide access Tel: +30 81 391741 toolbar on the screen keyboard that to the Web for a range of user categories E-mail: [email protected] provides feedback on the current and different contexts of use. Ongoing selectable item. work involves the support of different platforms and different terminal devices, Run-time adaptations in the AVANTI as well as the compliance to new Web browser aim to enhance the quality

Vector Graphics by David Duce

This article describes two recent developments in 2D University exploring the translation of higher level web graphics, WebCGM and Scalable Vector Graphics. languages into Scalable Vector Graphics is also Work in progress at CLRC and Oxford Brookes described.

Ever since the introduction of images via other than to note that VRML is being WebCGM in the Mosaic browser, the only way further developed within the CGM, the Computer Graphics Metafile, of including simple vector graphics (flow consortium, see http://www.web3d.org/ is an ISO/IEC standard for the storage diagrams, floor plans, simple layouts etc) for details. and transfer of picture description in a web document has been to compose information. The first edition of CGM the vector graphics using some other Vector graphics offers a number of appeared in 1987 and a number of system and translate the result into a pixel- advantages over image-based graphics: extensions have been standardized since based image. This leads to an inefficient vector graphics are fully scalable without then. CGM is widely used to describe representation, inability to search for the loss of resolution, file sizes are technical illustrations and has a strong content in the diagram (particularly the independent of resolution (leaving aside following in fields of engineering text), and an inability to make changes. considerations of level of detail), a rich including the aeronautical and automobile Furthermore, the resulting image is set of primitives is typically provided, text industry and the defence industry. CGM completely inaccessible to people with is represented as text strings, hotspots can provides a rich set of primitives and disabilities or users working in be defined by drawing elements. There attributes for describing pictures. Abstract environments where visual representations are both proprietary and open systems and concrete syntax are separated in a nice are inappropriate, unless alternative around, for example the proprietary way, the functionality of the drawing descriptions are included. Two recent systems PDF (Adobe), WMF (Microsoft) elements are defined in Part 1 of the developments are changing this picture: and DXF (Autocad). Open systems standard and three concrete WebCGM and Scalable Vector Graphics. include the ISO/IEC standard CGM representations (binary, clear text and Both address 2D graphics. Some of the (Computer Graphics Metafile) and the character encodings) are defined in Parts requirements for the inclusion of 3D recommendation under development in 2 to 4. Another important concept in graphics in web pages have been the World Wide Web Consortium SVG CGM is the idea of a profile. Profiles addressed by the VRML (Virtual Reality (Scalable Vector Graphics). CGM and enable the standard to be tailored to the Modelling Language) standard; 3D SVG are briefly described below. needs of specific application areas in well- graphics is outside the scope of this article, controlled ways, for example by

ERCIM News No. 41, April 2000 25 SPECIAL THEME: WEB TECHNOLOGIES

Web Schematics markup of a diagram and the corresponding visible representation generated by translation to Scalable Vector Graphics.

restricting the range of primitives and submitted a proposal called PGML and a Unix pic language for picture description attribute values that need to be supported few weeks later a consortium led by has been designed and its translation into in order to meet the specific needs of a Microsoft submitted a proposal entitled SVG using the DOM and a scripting particular area. VML (Vector Markup Language). W3C language has been demonstrated. The responded by establishing a working language provides a more declarative way WebCGM is a profile of CGM designed group on scalable vector graphics and to describe diagrams than SVG. to meet the needs for scalable 2D vector after an initial requirements phase, the graphics in electronic documents on the first working draft of SVG appeared in Links: World Wide Web. CGM was developed February 1999. WebCGM: by the CGM Open Consortium in http://www.w3.org/Graphics/WebCGM/ collaboration with W3C staff and with Since that time SVG has matured SVG: the support of the W3C-LA leveraging considerably and prototype http://www.w3.org/Graphics/SVG/ action funded by the EU. Staff at a implementations have started to appear. SVG plug-in viewer: http://www.adobe.com/SVG number of ERCIM institutes have been Unlike WebCGM, SVG is represented in engaged in this activity. The publication XML and fully utilizes the style sheet and Please contact: of WebCGM as a W3C Recommendation DOM mechanisms provided by other represents a significant interoperability W3C Recommendations. In terms of David Duce – Oxford Brookes University and CLRC agreement between major players in this application areas, it is fair to say that Tel: +44 1865 48 4379 field. As well as providing a rich set of WebCGM is slanted more towards the E-mail: [email protected] primitives, WebCGM also includes CAD and engineering areas, whereas elements to group graphical objects for SVG is slanted more towards business linking, layering and searching. graphics and graphics arts. SVG is based on the painter’s drawing model. Shapes There are two points to note about defined by paths may be filled and stroked WebCGM are that there isn’t an XML in a variety of styles. Fill styles include representation of WebCGM and also it rich gradient and radial fills and filter does not use the web style sheet effects such a Gaussian blur, specular and mechanisms. For some applications, these diffuse lighting may be applied to shapes are significant limitations. or groups of shapes. Animation facilities are also under development in Scalable Vector Graphics (SVG) conjunction with other W3C Working SVG is a Recommendation under Groups. development in the World Wide Web Consortium. The SVG activity started Web Schematics around 1998 with the submission of a This is work in progress at RAL and now proposal called Web Schematics by Bob Oxford Brookes University to explore Hopgood and David Duce at RAL and how higher level (in a sense modeling) Vincent Quint at INRIA. Shortly languages, can be translated into SVG. afterwards a consortium led by Adobe So far an XML language based on the

26 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

More Quality for Services in Internet by Juha Koivisto

Internet traffic consists of packets. How well the Technology is building test networks to evaluate QoS packets flow can be characterized by Quality of technologies in various configurations and Service (QoS) parameters. Several approaches to combinations. Particularly interesting are provide better service for selected applications have configurations where QoS support is available only been defined. A research team at VTT Information in very limited parts of the network.

Internet traffic consists of packets. These is not always sufficient. Users want better applications can get guaranteed QoS, Internet Protocol (IP) packets carry all the quality for important applications. They which is independent of the load other data of, for instance, Web, e-mail and IP are even ready to pay higher prices to users put on the network. However, phone applications. How well the packets guarantee the smooth operation of critical establishing and maintaining such flow can be characterized by Quality of applications. Therefore, several reservations is complex, and classifying Service (QoS) parameters. QoS approaches to provide better service for the packets is heavy for network parameters include bandwidth selected applications have been defined: components. This is why resource (bits/second), average delay (transmission • Network managers may configure reservations are usually not supported. time from sender to receiver), delay routers to favour certain applications, Before resources can be reserved, it variation, packet loss rate, and error workstations or sub-networks. The must be decided who is allowed to probability. Which parameters are the routers set priorities based on the make the reservations, and who is going most important depends on the addresses and port numbers of the to pay for them, etc. applications. WWW users want as broad senders and receivers. Such priorities • Smart applications that adapt to bandwidth and low packet loss probability tend to be fairly static. available QoS can improve perceived as possible in order to shorten download • Applications may categorize IP packets service quality. For instance, an times. Users of interactive voice into different classes. Then the network adaptive video encoder may decrease applications cannot take advantage of the knows which packets are, for example, resolution and send fewer bits per bandwidth exceeding a certain limit, but less important and drops them in second when the network is heavily they find short delays important. In overload cases. This kind of loaded. Adaptive applications need to general, real-time applications are more differentiation of services does not get feedback about the actual QoS. sensitive to QoS than other applications. require end-to-end signalling. However, • The network may be enhanced to there is no general agreement about the provide more bandwidth, or separate When no special actions are taken, IP exact meaning of the traffic classes. networks may be used for critical packets are transmitted as fast as the • Resources may be reserved in advance applications. This may be the simplest bandwidth and load on the network through, for instance, the Resource but not necessarily the cheapest permit. However, this best-effort service Reservation Protocol. This way, some approach.

Internet traffic between two users in different organizations goes through several administrative domains. A combination of various approaches can be used to improve QoS. For example, resource reservation in an access network can be combined with priorities in the core network. It is not necessary to control QoS on all the nodes on the path from sender to receiver, but it is important to control the bottleneck of the connection. We are building test networks to evaluate QoS technologies in various configurations and combinations. The goal is to find out when it is worthwhile to use QoS mechanisms.

One problem with many QoS mechanisms is that they require extra network infrastructure, modifications to Important applications can maintain high quality by having high priority protocols, or co-operation between the in the network bottleneck. sending and receiving applications. Often

ERCIM News No. 41, April 2000 27 SPECIAL THEME: WEB TECHNOLOGIES

such an infrastructure is not available, or applications and it gives a larger share of cation software development methods. end-users are not allowed to dynamically the limited bandwidth to important Moses is a co-operation project between configure it. However, there are also applications. The solution is useful when VTT, Lappeenranta University of simple mechanisms that set very little no other differentiation mechanism is Technology, HPY Research, Nokia requirements on the environment. One available and the network bottleneck is Research Center, and Tekes (the National example is the process-priority-based not loaded by other users. This situation Technology Agency). method that was developed in the Moses is quite common with slow wireless and project. The method slows down selected modem access links. For example, a user Please contact: applications by reducing the processor may want to give a larger share of the Juha Koivisto – VTT Information time they get. Through the properties of access network capacity to an interactive Technology the Transmission Control Protocol (TCP), voice application, and less capacity to Tel: +358 9 456 5664 this will eventually slow down the WWW downloading. E-mail: [email protected] sending rate on the other end and leave more bandwidth for critical applications. This work is part of the Moses project. This method does not require any extra Other research topics in Moses include support from the network or the sending mobility support technologies on network end or any modifications to existing and application layers, and communi-

The Social Web Cockpit: A Tool to Support Knowledge Sharing Communities

by Wolfgang Gräther and Wolfgang Prinz

The Social Web Cockpit integrates different services community. This includes a web-search based on that have been developed in the Social Web research recommendations, a concept index to develop a programme at GMD Institute for Applied Information community vocabulary, shared workspaces for Technology to support a knowledge sharing knowledge sharing, and an awareness service.

Computer supported co-operative Work • LiveMarks, a system of information knowledge by recommending pages and (CSCW) research yielded a number of processing software agents running on related links, by annotating and discussing groupware systems for the support of our agent platform SoaP. them. People may add their own cooperative work such a shared documents in a shared workspace, they document-spaces or workspaces, shared We have developed a service, called the may highlight key phrases and develop applications, and workflow systems. ‘Social Web Cockpit’, or ‘cockpit’ for their special terminology. These concepts These applications support primarily short, that integrates these service and that and terms are automatically cross- work or knowledge sharing processes that turns websites into meeting places where referenced with the site and related pages are pre-planned or that involve well- visitors may become aware of each other, to support content-based navigation. defined, closed, or small groups of people. open communication channels, and Search facilities take into account concepts This paper describes systems that aim at exchange information and knowledge with and collaborative ratings. Visitors may the support of knowledge sharing within each other or with experts. Usually people select pages from the site to be monitored larger and more loosely coupled groups visit a website in search of some for changes that will be indicated upon of people that share a common interest, information. So there is some chance that their next visit. ie knowledge sharing communities. The its visitors may have overlapping interests, presented systems originate from our if the site matches their expectations, We expect that web-sites that are ‘Social Web Research Programme’ at backgrounds, or motivations. These augmented with the cockpit and the Social GMD-FIT which aims to explore and persons might well profit from each Web Services may become very attractive demonstrate how we can turn information other’s knowledge by sharing opinions or for knowledge sharing communities environments into rich communication experiences or offering advice. Some because they offer the chance to meet and interaction environments. These are: visitors might enter long term relationships again, to see what is new, and to inspect • BSCW, a Web-based shared workspace and eventually evolve a community - if the growing number of contributions by system only they became aware of each other. its visitors. Eventually, the visitors may • Nessie, an application-independent Similar functions are provided by systems consider themselves as a virtual community awareness infrastructure such as CoBrow or ThirdVoice. The that is held together by their common • Concept Index, a system for concept- Social Web Cockpit goes beyond that. interests and this special place for meeting based indexing of Web documents Visitors may enrich the website with their and trading information and knowledge.

28 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

Many websites would benefit from added returned to the cockpit and displayed a query to a search engine, software community support: commercial sites that using the corresponding icon. Ratings are agents in the background look for similar want to attract visitors, sites of used in LiveMarks to help users to queries of the community with related and professional organizations, self-help retrieve useful information from the Web. recommended web pages. The result of groups, clearing houses, digital libraries, the search is a mixture of web pages project portals. Nowadays, however, The cockpit uses the NESSIE awareness originating from the recommender community support is mostly restricted infrastructure to turn websites into database and pages coming from the to chat, annotation and rating facilities, meeting places. Members of a community search engine. When the member discussion forums, or forms to submit become aware of each other and get in recommends such a web page (by rating new links. The combination of a powerful touch. While visiting web pages, users it good or excellent), the page is put into shared workspace system, a collaborative get informed about the community to the link collection of this community. recommendation service, a community which the web page belongs. The median vocabulary for knowledge organization, of all ratings of a web page – made by Summary and a notification service, as offered in members of the community – is also The different services developed by the our Social Web Cockpit, is unique. shown. Users can easily contact experts Social Web Research Programme are of this community. Awareness information combined to provide a comprehensive set Cockpit Services is visualized by different indicators; f o r of services for a knowledge sharing The cockpit is a small window that example, changes of icons, detailed community. This includes a web-search occupies only little space on screen; all listings of present members in a specific based on recommendations, a concept actions and controls can be dragged out community, etc. Members of a community index to develop a community vocabulary, in separate tiny windows. Therefore, most may easily start an agent which monitors shared workspaces for information users don’t need to reorganize their desktop when using the cockpit (see figure 1). After the cockpit has been launched and when a web-page is visited, a user becomes immediately aware of other visitors and the communities at the site. A cockpit is a place for social Cockpit User Interface. encounters so users can get in contact with each other using different communication tools. A chronology of community visits web pages. The monitoring agent adjusts exchange, and an awareness service. We supplies further details about the to the users’ preferences: the user may believe that the extension of a web-site distribution of visits over the last hours, specifiy what web page should be with these services will foster the days, or month. This helps users to decide monitored, for what keywords and at what community building process because it on the best time for visiting a community. dates and in what frequency this should enables a web-site to become a meeting be done. Users are notified of changes in place for a knowledge sharing community. The Social Web Cockpit uses BSCW to a web page by a little yellow asterisk in enable cooperation over the Web. BSCW the monitor icon. Acknowledgement (Basic Support for Cooperative Work) is We would like to thank the teams of a shared workspace system which The Social Web Cockpit uses the BSCW, Concept Index, LiveMarks and supports document upload, event ConceptIndex system to enable members Nessie of the FIT.CSCW research group notification, group management, of a community to construct their own who all cooperated to make the Social discussion forums, meetings, and much shared vocabulary. This vocabulary Web Cockpit operational and A. Voss more. In the context of the Social Web captures characteristic words and phrases; who contributed to an earlier version of Cockpit a community is a group of people an index based on this vocabulary cross- this article. sharing similar interests and tasks. In our references all web pages of this approach, the documents which the community. Web pages may be viewed Links: community collects and produces are the with the community vocabulary; phrases Social Web Cockpit: main carrier of knowledge. Community in the web page are highlighted and links http://orgwis.gmd.de/cockpit/ knowledge is stored in BSCW shared to the index are provided. A word or Social Web Research Programme: workspaces. Creation of communities and phrase can easily be added to the http://orgwis.gmd.de/projects/SocialWeb/ community membership is implemented vocabulary. The index with all the cross- Please contact: by BSCW functions. Users can rate the references is automatically updated. current page on a scale of 5 (poor, Wolfgang Prinz – GMD passable, fair, good, excellent). This The cockpit uses a recommender system, Tel: +49 2241 14 2730 personal rating is sent to the BSCW which enables members of a community E-mail: [email protected] system, where it is added to other ratings to share queries and to collaboratively of the page within the present community. retrieve and collect useful web pages. The (changed) median of all ratings is While a member of a community submits

ERCIM News No. 41, April 2000 29 SPECIAL THEME: WEB TECHNOLOGIES

Eliminating the I/O bottleneck from World Wide Web Proxies

by Evangelos P. Markatos

World Wide Web proxies are being widely used today, matters worse, this load is expected to increase in order to reduce network traffic, web server load, sharply in the near future when users will be able to telecommunication costs, and to provide Internet connect to the Internet via fast modems. A research access to users behind a firewall. Web proxies are team at ICS-FORTH investigates methods to enhance typically required to handle a heavy load which the performance of Web proxies thus enhancing web reaches up to millions of requests per day. To make server performances.

Although the network bandwidth has been they will usually only be replaced by reference by serving several clients at traditionally thought to be the main other URLs when the proxy runs out of a time, and effectively storing bottleneck in a proxy’s performance this space. documents requested by different is not true anymore. Recent research clients in nearby disk locations. The results suggest that a proxy server may Thus, traditional file system technology locality of reference that apparently spend a significant percentage of its time can not keep up with the disk I/O needs exists in web client requests would be on file-system related activities. of busy web proxies. Our current work preserved and exploited to achieve a focuses on identifying and eliminating better disk data layout. Experimental evaluation of the web proxy the sources that create the disk I/O at Digital Palo Alto firewall revealed that bottleneck. We investigate the effects of We have started testing our approach the disk I/O overhead of caching turns out storage management policies for web using a combination of experimental to be much higher than the latency proxies based on the following principles: evaluation and simulation. Our initial improvement from cache hits. Thus, to • Avoid file management operations: results are very encouraging: they suggest save the disk I/O overhead the Digital current proxies store a single URL in that our method breaks through the I/O Palo Alto proxy server is typically run in each file. Given that each file bottleneck and effectively improves its non-caching mode. This disk I/O creation/deletion operation requires 1- performance over traditional proxies (like overhead can be partially attributed to the 2 disk operations, and that most disks SQUID) by more than an order of fact that file systems were designed and can usually perform no more than 50 magnitude. optimized for an every-day pattern of use operations per second, then an average that is completely different from the needs proxy can not deliver more than 20-30 Links: of a web proxy: URLs per second, or roughly 100 http://archvlsi.ics.forth.gr/OS/www.html • Contrary to read-dominated traditional KBytes/sec, a rate of execution that is file systems, disk accesses in web orders of magnitude lower than the data Please contact: proxies are dominated by write transfer rates that current disks can requests: Each URL request that does sustain. This data rate is even lower Evangelos P. Markatos – ICS-FORTH Tel: +30 81 391655 not hit in the proxy’s cache must be than most Internet connections. We E-mail: [email protected] fetched from the network and stored advocate that proxies should not store into the proxy’s local disk. Thus, each one URL per file, but use a small URL miss usually results in a disk write number of large files to store all URLs. operation. Disk read operations are This approach reduces the number of invoked when a URL request misses operations associated with creating, the proxy’s main memory cache, but opening, closing, and deleting files. hits the proxy’s disk cache. Our • Avoid disk operations: proxies should measurements suggest that for a web write data in large chunks and reduce proxy file-system write operations are the aggressive prefetching being usually 3-5 times more than file-system employed by modern operating read operations. systems: prefetching is likely to hurt • Although file-update operations are performance since proxy data are not frequent in traditional file systems, Web accessed sequentially. proxies rarely update their files: Most • Preserve locality of reference: each of the data on the web change rather stream of URL requests that arrives in infrequently. Published results suggest a web proxy from a single client has a that the average life of a web document significant amount of temporal locality can easily be several days up to several (eg every time a client requests an weeks (or even months) long. This fact HTML document (s)he will request the implies that most URLs stored on a document’s embedded images as well). proxy’s disk may never be updated; Current proxies destroy this locality of

30 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

WEBPATH — a 3-D Browsing History Visualisation by Emmanuel Frécon

WE BPA T H is a virtual reality based application that space are used to position representations of provides an easily recognisable presentation of your documents using different metrics, together with their browsing history. The different dimensions of a 3-D relationships.

Only a few years ago, Web browsing was seen as an interesting possibility. It is now part of our working and leisure activities, and the time we spend looking for information ‘out there’ increases. As it becomes easier to publish documents, as the number of users, and thus publishers, increases and as the number of documents grows, searching for information is turning into a cumbersome and time- consuming operation. Moreover, because of the loose inter-connection among documents, people have difficulty returning to previously visited pages and remembering where they have been.

SICS and Lancaster University have prototyped W E BPA T H, a virtual reality (VR) based application to be used alongside traditional Web browsers. A driving motivation for this work is specifically to assist users in scenarios where they are searching for pages previously visited. These scenarios A 3D network of documents that have been visited during a number of include requests such as: “I want to find browsing sessions.The vertical axis portrays the time. a page I visited this morning, all I know is that it had no images in it”; “I remember etc.), WE BPA T H will create a (y) is always reserved to portray the time the page was from Asia somewhere”; “It representation for that document and the at which the documents were visited, the had many links and the latest news in it”; user will be smoothly transported in front last visited document is always the top- and “It took a very long time to load”. of it. Each Web document is represented most one. If a document happens to be as a cube that is labelled with the revisited, it will generate a new cube, but WEB PAT H provides a 3D structure to the document’s title and appears as any one all representations for the same document network of documents that have been of: the page’s background image, an are vertically linked to ease recognition. visited during a number of browsing image in-lined in the HTML document sessions. A three-dimensional or simply the background colour of that Clicking on a cube within the visualisation presentation was chosen for maximum page. In order to indicate the structure of will cause the user’s browser to re-load flexibility in data visualisation. The the user’s browse path, WE BPA T H the associated Web page. When the user different dimensions of a 3-D space are connects the cubes with arrows. The selects a new document location using the used to position representations of colour of the arrow indicates the nature different direct methods (ie bookmarks, documents using different metrics, of the link, for example, does the history facilities or manual entry), he/she together with their relationships. The destination document originate from the starts a new browsing session and a result is an easily recognisable same site as the previous document? semitransparent horizontal plane, covering presentation of your browsing history. the whole visualisation area is generated WE BPA T H uses three orthogonal just below the new representation. When The presentation can be tailored in terms dimensions to visualise the browsing many planes have been created (ie when of its overall structure and the appearance history. A number of metrics are a user has been engaged in several Web of individual pages. When a user loads a available, such as page size or number of sessions) older documents, towards the new document (using any of the methods images contained in a document. Any of bottom of the visualisation space, will offered by existing browsers: bookmarks, these metrics can be applied to the two appear through an increasingly dense fog, history facilities, manual location entry, horizontal axes (x and z). The vertical axis thus, reducing the user’s awareness of their

ERCIM News No. 41, April 2000 31 SPECIAL THEME: WEB TECHNOLOGIES

previous results. When looking back for WE BPA T H provides users with a Links: previous documents, all the cubes graphical view of their recent activities, resulting from a browsing sub-session are enabling them to for example, see how a http://www.comp.lancs.ac.uk/computing/ users/gbs/webpath/ “isolated” in-between two semitransparent given document was reached and visually layers, which visually separates them from perform complex searches for previous Please contact: the rest of the data. documents visited. WEB PAT H has raised substantial interest, ranging from casual Emmanuel Frécon – SICS Tel: +46 8 633 15 34 WE BPA T H is a tool that unobtrusively users to scientific researchers. The authors E-mail: [email protected] visualises the users trail of URLs as they are currently looking into commercial browse the Web, which can be tailored to ways to distribute an improved version best suit the users working needs. of the software.

The sView System

by Fredrik Espinoza and Markus Bylund

The sView system developed at SICS makes services to collaborate and further research will be networked services available to users in an done in the area of user assisted service collaboration environment that is close to the user. It allows using sView as a platform.

The role of the Internet is about to change user, enables user interaction with with their users (whichever channel is again. After the hype of web portals we networked services. The PSE can store and available for the moment), eg HTML over look forward to an Internet in which execute service logic and data, and it can HTTP, WML over WAP, ASCII over networked services dominate. However, move from computer to computer as the SMS, GUIs via SWING, etc. problems with the currently prevailing user moves between access nodes within infrastructure for networked services, the the network. Preferably, the PSE is run Implementation World Wide Web, slow down this process. on a computer under the user’s immediate The sView system is a complete system We argue that there are three main control (such as his/her workstation at for user-service interaction that implements problem areas: limited support for service work or personal computer at home), but PSEs. The system is authored exclusively collaboration, producer dominance, and if the user is not directly represented on in Java 1.2, which brings several the poor support for service interaction in the network, the PSE can run on a server advantages. The choice of implementation today’s web browsers. that is dedicated to running PSEs. language makes it easier to use the system on different platforms. The popularity of A Personal Service Environment (PSE) The choice of method for interaction Java makes it easy to reach a broad user enables a rich interaction between between service logic within a PSE and group, and finally, Jini technology can be networked services and their users. A its user is open. The PSE can provide a used for naming, searching and PSE, which is private to an individual range of channels for services to interact distribution of service components.

The sView system constitutes a very thin layer of infrastructure between the service providers and their users. Nothing is assumed about communication protocols between base services and service components within a PSE, or about interaction protocols between service components and the devices that render service presentations. Similarly, no languages, interpreters, or interaction styles are forced upon service components that wish to collaborate within the PSE. The thin properties of the sView system are a deliberate design choice. We want the system to be specific in its overall goal Figure 1: (to enable user-service interaction), but A schematic overview of the described usage open when it comes to methods of scenario. achieving the goal.

32 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

Sample Scenario Below we describe a usage scenario that illustrates some features of the PSE. A man is about to make a business trip to a foreign city. Using his Personal Service Environment, he locates an electronically mediated travel agency and initiates a dialog with it. The travel agency uploads a travel service component to the user’s PSE. Once in the PSE the travel service receives the man’s instructions, via a standard graphical user interface (GUI), to make a flight and hotel reservation for his planned trip. Then the man turns his attention to something else and leaves the office bringing his cellular phone. Figure 2: Screenshots from interactions with service components in the sView The travel service now makes use of a system. The calendar service and the travel service from the scenario above number of information sources in order interact with the user through a SWING GUI. The payment service interacts via to accomplish its task. It searches the PSE a cellular telephone. The snapshot of the cellular phone is taken from the for a preference manager and asks it about Nokia WAP Toolkit 1.2 that is used to simulate WAP enabled cellular phones its client’s complete name and address, in absence of real ones (see http://www.forum.nokia.com/developers/wap as well as his seating and smoking /wap.html). WAP phones are available for European 900MHz/1800MHz digital preferences. It also locates a calendar networks, but not yet for American 800MHz/1900MHz digital networks. within the PSE and checks when the man must be back and if the trip conflicts with The above scenario illustrates three • sView allows services to collaborate. any of his other appointments. important aspects of the PSE: In itself it only provides the locale for • the PSE allows, and actively supports, this collaboration, a locale that is very Having collected all background service collaboration to take place intimately tied to the user. However, it information, the travel service turns to its within the environment is our intent to do further research in base service trying to find an appropriate • since the infrastructure does not require the area of human assisted service flight and hotel. The service finds three HTML/HTTP, interaction through a collaboration using sView as a alternatives that all match the man’s number of different devices without platform. request, preferences, and schedule. The loss of interaction state is possible travel agency is now ready to get back to • the user shares personal information A Personal Service Environment (PSE) the client with the result of the search. (such as preferences) with a special enables a rich interaction between However, since the man is no longer purpose component within the PSE (the networked services and their users. We available via the desktop computer (as preference manager). This makes it have introduced the concept of a Personal can be concluded since the screen saver easy for the user to add, change, inspect, Service Environment as a solution to has been on for quite a while), the service and retract information without having many of the problems with using the contacts him via his cellular phone. The to contact every service that is used. At World Wide Web for mediation of man, now on the train on his way home, the same time, services have a central networked services. The main selects one of the alternatives and instructs source for such information for every contribution of the PSE is that it increases the travel agency to go ahead with the user. its user’s control of networked services. reservation. The travel service accepts the Services are offered support for request and starts searching the client’s Features spontaneous collaboration between peers. PSE again, this time for a service that In addition to what the above scenario The sView system will also serve as a provides payment. One of the man’s illustrates, the sView system demonstrates platform for further research in the area services, a bank service, is willing to the following noteworthy features: of human assisted service collaboration. provide payment, but only after a • service components can be downloaded confirmation by the client (this is also to the PSE for local execution. Links: done through the interface of the cellular • the use of three different interfaces for sView home page: http://www.sics.se/ telephone). Having everything that is interaction with the same service (HTML/ ~espinoza/sview/index.htm needed, including payment, the travel HTTP, WML/WAP, GUI/ SWING). service now executes the man’s request by • the interaction state can be preserved Please contact: instructing its base service to buy the flight between usage sessions and throughout Fredrik Espinoza and Markus Bylund tickets and make the hotel reservation. the transfer of interaction between Tel: +46 8 633 15 15, +46 18 471 70 53 devices E-mail: [email protected], [email protected]

ERCIM News No. 41, April 2000 33 SPECIAL THEME: WEB TECHNOLOGIES

Building a Multinational R&D Proposal through the Web

by Andrew Hallan, Alexander M. Samsonov and Wolfgang Ziegler

Last December, 20 scientific institutions and industrial European Commission’s formats and procedures. The companies led by GMD set up a Europe-wide various parts of the proposal were prepared at the consortium to write a proposal for the 2nd call of the widely dispersed partner sites and distributed, EU’s Information Society Technologies R&D discussed and combined in a shared workspace on programme. The preparation of a successful proposal the Web. This workspace was created and managed required close collaboration between consortium using BSCW, a tool developed by GMD. The members, together with strict adherence to the consortium never held a meeting.

Scientific research repeatedly exposes • no simple way to add comments or for creative work. Negotiation and new problems. Some of these, such as the ranking to documents. preparation of the proposal was done investigation of non-linear waves in almost entirely via the Internet, and fax solids, require exceptional computing We decided to set up a shared workspace and mail were negligible. Moreover, the resources, beyond the capacity of any one for the partners using the BSCW tool positive experience with long distance computing centre. The aim of our developed by GMD. BSCW allowed on-line document preparation and proposal was to make it easier for partners to use any Web browser to check collaboration encouraged us to reduce the scientists working on these problems to the status of all documents, to upload and budget for travel and other expenses in combine resources from a variety of high- download files, and to add their own the project itself. performance computing centres. comments. It also provided smooth integration with word processing, Once the proposal is accepted by the Our consortium was a large one, spread database and scheduling applications on European Commission, we will continue across Europe from Spain to Russia and the partners’ own workstations. The to use BSCW Web technology. This will from England to Armenia. By the time security features and versioning enabled allow a smooth and efficient transition that we had assembled the right partners, the editors to maintain a complete and through contract negotiations and into the we had only two months to write the consistent view of the whole proposal, management of the project itself. proposal. Writing an EU R&D proposal accessible to all the partners. With the Furthermore, part of our project will be is a complex process, where scientific underlying clarity and security provided devoted to the further development of objectives and approach must be by BSCW, ad-hoc e-mails and telephone Web technology to manage user accounts reconciled with the distribution of tasks calls were used just to resolve difficult and to ensure security in geographically between partners and an overall budget. points. dispersed high-performance computing. The geographical spread of our partners Results of the project will also be and the lack of time effectively prevented In less than two months, we succeeded in available in the European Grid Forum us from holding the meetings typical of building a complex, high-quality proposal EGRID. R&D consortia. of 100 pages plus an administrative and financial database, all through the virtual Links: The solution was to organise all the workspace provided by BSCW. Although BSCW website: http://bscw.gmd.de/ collaboration through the Internet. But almost all partners were using BSCW for from the beginning it was obvious that the first time, it was accepted and used Please contact: just setting up e-mail exploders and an from the beginning thanks to the intuitive ftp-site would be insufficient for GUI and the comprehensive on-line help. Wolfgang Ziegler – GMD Tel: +49 2241 14 2258 discussion and management of As a result, the consortium as a whole E-mail: [email protected] documents. There would be confusion never had a meeting. The four partners and wasted work because of missing managing the proposal held just four functionality, such as: working meetings face-to-face. At the end • no reasonable synchronisation and of the process all the documents required versioning by the Commission were available online • no document description available in the workspace, allowing electronic along with each document submission of the proposal to Brussels • no history function, showing who within a couple of minutes. changed what and when • no easy, platform independent facility The advanced Web technology of BSCW for navigating the document store allowed successful and cost-effective • no efficient and consistent way to collaboration on a pan-European scale. It manage the access rights for the reduced dramatically total travel expenses documents for partners and saved considerable time

34 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

The Collaborative Web by László Kovács and András Micsik

The omnipotent presence of World Wide Web the development of collaborative Internet services information sources and technologies made it feasible integrating ideas from Computer Supported to build groupware applications using Web Cooperative Work: rating within a digital library, a technologies. At the Department of Distributed collaborative information filtering project and an open Systems of SZTAKI, three projects are dealing with voting Web service.

AQUA-2, User Interface for Rating within Digital Library CommunitiesTechniques for querying digital library collections ranging from searching and browsing to relevance feedback, result ranking etc. are usually combined and users construct several queries until they get the desired results or the final query expression. A query interface that eases the building of the query expression is developed within the AQUA project (Advanced Query User Interface Architecture). The first version of AQUA was integrated with the Networked Computer Science Technical Reports Library (NCSTRL) using Dienst distributed digital library system.

The system builds upon the query chain paradigm: a sequence of alternate search and browse operations to filter the result sets of the preceding operations. Individual query or search operations are displayed via a GUI panel, the query A query in AQUA, and the panel for rating a document. chain is given by a series of panels. A great variety of query methods is offered. During an iterative query refinement, users build query chains that can be ratings to documents they have visited. users, especially those who have modified/refined at any point. Associated ratings can help other users to displayed similar tastes and interests in find the most relevant (high-)quality the past, are built within the SELECT The AQUA framework and environment documents. Ratings given by the users project supported by EC Telematics can ease future research as well as user are collected in a SELECT server Applications Programme. experiments towards the development of (SELECT is an EC funded project for advanced digital library query techniques social and collaborative filtering Automatic methods where the system and visualization methods. described below). The development of derives the filtering conditions from user AQUA-2 is partly funded by KNIXMAS actions or evaluations of documents using AQUA is further developed towards a INCO project supported by EC. NLP (natural language processing) collaborative tool for digital library techniques are also being developed. Two community use. AQUA-2 helps sharing SELECT, Social and Collaborative different user interface techniques were knowledge and opinions about documents Filtering within WWW developed: a bookmarklet-based thin while AQUA (the first version) helped and Usenet News SELECT client and a local proxy-based only sharing queries as well as exploration Internet users need tools to find the fat client. The thin SELECT client has the strategies among named or anonymous information most valuable to them within advantage of requiring virtually no community members. the limited time they may have available. installation and being accessible from any Information filters that make use of standard WWW browser. The proxy- AQUA-2 supports digital document recommendations derived from individual based approach allows the content of a rating. Peer rating groups (reviewers) can users’ past choices and/or recommen- document (WWW page or Usenet News be organized. They can attach their own dations derived from the behavior of other article) to be modified in different ways

ERCIM News No. 41, April 2000 35 SPECIAL THEME: WEB TECHNOLOGIES

presenting the actual rating information of the document as well as other kinds of quality indicators. Thus, the SELECT architecture may be used as a gateway for the associated quality metainformation. The proxy-based client may also perceive/interpret the actual user behavior and translate it into implicit rating information.

The server consists of several modules for passive filtering (recommendations are based on the submitted ratings of documents), for active filtering (recommendations are based upon the ratings of documents submitted by people with interests and/or rating histories that match those of the user), a rating database containing ratings of registered and/or anonymous users, as well as a profile database containing information about users and their interests.

The prototype implementation of SELECT has recently been completed, user groups evaluations as well as the A vote process for the standardization of Dublin Core element modifiers. development of a new open SELECT framework for public testing of new collaborative filtering ideas and service was developed which has been new graphical design of one of the WWW techniques are being initiated. operating as a public service of SZTAKI sites of the Hungarian government, for more than one year allowing any user surveys of a newspaper and a public WWW-based Open Voting or closed group to create vote processes library. It was used for decision polls and Survey Service of any type. within closed groups (selecting a logo for The importance of computerized votes a digital community, a survey about a and related activities (polls, surveys, etc.) Besides the basic functions of the service university lecture, deciding on the new have been increasing since the it has some features not found in other name for a company, the approval of draft globalization of the world’s information similar systems: recommendations by the Dublin Core societies. Computer supported decision • user access via WWW as well as E-mail international standardization board, etc.). making processes can have beneficial • many built-in question types and The system allows also to investigate the effects on society; these tools may be able evaluation algorithms beneficial effect of multi-round votes in to open and clarify decision making • support of multilingual votes reaching the best consensus (two phase processes, to help the evolution of • user identification, user and group poll for the new state slogan for Victoria people’s democratic attitudes. definition for different access state (Australia) number plates). permissions, A Reference Model for voting activities • flexible control scripts to control the Links: was defined as the first step. The voting vote process events, AQUA project: process is split into the following • replaceable graphical layout. http://www.sztaki.hu/sztaki/aszi/dsd/aqua subprocesses: initiation, collection, SELECT project: analysis, distribution, and decision A version of the voting system is also http://www.omega.it/research/select making. A high-level vote control used in a spin-off company of Kapsch SZTAKI vote service: http://www.sztaki.hu/services/voting language was created for the definition AG., which promotes the Web4Groups NCSTRL: http://www.ncstrl.org/ of vote subprocesses. WWW-based groupware application. ETRDL (ERCIM Technical Reference Digital Library) user interface: A prototype voting and survey subsystem The public voting service at SZTAKI has http://www.sztaki.hu:8000/ was developed at the Department of many applications in various types of Please contact: Distributed Systems of SZTAKI under voting activities. It served as a public the umbrella of Web4Groups of EC opinion collection tool in polls such as a László Kovács – SZTAKI Telematics Application Programme. As questionnaire on user requirements by Tel: +36 1 209 5286 a follow-up activity an Internet voting European R&D projects, a survey on the E-mail: [email protected]

36 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

Separation of Concerns in Multimedia Services by Olle Olsson

What can you do to make your customer feel good to his/her interests, and adapt the form of presentation when (s)he accesses some service you offer via the and interaction to the situation at hand. web? A good approach is to personalize the service

The ESPRIT project MAPPA packaging these services in a way that the desire to be able to support access (Multimedia Access through Persistent makes them accessible on a wide range from WAP-phones, as Java applets are Personal Agents) is targeted on supporting of access devices. The technologies range not (yet) supported on such phones. multimedia-based services adapted to the from public kiosks, desktop PCs and individual user. The two-year project laptops, to palmtops and mobile phones; A goal pursued in this project is to enable (running 1998 - 2000) is performed by a connected through broadband WANs, service tasks to be automatically consortium of six partners, from Sweden, called-up connections, or wireless configured and distributed to the user’s UK and Ireland. connections. access device. There are a number of constraining factors influencing this, eg The goal of the project is to develop The emergence of XML offers relief in what representation, presentation and technology and methods that enables certain aspects; the markup languages execution resources are available in the adaptable systems to be built. Such used for information delivery can be access device? What communication systems offer personalized service access, handled within a common framework. resources are available? What behavioral based on the paradigm of a automated But having a common basis for patterns are expected of the user? ‘How personal agent interacting with the user representation languages does not solve do the tasks map to response-time via multimedia assets. The personal agent the real problem; how to optimally priorities’? What are the sizes of the is able to build and evolve a model of user provide individualized services. different kinds of data-sets that support preferences based on the user’s the tasks and to what extent should data interactions. As the personal agent is One of the aims of the MAPPA project and task implementations be protected persistent, it is able to track changes in is to explore ways of dynamically from exposure in the access device? user preferences over time. adapting user interaction to the context in which interaction occurs. The approach We have developed a set of models The application goals are to provide is to provide a clear separation between corresponding to different perspectives services to registered customers to a the semantics of the services and the way on an interactive service. These include vendor, support their explorations of the services are presented when they are domain models, view models, interaction available products and provide communicated across the web. models, presentation models, and action announcements of new products that models. A representation of interactive match the interests of the customer. A second way of decomposing the access to a service is constructed by problem is the separation of Multimedia instantiating these models, adapting them The technical goals are to develop generic asset representation of information from to the constraints of the delivery models of multimedia product the tasks that the user performs on the environment, combining them to a presentations, of user preferences, and of basis of this information. The tasks complete models, and then producing a the tasks that the personal agent can correspond to what happens in the system representation that can be transported to perform. when the user interacts with the the client’s access device for execution. multimedia assets. Information Part of the agenda of the project is to representation and presentation is Service tasks are classified into a set of explore how customers can access the basically what XML and other markup categories, ranging from ‘interaction’ offered services via a range of devices languages support. Support for tasks are tasks (tasks that the user performs while and network technologies. typically provided through scripts exploring the information space and embedded in the document itself, or via preparing his fundamental service Individualized Services cgi-bin processing. request) to ‘service-effect’ tasks (tasks The broadened acceptance of the World that involve effects/actions that take place Wide Web as a medium for access to Individualised multimedia access depends in the service-provider’s environment). information and services has recently heavily on dynamic adaptation at the user As an example from the application been accompanied by a growth in the interface. Ideally, the full service domain we are working in, typical spectrum of commercially available implementation could be sent to the user’s interaction tasks are finding, inspecting, devices through which the web can be access device, eg in the form of Java and evaluating available products, and accessed. Service providers are becoming applets managing/controlling both the composing the content of a ‘shopping aware of the need to offer services that multimedia presentation and the actual cart’. All such interactions have no provide satisfaction to their users, and service tasks. But this is in conflict with semantical persistent side effects in the

ERCIM News No. 41, April 2000 37 SPECIAL THEME: WEB TECHNOLOGIES

vendors environment. The prototypical The structure of the information model actually will take advantage of the fact example of a service-effect task is maps to/from XML, which enables that they will be ‘continuously on-line’. submitting a request expressing that a domain data and multimedia assets to be Technologies and services should adapt specific set of products are to be shipped generated, operated on by to the user, not the other way around, and to the customer and payment drawn from generated/configured tasks, and presented the user will expect the services to be his account. Obviously, service-effect using standard browsers. The delivery more ‘intelligently interactive’. This will tasks are extremely domain dependent, environments experimented with include take us away from the old paradigm of and are handled as wrapped components standard web-browsers with HTML or the Web as a document presentation with a declarative description of their XML capabilities and JavaScript or Java technology to the new paradigm of the software interfaces. On the other hand, applet support, WAP-phones with WML network (whether wireless or not) being interaction tasks typically exhibit certain and WMLScript, as well as an a platform for intelligent interactive kinds of patterns, which makes it possible experimental Java-based presentation distributed services. This means that we to model these in a semantically more system using a flexible protocol for will need techniques and methods for meaningful way rather than programming interacting with server-based functional smart dynamic distribution of a set of them. components. An XSL-based adapted tasks that together provide a set transformation system has also been of services to the user. Interaction tasks are modeled evaluated (Cocoon), to determine to what declaratively, at a level of granularity that extent this type of standard support tools Links: enables the desired service behavior to be can be used for producing the actual Mappa website: achieved on the user’s access device. To representation. http://www.sics.se/mappa each task there is an associated description of how this task relates to and has effects The rapid evolution in commercially Please contact: on the information model underlying the available access technologies that we see Olle Olsson – SICS service. This description is expressed in today, introduces problems that will Tel: +46 8 633 15 19 a representation that can be transformed persist for a long time. New technologies E-mail: [email protected] into a form that can be executed or enable new patterns of behavior, which interpreted in the presentation will change the expectations of users, and environment on the device used. this will change the ways in which people

Increasing Security in Virtual Enterprise Communication for Mobile Environment

by István Mezgár, Tamás Szabó and Zsolt Kerecsen

A research project to develop a secure communication thus a large amount of valuable technical-related data architecture for mobile communication in virtual moves on the network. Since the management of virtual enterprises started at SZTAKI in 1999, supported by enterprises will be controlled also from mobile devices the Hungarian Scientific Research Fund (OTKA). The in the near future, security problems of mobile operation of virtual enterprises is based on the Internet, communication are of vital importance.

Mobile devices (mainly mobile phones) communication, operation - is strongly Today, e-commerce represents the best can be used today for more purposes than connected with the Internet. During way in efforts of convergence and just as making phone calls. Additional communication in a virtual enterprise, a globalisation. From an information point functions like mobile Internet browsing, huge amount of extremely valuable of view, the future of VEs moves in the ATMs for banking transactions, etc. will technical data and information same direction. E-commerce is the ‘fore- be available for daily use and virtual (development-, product-, process data runner’ of VE, there are many similarities enterprises are candidates for using these beside business information) moves in their functions and in the techniques types of new functions extensively. through the network, making security a they use. vital concern. Although there are already A virtual enterprise (VE) is a co-ordinated many applications available, the lack of Increasing the security level of virtual network of autonomous production units proper co-operative, effective and secure enterprise communication in a mobile (factories, firms) with the goal of software tools in this field delays the environment is a vivid demand. Enhanced producing a product while exchanging all massive spread of real VEs. security can contribute to the growth of needed information via a computer the number of realized VE solutions. In network. The lifecycle of a VE - formation, the future, management activities in VE

38 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

Wireless carriers are known for their low bandwidth, high latency, and unpre- dictable reliability. Wireless devices represent limited processing power, memory and simple user interface. All these limitations are addressed by the WAP, making it the most suitable platform for supporting sophisticated telephony and information services on hand-held devices. The most important parts of the WAP protocol stack are Figure 2. SmartX encapsulates both outlined in Figure 1. the smart card and the terminal application. The security layer in the WAP is WTLS (Wireless Transport Layer Security). This protocol is based on the TLS protocol, in XML, since HTML based Internet is which makes client-server authentication moving towards XML. The description possible and performs cryptographic of VE in XML including the enterprise operations. In addition, the application resources, data transactions and product level security can be accessed using the flows is a new approach, it will make far WMLScript (Wireless Markup Language more easy the building of the connection Script). However, in order to address most between the content and the Figure 1. The WAP protocol Stack. of the security requirements of the communication channels. By combining customers, some of the security WAP and smartX technology we obtain functionalities have to be performed in a a very flexible, secure mobile architecture tamper-resistant device. For this purpose which together with VE represented in will be accessed and controlled from WAP uses WIM (WAP Identity Module), XML will provide a safe business mobile devices as well. This access has usually implemented by a smart card, communication environment. to be very secure, and a general tool for possibly together with the SIM safety could be the smart card (SC). Smart (Subscriber Identity Module) card. Conclusions cards can be used for a variety of HTML and HTTP/TCP standards are not purposes, such as cryptography, According to WAP, not only a mobile appropriate for mobile communication identification (combined with biometry), client is able to access the services, but devices. The WAP and WML have been authentication (digital signature). also the server side is no more localised, developed to bridge this problem in such so needs of the mobility-based new VE way that wireless devices can be The aim of our project is to define and concept are fulfilled. In the case of smart connected to traditional WEB servers, describe a software architecture that card applications a new technology is tools and techniques. Secure properly addresses mobility and security smartX that defines a complete framework communication is very important in related needs of VEs. This includes to develop smart card applications. By wireless communication, and as the collecting a protocol set where security separating the application process (the management of virtual enterprises moves is based on smart-card technology and logic of the application) from the towards the mobile world, the which provides mobility in accessing application protocol (card-specific layer), development of a secure mobile virtual enterprise services. smartX makes quick and efficient communication architecture is a vivid migration to a new smart card possible demand. In order to reach a flexible and The Applied Techniques (see Figure 2). adaptable communication infrastructure In order to fulfil the compatibility the representations will have to be written requirements, in developing the new SmartX relies on SML (smartX Markup in different XML-based languages. software architecture, standardized Language), a description language that communication technologies have to be describes the smart card application data Please contact: used. The most widely used protocol to and processes. SML leverages XML István Mezgár – SZTAKI deliver the Internet content is HTTP, technology with a DTD (Document Type Tel: +36 1 209 6143 however, it does not behave in accordance Definition), which describes the E-mail: [email protected] with the mobile environment. Therefore, semantics and the grammar to define a new protocol - a de facto standard - has application protocols. SML is an Tamás Szabó – Nokia Hungary Ltd. Tel: +36 20 984 9074 been defined for the wireless world, called implementation of XML for the smart E-mail: [email protected] WAP (Wireless Application Protocol), card industry. which makes reliable and fast access of Zsolt Kerecsen – BablakSoft Ltd. data and services possible. Our activities in the close future will Tel: +36 1 327 5821 E-mail: [email protected] include the description of VE functions

ERCIM News No. 41, April 2000 39 SPECIAL THEME: WEB TECHNOLOGIES

Creating Hypertextual Dossiers in XML for the Italian Parliament

by Andrea Marchetti and Patrizia Andronico

During the last two years, a working group at the Deputies to collaborate on the creation of a website Institute for Telematic Applications of the Italian to be presented during the course of the Annual National Research Council (IAT-CNR) has been European Parliamentary Technology Assessment following the evolution of XML-related technologies. (EPTA) Conference held in Rome, December 1999. This group was requested by the Italian Chamber of

The EPTA Conference signalled the order to be able to extract phrases from Layout of the Website introduction of new high-tech equipment the texts, at the moment of their The information present on the browser into the Italian Chamber of Deputies. A presentation, and insert them as had to include: meeting-room for members of the Italian enrichments of the thesaurus entries. • the programme of the day; parliament was equipped with desks, each • the documents available for each with their own built-in PC and Internet For each of the topics identified, the speaker (full paper, abstract, access. The Conference website provided logical schemata used to structure the bibliography, others), directly access to the texts of the speakers’ material made available (presentations, accessible from the programme; presentations and documen-tation on the articles, thesaurus with its associated • the index of terms from the EPTA themes of the Conference. Links were documentation) were defined in the form Conference Thesaurus; built between the various documents on of XML DTDs. All the material was • the associated documentation. the basis of a thesaurus, consisting of edited in XML and experts then inserted This requirement led us to adopt a frame topics identified by experts in the fields further markups to signal thematic architecture. Each frame contained a type of interest. associations between different items. The of document. The frames were organised For each term of the thesaurus, material was then sent to IAT for so that the page was read in a circular documentation was created containing: validation and for an initial processing direction, following the order given in the • a definition aimed at transferring information above list. At the centre of the page, the • associated terms in the thesaurus extracted from the thesaurus and inserting thesaurus documentation frame formed • one or more quotations it in the documentation associated to the the central point of navigation. • pointers to related websites terms of the thesaurus. Selected images For each document type, a specific CSS • bibliography. were also associated with this stylesheet was defined. documentation. This documentation had to be collected, Conclusions processed and rendered homogeneous in At present, very few browsers allow the This application involved defining a a very short period (approximately 10 visualisation of XML pages. For this complex model to organise the days) by people working in different reason, all of the material produced was information. The IT and document geographical areas. A requirement on the translated into HTML and inserted into experts worked together on the application was that it should allow real- the website. development of this model: the IT time processing on the documentation, in operators were involved in the choice of

The Home Page of the Hypertextual Dossier. The Hypertextual Dossier during navigation.

40 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

documentation, and the documentation not yet fully support XML and associated Links: experts were similarly consulted with standards. Currently, the conversion into respect to the choice of technologies. This HTML is performed off-line. In the near EPTA: http://www.atkinsoft.ndirect.co.uk/epta/ integration contributed greatly to the future it is expected that it will be possible Apache project: success of this experimental application; to develop a version of the site in which http://xml.apache.org the website was set up much more quickly the documents remain in XML format, than could have been expected. being translated into HTML only when Please contact: requested by a browser that does not Andrea Marchetti or Patrizia Andronico – The site has functions (such as the support XML. The results of the XML IAT-CNR consultation of the documents in various project Apache may be used for this. The Tel: +39 050 315 2649 or +39 050 315 2090 languages) which have not yet been site can be consulted via Internet at: E-mail: [email protected] or [email protected] activated, partly as a result of lack of time http://epta.camera.it, selecting the link and partly because available browsers do ‘Hypertextual Dossier’

Providing the Context for Business Decisions by Brian Matthews

Information is the major asset of business today. relevance and as a consequence, may not be able to However, if there is insufficient knowledge about the make correct business decisions. The problem is to business context in which information is supplied, provide the correct information to the correct user at then the user may not be in a position to judge its the correct time.

Information Systems Engineering at The project is using an XML be in audio or video formats. The RDF CLRC is working with a major representation of this UML model to description can also be used to apply international communications company transmit it via the corporate Intranet. This constraints on the data model, or to explore ways to enhance the gives the opportunity to present it to the constraints on the contextual information dissemination of the corporate knowledge user within widely available XML enabled itself. As the data model becomes subject assets. As the Web has become a browsers, providing a universal and to extension and restriction according to universal medium for both external and economic interface onto the data model. the circumstances in use, additional internal communication, it is natural to The use of an XML format combined with context can be provided to track the use it for this project. The emerging Web the scripting languages such as JavaScript, changes and record the additional technologies, especially the Extensible and the Extensible Stylesheet Language reasoning behind its evolution. Thus an Markup Language (XML) and the (XSL) allows the user to move between audit trail can be established through the Resource Description Framework (RDF) different views of the data within the variants on the model. Again all this can offer a powerful means for managing and browser, without reference to the server. be presented and edited across the WWW distributing data as well as the context in This allows the possibility of user profiles using standard browsers. which it is presented. providing customisable views. Further, the data model can also be modified in Please contact: The company maintains a single data XML and transmitted back to the central Brian Matthews – CLRC model for corporate information. This repository. Thus an easy to use editor can Tel: +44 1235 44 6648 forms the basis for generating particular be distributed to the user via their desktop E-mail: [email protected] data models to be used for particular br o w s e r . situations, such as a product management system, marketing system, or customer As the model is developed, decisions are handling system. By conforming to a taken on the design of the model. These common basis, the exchange of decisions and the reasons behind them are information between these systems vital to the understanding of any future becomes easier, allowing rapid and user. RDF is being used to provide such consistent information to be transmitted. contextual information, which is supplied However, maintaining and using this to the end user who can then discover the model is a real problem. If the model is context in which the model has been hard to understand and explore, then it is designed. Thus, links to explanatory text, hard to persuade systems developers to constraints, and company standards can use it. be provided. Such explanations can also

ERCIM News No. 41, April 2000 41 SPECIAL THEME: WEB TECHNOLOGIES

Virtual Physics and Astronomy

by Detlef Skaley and Udo Zlender

The Virtual Physics and Astronomy (VPA) project is developed at GMD. It is intended for schools, a Web-based innovative learning platform, with universities, or self-study to learn physics and multimedia capabilities and intuitive user interfaces, astronomy.

Today’s life requires accessing Examples Interesting experiments can be performed information and knowledge in relatively The virtual optical bench provides by modifying the optical properties of short time scales, which means new examining the laws of geometrical optics. complete instruments by changing the methods of knowledge transfer must be The student may place lenses, mirrors, properties of its elements eg, changing established. The Internet has brought screens, and light sources on a holding the refraction indices or radii of curvature forward enormous technical possibilities bar to study the light rays in detail. of the lenses (see figure 1). which not only can compete in many ways with classical text books, but also provides new grounds through it’s multimedia capabilities. Visualisation has always been regarded as an important element to aid in understanding complex scientific theories and experiments which up to recently required rigorous special soft- and hardware requirements. Nowadays, however, graphics, videos, animations, together with appropriate description texts can be accessed directly on the Internet using an ordinary personal computer. Moreover, new learning tools can be established in combination with exercises. In this way the student can learn by experimenting in the virtual worlds. For instance measurements with a simple stop watch could be sufficient to carry out some of these exercises such as computing the mass of a planet.

The project Virtual Physics and Astronomy (VPA) is an innovative Figure 1: Maksutov telescope on virtual optical bench. learning platform, primarily considered for schools, universities, or self-study to learn physics and astronomy. Through its multimedia capabilities and intuitive user interfaces, VPA enhances the learning effects. The latter is further amplified when visualisation is combined with interactivity which allows the learner to explore the virtual environments individually. Next to 3D representation accurate programming of scientific simulations are essential for such a learning platform. VPA simulations are carried out in Java, hence a standard Web browser, and some additional and freely available plug-ins are the only requirements for using the program on the Internet. Figure 2: The Foucault Pendulum.

42 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

Physical experiments like the one relative to the ecliptic plane from different of real images, videos and VRML models performed by Jean Leon Foucault in 1851, viewpoints. In addition calculations of of those examples clarifies the important by which he proved the earth’s rotation physical laws in real-time enable the results from that recent field of research. can be understood using the Virtual realistic representations of the planets’ In here a student might be instructed to Reality Modelling Language (VRML). motions. The latest development of this observe and measure the distance of a Virtual Foucault’s Pendulum may aid the software provides Java3D stereoscopic galaxy from the oscillation period of a user to determine the geographical 3D-presentation using recent of-the-shelf cepheid star. Or perhaps consider two latitude by measuring the rotation velocity 3D-graphics hardware (see figure 3). penetrating galaxies, where the mass of of the pendulums plane of oscillation (see the inner galaxy can be determined from figure 2). Discovering supernovae explosions, the rotational period of the outer galaxies’ neutron stars, black holes, or quasars are debris. The examples mentioned here can The Virtual Planetarium allows facts of some of the outcomes of modern be found among many others under the our solar system to be explored. By astronomy. Virtual astronomy can help following URL: http://imk.gmd.de/delta. navigating freely in 3-dimensional space, to illustrate and explain some of the the student can study the planets’ orbits exciting cosmic facts. The combination Techniques Although the project consists of a wide range of different topics in Physics and Astronomy, the underlying structure of each individual application is very similar. Most of them consist of three major parts: • Computer Simulations and Calculations: Are required to give insight into complex dynamical processes. They might be written in any desired programming language like C, C++, Java etc. • Network Connection: Interfacing between client (visualisation front-end) and server applications (simulation providing site) can be realised by Common Gateway Interface (CGI, Perl) or Remote Method Invocation (RMI, Java). • Visualisation and Animations: Are done local on the users machine in Java, Java3D or with help of the Virtual Reality Modelling Language (VRML). Static examples are just downloaded and displayed (VRML), animation of Figure 3: Virtual Planetarium. parts of the scene require Java, or Java in combination with extensions to VRML eg External Authoring Interface (EAI).

Promising future developments like MPEG-4 will be used as they become generally available.

Links:

http://imk.gmd.de/delta

Please contact:

Detlef Skaley – GMD Tel: +49 2241 14 2987 E-mail: [email protected]

Udo Zlender – GMD Tel: +49 2241 14 2485 Figure 4: E-mail: [email protected] Virtual Astronomy.

ERCIM News No. 41, April 2000 43 SPECIAL THEME: WEB TECHNOLOGIES

CommonGIS – Common Access to Geographically Referenced Data

by Hans Voss, Natalia Andrienko and Gennady Andrienko

Public access to the immense volume of existing geo- available geo-data is, however, of little use unless people and geo-referenced thematic data and their exploitation can easily access and easily exploit it. CommonGIS, a is of significant value for the development of an open research project coordinated by the Knowledge and democratic ‘information society’ and a true global Discovery Team of GMD Institute for Autonomous market. The widespread use of such data and GIS Intelligent Systems, is developing Web-based tools for (geographical information system) will promote general access and analysis of geo-data that can be utilized public awareness and further social cohesion. Publicly likewise by skilled and casual, non-expert users.

Geo-data encompass various thematic or Web-based, intelligent GIS for supported by suitable editors and statistical data on demography, economy, automated thematic mapping have been checkers. So one specific task of education, culture, history, etc., which are implemented. CommonGIS was to develop a language associated with objects and locations in • Several demonstrator applications that is applicable to any domain of space. Probably the best way to explore showing the potential of the new spatially related thematic data, and to such data is to visualize them on maps. technology to data providers and users provide tools that make it easy for data The key-thought of CommonGIS were put in place. providers to build their own applications. (http://commongis.jrc.it) is thus to make geo-data commonly accessible and usable In the following we will address these As a result, a so-called data characteri- for everyone, from everywhere, by achievements in some more detail. A zation scheme (DCS) was developed. In providing a Web-based Geographical basic requirement was that CommonGIS contrast to other work known from Information System (GIS) with should easily connect to existing data sets, literature, the DCS provides a rich arsenal knowledge-based functions for the which could be stored in large databases of concepts, data structures and operators automatic generation of thematic maps. or kept in spreadsheets, just to name for defining semantic domain models of To a very large extent, the user should not extreme options. For this purpose, the given data. The reason for this at all need to worry about visualization CommonGIS provides a set of converters particular emphasis is that the focus of issues but rather focus on the problem for various table data formats as well as previous work was mostly on the visual solving process and analytical task. data base connectivity. Internally, all presentation of data on then static maps, geographic and thematic data are stored while in CommonGIS we are dealing with Ordinary GIS are not well suited for this in a data base (ODBC/JDBC interface). very interactive maps. In addition, we want because they suffer from at least one of Aside from this base functionality, the to support users in handling more complex three intrinsic pitfalls: they tend to be proper goal of providing smart support analytical tasks. For example, in the architecturally closed, monolithic, and also to non-expert users can only be context of exploring some demographic costly environments; their use requires satisfied if the application program itself data the CommonGIS system should specific technical skills; they require the incorporates knowledge: knowledge become able to automatically identify, user to think more about obtaining about the data of the specific application, formulate, and support analytical tasks like visually nice presentations than about the and knowledge about principles of “Compare gender structure of population selection of appropriate data and solving cartographic visualization and analysis. in different age groups” or “Look at the problems with their use. The distribution of a specific age group (0-14 CommonGIS project is now (as of As basis for the latter we utilize the years, 15-64 years, or older than 64) across February 2000) in its 16th month, and knowledge base for cartographic the countries”. In the latter example, it current achievements include: visualization from Descartes could also propose that using relative • A data characterization schema (DCS), ( h t t p : / / b o r n e o . g m d . d e / a n d / j a v a / i r i s / ) , values (age group in percentages of the and a data characterization language which was developed at GMD’s AiS total population) would probably make (DCL) as an instantiation of the schema institute (see ERCIM News, July 1998). more sense than using absolute numbers. has been defined. The DCL is used to In addition, some ideas are being describe properties of and relationships incorporated from the cartographic The DCS was defined using UML, and between attributes of a set of thematic visualization system VIZARD from the was studiedly kept on a conceptual level. data set that shall be published. The Fraunhofer Institute FhG-IGD. Regarding It can thus be used as a schema that can semantics, in particular, is utilized to knowledge about the application data, the be instantiated by different syntactical further automated support of data original Descartes system also included variants. For use in the CommonGIS visualization and analysis. some means to describe the semantics of software one such language was XML, • A prototype for the support of building the data. However, the language used for and corresponding parsers were new applications (application builder) defining characteristics of and relationships generated. This instantiated schema is has been realized, and the resulting between attributes was somewhat ad-hoc, called DCL (data characterization application modules comprising a new not documented, and its usage was not language).

44 ERCIM News No. 41, April 2000 SPECIAL THEME: WEB TECHNOLOGIES

Figure 1: User interface of CommonGIS in the Internet. Figure 2: Pie charts are generated for showing the age structure of the Figure 3: Pie charts: focus on areas population. with relatively small population.

The CommonGIS software is developed A demonstrator of CommonGIS using this basis, one of the proposed as an open, object-oriented, distributed Portuguese Census Data is available in visualizations is a pie-chart presentation system with a client-server architecture. the Internet. It may be run from either we b (see figure 2). The sizes of circles are The user interface is realized as Java- site, http://commongis.jrc.it/ commongis/ proportional to the total population in the applets, thus providing comfortable sw/commongis_first/CommonGIS/ or municipalities of Portugal, and the sizes access and interactivity while only h t t p : / / b o r n e o . g m d . d e / d e s c a r t e s / of the segments show proportions of requiring a standard (Java-enabled) Web- CommonGIS/. After the applet is loaded, different age groups. browser. Currently, what the user sees it connects to the server, retrieves after starting CommonGIS in the Internet information, and displays a map. The From the sizes of the pie charts one can looks like something that one would HTTP protocol is used for communication observe that population numbers in the expect to see when running a typical GIS between the client and the server. This capital, Lisboa, and in its surroundings on a local PC (see figure 1). In fact, what allows the system to run in the Internet are much larger than in other districts. was taken as the base technology is the or in Intranets without being disabled by This makes the signs in those districts too Java-based Lava/Magma GIS from PGS firewalls. If a geographic layer is small for seeing proportions of different (see. http://www.pgs.nl). A salient feature associated with thematic information, this age groups. As a remedy, all visualization of Lava/Magma is that it performs information can be visualized on a techniques are supplied with specially sophisticated caching of data for thematic map. For this purpose one designed controls for the interactive optimization of performance. This is pushes the ‘T’ (Thematic data) button. In manipulation of the display. In the particularly needed when using larger response the system will list all available example one may decide to not show maps that would take too long to thematic variables. After the selection of signs for the largest districts just by download as a whole. One achievement one or several variables CommonGIS will moving a slider, and thereby implicitly of CommonGIS was to make the propose one or more suitable visualization enlarge the small signs so that their Descartes system running from within methods that are selected on the basis of structure becomes better visible (see Lava/Magma. The user can thus select the available semantic information about figure 3). and define a desired map with ordinary the data. GIS operations, and then select certain Currently, two pilot applications are thematic variables for visual display on Using the Portuguese example, one may running. The above mentioned one was this map. Thereby the full interactivity of select four variables with population developed by CNIG (Centro Nacional de Descartes stays available as the systems numbers of different age groups. Informacao Geografica, Portugal), which were redesigned so that Descartes can use CommonGIS will take into account is also responsible for the overall usability the Lava/Magma display methods for relationships among the variables: it analysis of CommonGIS. Another presenting the results of interactive knows that they are comparable and all application is under development by manipulations. together make the total population. On Dialogis GmbH together with the City of

ERCIM News No. 41, April 2000 45 RESEARCH AND DEVELOPMENT

Bonn. Other applications for customers of Descartes. Furthermore, the consortium Links: of PGS and InGeoForum (which is is taking efforts towards the standardization affiliated to ZGDV – sister institution to of the data-characterization schema. It is Homepage of the project: http://commongis.jrc.it/ FhG-IGD – and to the Hessian State expected that a standardization in this Demonstrators: Surveying office) will soon follow. The field will promote ‘intelligent’ GIS http://commongis.jrc.it/commongis/sw/ two industrial partners of the consortium software interoperability, thus broadening commongis_first/CommonGIS/ (PGS and Dialogis) are developing and a potential market. The Joint Research ht t p : / / b o r n e o . g m d . d e / d e s c a r t e s / C o m m o n G I S / Further publications about the project: will commonly market a commercial Center of the European Commission EC- http://borneo.gmd.de/descartes/ version of the software. They will also JRC, Institute ISIS, is co-managing the incorporate results of CommonGIS in whole project effort with GMD, and Please contact: their current commercial products, together with its subcontractor GISIG, Hans Voss – GMD Lava/Magma and DialoGIS, respectively, will be instrumental in the dissemination Tel: +49 2241 14 2532 where DialoGIS is a commercial version and standardization process. E-mail: [email protected]

StraDiWare: Linking Business and IT Models using Graph Theory

by Tony Conway

The aim of StraDiWare is to formalise the linkage This will be achieved by a heterogeneity-reconciling between quantified business objectives, decisions to linkage mechanism between the various tools and implement IT systems and the IT systems themselves. information representations. The use of graph The delivered system will also provide formal linkage language formalism is seen as the essential feature descriptions between dependencies within IT projects. of this mapping.

At present IT systems rarely satisfy the is typically a long duration process that their development methodologies, can aim of supporting business objectives: is at present partially supported by several nevertheless use StraDiWare to there is thus a high level of business discrete methodologies. The benefits of understand the linkages between business dissatisfaction. The StraDiWare project using the StraDiWare tool set and objectives and IT projects. A major goal aims to reduce this by coupling business methodology will be that business and IT of the project is to produce an objectives to IT support. There is a large managers can see at a glance: implementation of the CASSP (Computer and growing market for analysis and • the impact of a change in business Aided Support for Strategic Planning) recommendation of remedial actions for objectives on existing IT systems and software tool. pre-existing (‘legacy’) IT systems and for planned future developing IT systems the development of new systems strictly • the effect of change in the design / StraDiWare brings together appropriate controlled by quantified business implementation of an IT system on the end-users in realistic business situations objectives. achievement of business objectives in the IT support development cycle with through a multi-media representation IT SMEs in both CEE and EU countries Thus the StraDiWare project supported by a formal model. This will supported by one CEE and one EU complements Action 18 "Euromethod provide monitoring and control of the research Institute each with appropriate pilot projects in the CEE countries" in the developing IT project in terms of (commercially tested and developing) PECO Action Plan by providing the quantified business objectives. expertise to transfer to the IT SMEs. essential linkage from IT projects (managed using Euromethod) to higher The proposed delivery mechanism for the StraDiWare is a Research & level business objectives and processes methodology as realised in the Technological Project supported by the (outside the scope of Euromethod). StraDiWare software product range will Commission of the European Union Nevertheless, StraDiWare is not simply be via Web technology. The universality under the Copernicus scheme. an extension of Euromethod: it is equally of browser availability and advances in applicable to other IT development technology recommendations from W3C Links: methodologies, such as SSADM. (notably XML and RDF) combine to StraDiWarehome page: make this an appropriate and cost- ht t p : / / w w w . i t d . c l r c . a c . u k / A c t i v i t y / S t r a D i W a r e StraDiWare is concerned with the effective delivery mechanism. business process in which enterprise StraDiWare will use this advanced multi- Please contact: objects interact with the implementation media user interface to produce outputs Keith G Jeffery – CLRC of IT projects: this is referred to as that are easily understood by non-expert Tel: +44 1235 44 6103 Strategic Planning (SP). SP is a rather users. This is necessary so that business E-mail: [email protected] complex task that has several basic stages managers, who may not be familiar with and which produces various products. It the technical aspects of IT projects and

46 ERCIM News No. 41, April 2000 RESEARCH AND DEVELOPMENT

Computer Modelling in Geosciences by Stanislav Saic and Josef JeÏek

The progress made in computer technology during numerical solutions and using powerful computers. the recent years allows to solve more difficult tasks A typical example is the three dimensional dymamical also in the field of geoscience. Many new results can modelling of rigid objects moving in viscous fluid and be obtained by reconsidering old mathematical/ the flow around them.

In ductile shear zones and in metamorphic Analytical or numerical integration of Jeffery contained some errors (most rocks in general, the geometry of fabrics these equations give the trajectory of a probably misprints) that did not affect the of rocks is a valuable source of particle in a flow described by a velocity equations of rotation of the inclusion, but information on the kinematics and gradient tensor. Jeffery’s work has been significantly affected those for the dynamics of flow. In order to understand broadly used and discussed by those who surrounding velocity field. We present a the fabric development, structural studied the development of shape- procedure based on the equations of geologists use numerical or analogue preferred orientation in rocks. Jeffery also Jeffery that were re-developed and modelling of the behaviour of rigid developed equations that can be used for corrected. We found an efficient way to objects embedded in a viscous matrix. numerical evaluation of the surrounding compute the elliptic integrals that are Numerical models are based on velocity field. It is surprising that the needed for the solution and produced continuum mechanics and, with more or equations have not been used for this software for computer modelling and less accuracy, describe the motion of a purpose up to nowadays. The possible visualisation of the problem. In the set of rigid body and/or the flow around it. reasons are a rather complicated three figures we demonstrate a result of procedure of development, the need for modelling with an application to structural A pioneering work in this domain was an efficient algorithm for the numerical geology. carried out by G.B. Jeffery in 1922, who evaluation of elliptic integrals in Jeffery’s developed equations of the rotation of an formulas and a possibility of computer The porphyroblast is approximated by a isolated rigid ellipsoidal particle in a implementation of the whole problem. growing ellipsoidal inclusion rotating slowly moving Newtonian viscous fluid. Moreover, the original manuscript of around its longest axis parallel to the rotation axis of simple shear. Markers representing external foliations were initially arranged in lines and moved with the viscous fluid until they were captured by the growing porphyroblast.

We conclude that, in comparison to recently used procedures, the procedure and computer program we have developed represent a fully three- dimensional hydrodynamical model. It permits considering ellipsoidal inclusions and a general type of flow, and therefore to study how the asymmetry of the inclusion and different kinematic conditions influence flow patterns and structures around rigid objects in a ductile matrix.

The research was supported by the Grant Agency of the Czech Republic under Grant No. 102/96/0419.

Please contact:

Stanislav Saic – CRCIM Tel: +420 2 6605 2211 E-mail: [email protected] A two-dimensional cut through a rigid ellipsoidal particle rotating in a viscous flow in a regime of simple shear. Arrows indicate the velocity field in a set of chosen markers (point) of the moving fluid.

ERCIM News No. 41, April 2000 47 RESEARCH AND DEVELOPMENT

Efficient Use of Networks through equal Bandwidths

by Annette Kik and Henk Nieland

Network providers offering integrated services aim at transmission an equal portion of the bandwidth. This optimal use of the available capacity. The best method result follows from research at CWI on queueing to satisfy the very diverse user requirements also models for performance analysis of telecommunication when the network is almost full, is to allot to each data systems with integrated services.

In integrated services like (mobile) number of phone calls), so does the opposed to using the ‘law of the jungle’, telephone with Internet, Web browsing capacity for elastic traffic. Efficiency is which gives every user instant access if or Internet with video communication, improved by applying the method of capacity is available. Implementation in the various traffic flows need different ‘processor sharing’, in which every user the switches of a network both guarantees treatment. For example, speech should is immediately allotted an equal portion quality-of-service and achieves a high not suffer much delay, lest it becomes of the bandwidth. Here the average network efficiency. incomprehensible, whereas for e-mail this waiting time is determined by ‘first order’ is less problematic. In dealing with these factors like the mean size of the data Integrated services in mobile problems it makes sense to distinguish rather than the variance (in First-Come- communication is at present a focal point between ‘stream’ traffic and ‘elastic’ First-Served models the dependence on of interest, as is demonstrated, for traffic. The first mainly consists of real- variance can be detrimental to example, by considerable media attention time connections (telephone, video), performance). CWI research showed that for the Wireless Application Protocol which are prone to delay in the this favourable feature of processor (WAP). Current research at CWI transmission. This requires some sharing is preserved under varying concerns comparison of several guarantee in capacity. Elastic traffic (bulk network capacity. integration strategies of data and real-time data, e-mail) allows fluctuations in in a mobile network. Here performance transmission speed as long as the total In collaboration with Dutch Telecom analysis is even more important than in transmission time remains ‘acceptable’. (KPN Research) the optimal integration fixed networks, because the capacity is The present Internet is focused on strategy for stream and elastic traffic was much more limited. handling elastic traffic: transmission time found to be a division of the bandwidth increases with the number of users. into two distinct parts. One part is Please contact: reserved for elastic (data) traffic only, in Sem Borst – CWI The network can be used more efficiently the other preference is given to stream Tel: +31 20 592 4092 by integrating stream and elastic traffic. traffic. If the latter needs less capacity, E-mail: [email protected] For example, pauses in a telephone the unused capacity is given to the elastic conversation can be used for sending e- traffic. This strategy provides some Sindo Núñez-Queija – CWI / INRIA Tel: +31 20 592 4168 / +33 4 9238 7637 mail. Since stream traffic varies (eg, the protection for both types of traffic, as E-mail: [email protected] / [email protected]

Network capacity is efficiently used by simultaneous transmission of stream traffic (speech, video) and elastic traffic (data). At time t the available bandwidth for data (ct ) is divided among nt packets to be transmitted.

48 ERCIM News No. 41, April 2000 RESEARCH AND DEVELOPMENT

Interactions in Virtual Environments by Lakshmi Sastry and David Boyd

At CLRC Virtual Reality Centre, located at Rutherford reviews, operational training and scientific Appleton Laboratory, work is underway to implement visualisation over the period of the last four years. A a 3D interaction toolkit as part of the INQUISITIVE VR system specific subset of the toolkit is an project. The toolkit architecture has been derived from Engineer’s Toolbox (ET) which is being developed as a task-based analysis of a range of immersive VR- part of the smaller project Delivering Advanced Virtual based applications for large-scale engineering design Interactions to the Desktop (DAVID).

Virtual Reality (VR) technology and techniques have the potential to deliver intuitive user interfaces with more natural styles of interaction between the user and his application which will enhance productivity and performance. Facilities for user interface and interaction development within today’s VR systems are rudimentary, limited and limiting. The availability of a generic interaction toolkit with portable modules makes a significant contribution to the rapid development and successful application of 3D VR interaction techniques to a wide range of virtual environments.

The aim of the RAL component of the INQUISITIVE project is the development A virtual environment for Atlas Inner Detector alignment checking. of a portable interaction toolkit for VR applications which will improve support those techniques. The application tasks Interpreter is independent of the host for developing user interaction within are then achieved by suitable VR system. task-oriented virtual environment combinations of these interaction • An Interaction Manager which applications. The design of the interaction processes. Based on this analysis, the monitors the changing state of user toolkit is based on our extensive main functional components which the interaction within the VE. It receives experience in developing and successfully toolkit must provide are: the event tokens from the Contextual deploying immersive end user • a set of interaction processes for the Interpreter. It finds out the current state applications for large scale engineering four classes of basic interaction tasks within the VR system’s runtime object design reviews, operational training and • a set of generic virtual interaction database through the VR run-time visualization within virtual environments. objects such as toolbox manager. The operation of the • a run-time interaction framework. Interaction Manager must be Toolkit Architecture customised for each host VR system. All application tasks, however complex, The components of the runtime can be implemented in terms of a interaction framework are: This architecture can be seen to map on combination of tasks from the four basic • A Contextual Interpreter which to existing VR system kernels such as classes of user interaction – navigation, interfaces to the device drivers to obtain Maverik from the University of selection, manipulation and data input, in the measures and triggers from the input Manchester, UK and dvMockup from virtual environments. Each basic devices and convert these into the VE Parametric Technology. interaction task can be realised using a co-ordinate system. It then takes the number of possible interaction techniques. dynamic constraints and current state Requirements Capture, For example movement can be of the interaction objects provided by Analysis and Modelling Methodology implemented using the magic carpet or the Interaction Manager and interprets We are using UML, the Universal point-fly techniques. Each application will the measures and triggers in that Modelling Language, to capture and identify one or more interaction context. Finally the Contextual transform high level user interaction techniques appropriate for carrying out Interpreter calls the appropriate techniques and tasks into detailed the tasks required in that application. This interaction processes to generate the interaction processes between the user, in turn will guide the definition of the event tokens which it passes to the the system and the underlying application interaction processes needed to realise Interaction Manager. The Contextual leading to an implementable design. The

ERCIM News No. 41, April 2000 49 RESEARCH AND DEVELOPMENT

use case methodology is being used to Proof of Concept – ‘virtual engineering toolbox’ which will identify the core classes that each basic An Engineer’s Toolbox include intuitive handling operations on interaction task will require and the Over the past four years we have model parts for simulation of assembly derived classes and class variables and developed many VR-based applications and maintenance tasks as well as sets of methods for each interaction technique. for engineering design review, operational tools for measurements and alignment We have realised that even the simplest training and computational data checking tasks in immersive and desktop interaction process in VEs requires visualisation in collaboration with other environments. In addition, it can include several agents co-operating together and departments as part of supporting RAL’s facilities for superimposing and monitoring various system states engineering and scientific activities. The visualising analytical results from concurrently for task achievement. This figure shows the VR model of the Atlas computational simulations such as metal places significant requirements on Inner Detector with lines-of-sight laid out deformation, fluid flow, thermal and communication between the components to check clearance and clashes. Such stress analysis simulations. The of the interaction toolkit and between the applications have demonstrated the implementation is currently under interaction toolkit and the VR system potential of virtual reality technology and progress using the commercial VR system manager. The interaction, activity and interaction techniques to deliver cost dvMockup which links closely with the class diagrams are helping to clearly effective productivity benefits and greater ProEngineer CAD software. identify the appropriate classes and insight into engineering design and entities and how they collaborate within computational data. The utility of the Links: each process. Its ease of use and its ability INQUISITIVE Interaction Toolkit is to The INQUISITIVEProject: to describe and analyse interaction be realised via the DAVID project to ht t p : / / w w w . i t d . c l r c . a c . u k / A c t i v i t y / I N Q U I S I T I V E / processes from both the perspectives of deliver an application toolkit to the user requirement and of communication engineers and scientists so that use of VR Please contact: between the user, the interaction toolkit, becomes part of their design and analysis Lakshmi Sastry – CLRC the VR system and the user’s application methodology. Thus the aim of the Tel: +44 1235 44 6892 is promising. DAVID project is to develop a portable E-mail: [email protected]

Alternative Processing of Verbal Data

by Milan Mare‰

The vague quantitative values which we call here of fuzzy numbers are well known. It is worth verbal data are usually processed by means of the mentioning that these properties are not completely theory of fuzzy numbers or fuzzy quantities. This identical with the properties of deterministic numbers theory is well elaborated and the algebraical properties ñ but they are well known and analytically managed.

The attempts to apply the theory of fuzzy The everyday experience with practical discussion, which was already done. Let quantities to the formal analysis of practical manipulation with vague data (and many us mention, at least briefly, its essence. situations like, eg, the CPM (critical path traditional techniques do manipulate with method) optimization techniques show them without admitting that they process It appears to be the crucial point that the another serious discrepancy connected fuzzy quantities – let us remember operations based on the extension with this, today already classical, quantitative units like ‘pinch’, ‘handful’, principle do not reflect the specific processing of fuzzy quantities based on so ‘step’, ‘two-days marching distance’, semantic structure of the fuzzy quantities called extension principle. It is used for ‘proportional weight’, ‘acceptable profit’, modelling the verbal data. Namely, these the summation and multiplication of fuzzy etc.) shows that the results of it are much fuzzy quantities are, in fact, combinations quantities and it has a very logical formal less dispersed and their realistic values of at least two components. Verbal datum structure consisting in re-combination of are concentrated near some periodic like, eg, ‘approximately 8’ consists of a possible values and their possibilities of achieved ‘expected’ values. As such ‘crisp core’ ñ its proper quantitative value the input fuzzy quantities. Anyhow, the practical vagueness is usually connected (in our case 8) and a semantic ‘shape’ extension principle enlarges the range of with verbal description of quantitative mathematically describing the structure possible values (namely in the case of values, we use the term ‘verbal data’ for of vagueness connected with the verbal summation) in a significant measure and, their characterization. The above practical expression (in our example with namely if it is applied repeatedly in experience shows that there exists a ‘approximately’). Meanwhile the crisp processing a large number of fuzzy contradiction between the fuzzy set cores are real numbers in the deterministic quantities, the extent of possible values of theoretical model and real state regarding sense and they can be processed by means the resulting fuzzy quantity becomes the processing of such verbal data. Its of classical deterministic operations, the enormously (and unacceptable) large. roots may be quite deep and worth for shapes are mathematically real-valued

50 ERCIM News No. 41, April 2000 RESEARCH AND DEVELOPMENT

functions mapping the real line into the leading to a deterministic real value, and structure of relation between vague and unit interval with modal value 1 achieved logical or semantical operation over crisp component (eg ‘something between for the argument 0. They can be shapes leading to a resulting shape. This 20 and 25’, ‘a bit more than 10’, etc.). In interpreted as some kind of normalized output crisp core and shape define the all these cases it offers the possibility to membership functions of the fuzzy fuzzy quantity which reflects the result reflect the logical relation between the quantities modelling the verbal data. As verbal datum. Due to the applied operation processed verbal data much better than the shapes are connected with the semantic over shapes the possible values of the the traditional approaches limited to the component of the fuzzy datum, they would result are more or less closely concentrated extension principle. be rather processed by fuzzy logical near their crisp core (it is worth methods using operators like maximum, mentioning that usually much more Please contact: minimum, or product, etc., even if the closely than if the extension principle is Milan Mare‰ – CRCIM formula of the extension principle is also ap p l i e d ) . Tel: +420 2 688 4669 one of the applicable methods. E-mail: [email protected] The above method can be very easily Each algebraical operation realized over completed also for the manipulation with such verbal data is to be decomposed, verbal data having ‘anonymous’ crisp core then, into operation with crisp cores (like ‘several’, ‘few’, etc.) or combined

Towards the Distributed Virtual Education and Training Environment by Igor Nikitin, Valery Burkin and Stanislav Klimenko

Interactive installations in a virtual environment, introductory course to correspondent discipline. supporting the educational process, are created at Related video-films, recorded in a virtual environment, GMD Institute for Media Communication. Each are offered as illustration material. installation is reflected in the Internet as an

The modern educational processes actively uses the methods of computer graphics and scientific visualization. Wide opportunities are opened by the emerging technology of virtual environments, which can be used for the creation of highly interactive virtual laboratories intended for teaching different disciplines. GMD’s Institute for Media Communication participates in the INTAS project ‘Visualization of Complex Physical Phenomena and Mathematical Objects in Virtual Environment’, supporting the development of virtual This cross-eyed stereopair shows a sample object from relativistic string laboratories in natural sciences. Three theory: world sheet. Arrow points to physical singularity (4D pinch point). installations are currently created in GMD’s immersive virtual environment system CyberStage, serving education in accompanied by a web-tutorial on the because it accumulates almost all higher mathematics, theoretical physics, topology of low dimensional manifolds, mathematical systems used in modern astronomy and astrophysics. introducing all necessary concepts, and theoretical physics. To attend this illustrated by 3 video-films. Internet-course, no special pre-requisites Topological Zoo are needed, except 1st year University This installation demonstrates the main In the Microworld lectures on higher mathematics. The properties of elementary topological This is an elementary introduction to a course includes a lot of graphical material, objects (Moebius band, Klein bottle, theory of relativistic strings, a branch of particularly, stereoscopic movies in the projective plane etc.) and gives the theoretical physics, exploring the form of stereopairs and anaglyphs. possibility to explore interactively these properties of elementary particles. String Several applications on visualization of complex constructions arising in higher theory is selected due to its increasing string dynamics are available, playing the mathematics. The installation is popularity among the theorists and role of interactive practicum. The course

ERCIM News No. 41, April 2000 51 RESEARCH AND DEVELOPMENT

is also supported by an installation in of all the objects in a scene, interactive the objects necessary for presentation. All CyberStage. It is a prototype for a virtual map of constellations and tools for search applications can be adapted to simpler lecture room of the future, where complex and navigation. projection systems for demonstrations on animated 3D forms, created by theoretical the screen of the University’s lecture hall physics, can be directly observed and Applications or the wall of real planetarium. interactively studied. The related course The described courses and videos were material can be displayed on virtual used in the educational process at Links: screens using virtual slide projectors. Moscow Institute for Physics and Introductory course on string theory: Technology, University of Helsinki, http://viswiz.gmd.de/~nikitin/course Virtual Planetarium University of Kaiserslautern and other Tutorial on topology of low dimensional Experimental course on basics of educational institutions in Europe and manifolds: astronomy, delivered inside the USA. Virtual laboratories in GMD’s http://viswiz.gmd.de/~nikitin/vismat_html Virtual Planetarium in CyberStage: immersive virtual environment system, high-end immersive virtual environment http://viswiz.gmd.de/~nikitin/stars/ and describing interactively all systems were presented at several local starsCB.html surrounding astronomical objects and remote exhibitions, students and Related research projects: (planets, constellations, stars, nebulae scholar excursions. http://viswiz.gmd.de/~nikitin/viz.html etc). Currently the installation includes Please contact: 3D models of 30 objects in the Solar The content of virtual laboratories can be System, 3200 nearby stars, a large integrated to Internet-accessible library Igor Nikitin – GMD database, containing textual descriptions of components, where the teacher can get Tel: +49 2241 14 2137 E-mail: [email protected]

Wavelets improve Information on Earthquakes

by Henk Nieland

Tremors caused by an earthquake contain valuable in the signal are revealed that otherwise would remain information on its nature. Hence, close examination hidden. The research was carried out jointly with the of these often very fanciful vibration patterns is highly Dutch National Meteorological Institute KNMI in a important. CWI has analyzed seismic signals with the project financed by the Dutch Technology Foundation wavelet method. It turns out that in this way structures STW.

Apart from natural earthquakes, which remain unpredictable so far, there are also those caused by man, in order to explore the structure of the earth’s crust (for example to trace possible oil sources). Vibrations reflected on the higher earth layers are recorded at an observation station and analyzed. Such a seismic signal consists of several waves (phases). Precise determination of the arrival times of two important waves, the P-phase (primary) and the S-phase (shear), is a Observed reflections on earth layers of seismic vibrations. On the left the crucial step in fixing the quake’s type and original distorted data, on the right the same data after cleaning up with the location. The real art is to extract these wavelet method. two waves from the total signal, in which other waves appear alongside several mini-quakes in Limburg and the set of ‘non-local’ quakes (Japan, types of noise. An algorithm based on Ardennes, etc.). The difference with California, etc.) will follow. wavelets was developed that manual determination of the arrival time automatically deducts the S-phase from remained far below the accepted norm A wavelet is a wave packet of small the signal, given the precise arrival time and was clearly smaller than what the best extension. By shifting it as a mould along of the P-wave (the S-phase always arrives method so far could achieve. Before the signal, and measuring at every time later). The algorithm was tested at KNMI implementation at KNMI for automatic point how well the fit is, certain features on more than threehundred ‘local’ quakes S-wave detection, more tests on a large may be revealed which are missed (vibrations of gas fields in Groningen, otherwise. The wavelet method,

52 ERCIM News No. 41, April 2000 EVENTS

developed since the mid 1980s, is (average over the whole signal). In several applied not only in the analysis, but also particularly good at detecting sharp cases this works well, for example to the processing of signals (efficient storage transitions or multi-scale (fractal) detect the influence of ocean waves in the and transmission of sounds and images). structures in a signal. It was proposed by wave pattern of the North Sea, but local Present research at CWI concerns so- the French geophysicist Jean Morlet as features are usually missed. Wavelets also called nonlinear wavelets, with an alternative to the traditional Fourier have a certain frequency and amplitude, application to information retrieval from method, which describes a signal but because of their small extension they multimedia databases. (function) as a sum of sine and cosine are able to describe local features in a waves of a certain frequency and signal. Mathematically they are more Please contact: amplitude. Because of their infinite complicated than the simple sine or cosine Patrick Oonincx – CWI extension the fit of such waves with the functions, but by now the theory is well- Tel: +31 20 592 4177 signal only gives global information established. Wavelets are successfully E-mail: [email protected]

SOFSEM’99 by Jan Pavelka

After celebrating its silver anniversary in Slovakia, international conference with international invited SOFSEM came back to Milovy in Moravia, Czech speakers and international refereed submitted papers Republic. SOFSEM’99, held from 27 November to 3 (known as contributed talks). As in several preceding December 1999, was true to its new character and years, SOFSEM received support from ERCIM and mission, established five years ago: a full-fledged several other sponsors from the IT industry.

From its previous history SOFSEM has from the Netherlands. The UK, France, Jan Staudek and his experienced inherited its generalist and Spain, and Sweden had 3 participants organising team. Together, they have multidisciplinary character. This is each. The remaining 7 countries were produced another successful SOFSEM. attested to by the themes of ‘99: Trends represented by smaller numbers. in Theory, Core Technologies, Software The history of SOFSEM, which now and Information Engineering, From Data Serving as a meeting ground for includes fresh information from to Knowledge, and Applications. These professionals from both the theory and SOFSEM’99, can be found under the were covered by 19 invited talks and the practice of computing has always been general SOFSEM Web page at reinforced by selected submitted papers. an important ingredient of the SOFSEM http://www.ics.muni.cz/sofsem/sofsem.html. mission. I am glad I can report that also The PC (program committee) chair and SOFSEM’99 has served this purpose SOFSEM 2000 will be held probably at co-chair as well as the OC (organising well. Closing the gap between theory and the same location, again with ERCIM committee) chair were approved by the practice was also the topic of a panel heavily present in its organization. The SOFSEM Steering Committee in 1997. discussion held on the first day of the new PC Chair is Vaclav Hlavac from the As usual, the Steering Committee conference. Czech Technical University, Prague. recruited the invited speakers, leaving the Keith G Jeffery (CLRC RAL) and Jiri rest (the composition and co-ordination The busy time schedule of SOFSEM’99, Wiedermann (CRCIM) will serve as co- of the PC, the preparation of the time in which invited and contributed talks chairs. Jan Staudek will again be OC table, and the supervision during the were supplemented by commercial chair. For the on-line information on conference) to the PC chair and co-chair. presentations and a cultural program, still SOFSEM 2000 please see the above left time for lively discussions, which mentioned web page. The 23 PC members – 10 Czech, 3 often focused on sketching research Slovaks, and 10 researchers from 6 other projects with newly discovered partners. Links: countries – reviewed 46 submitted papers ht t p : / / w w w . i c s . m u n i . c z / s o f s e m / s o f s e m . h t m l coming from 19 countries and selected For the fifth time, the proceedings were 18 contributions to be presented at published by Springer-Verlag in the Please contact: SOFSEM. Lecture Notes in Computer Science (LNCS) Series (1725) and contain papers Jan Pavelka – DCIT, Ltd. PC Chair SOFSEM’99 SOFSEM’99 was attended by 138 from the invited and contributed talks. Tel: +420 2 3536 3342 persons with ERCIM institutes well E-mail: [email protected] represented. Although the majority of I am greatly indebted to my colleagues in participants were Czech (82) there were the Steering Committee, the PC co-chair 21 from Slovakia, 8 from Germany, 5 Gerard Tel, all PC members, the OC chair

ERCIM News No. 41, April 2000 53 EVENTS

CALL FOR PARTICIPATION SPONSORED BY ERCIM QofIS’2000 – Quality of Future The Ninth International World Internet Services Wide Web Conference ERCIM News is the in-house magazine of ERCIM. Published quarterly, the newsletter reports on joint actions of the ercim partners, Berlin, 25-26 September 2000 Amsterdam, 15-19 May 2000 and aims to reflect the contribution made by ercim to the European Community in The workshop focuses on end-to-end Join five world leaders in the 21st Century Information Technology. Through short articles services over QoS assured Internet, on Web community as they deliver keynote and news items, it provides a forum for the exchange of information between the institutes service creation, configuration and speeches at The Ninth International and also with the wider scientific community. deployment. World Wide Web Conference (WWW9), ERCIM News has a circulation of 7000 copies. the only Web conference where leaders Sponsors: T-Nova (Deutsche Telekom), from academia, research, government and ERCIM News online edition is available at ht t p : / / w w w . e r c i m . o r g / p u b l i c a t i o n / E R C I M _ N e w s / Cisco Systems, Nokia, IST Programme industry meet on a global stage. The of the European Union, GMD FOKUS conference will be held May 15-19, 2000, ERCIM News is published by ERCIM EEIG, at the Amsterdam RAI, in Amsterdam, BP 93, F-06902 Sophia-Antipolis Cedex For more information: The Netherlands, with Web-related Tel: +33 1 3963 5303, E-mail: [email protected] ht t p : / / w w w . f o k u s . g m d . d e / e v e n t s / q o f i s 2 0 0 0 / hardware, software, and services ISSN 0926-4981 E-mail: [email protected] Director: Jean-Eric Pin expositions to open May 16-18. Central Editor: Registration for the conference is Peter Kunz CALL FOR PARTICIPATION available online at http://www9.org. E-mail: [email protected] Tel: +33 1 3963 5040 Local Editors: Cross-Language Evaluation The keynote speakers include: CLRC: Martin Prime Forum • Egbert-Jan Sol, Vice President E-mail: [email protected] Technology at Ericsson: ‘Making the Tel: +44 1235 44 6555 CRCIM: Michal Haindl A Cross-Language System Evaluation Internet Mobile for Everybody’ E-mail: [email protected] activity is now being launched in Europe. • Paul de Bot, Vice President Business Tel: +420 2 6605 2350 The activity is sponsored by the DELOS Development at Philips Digital CWI: Henk Nieland Network of Excellence for Digital Networks: ‘The Convergence of E-mail: [email protected] Tel: +31 20 592 4092 Libraries (see page x of this number) in Internet and Television Technologies: CNR: Carol Peters collaboration with the US National How will it Change the Living Room E-mail: [email protected] Institute of Standards and Technology Experience?’ Tel: +39 050 315 2897 (NIST) and the TREC Conferences. The • Charles W. Davies, Group FORTH: Constantine Stephanidis E-mail: [email protected] Cross-Language Evaluation Forum Development Director at Psion PLC: Tel: +30 81 39 17 41 (CLEF) will run three main evaluation Wireless Information Devices and the GMD: Dietrich Stobik tracks in 2000, testing multilingual, Mobile Internet” E-mail: [email protected] bilingual and monolingual (non-English) • M. Graeber Jordan, CEO of Tel: +49 2241 14 2509 INRIA: Bernard Hidoine information retrieval systems. There will GJordan.com LLC: “The Spirit of the E-mail: [email protected] also be a special sub-task for domain- WWW in the Corporate Intranet” Tel: +33 1 3963 5484 specific cross-language evaluation. • Lawrence Lessig, Berkman Professor SICS: Kersti Hedman of Law at Harvard Law School: E-mail: [email protected] Tel: +46 8633 1508 The results of the activity will be ‘Cyberspace’s Constitution’ SINTEF:Truls Gjestland presented during a two-day Workshop on E-mail: Multilingual Information Access, 21-22 The conference will last five full days tr u l s . g j e s t l a n d @ i n f o r m a t i c s . s i n t e f . n o September in Lisbon, Portugal, with Tutorials and Workshops on day Tel: +47 73 59 26 45 SRCIM: Gabriela Andrejkova immediately after the fourth European one, Keynotes and tracks for Refereed E-mail: [email protected] Conference on Digital Libraries (ECDL Papers, Poster Sessions, Web and Tel: +421 95 622 1128 2000: see http://www.bn.pt/org/agenda/ Industry track, Web/Internet and Society SZTAKI: Erzsébet Csuhaj-Varjú ecdl2000). track, Culture track, Panels and Invited E-mail: [email protected] Tel: +36 1 209 6990 Speakers on days two through four, and VTT: Pia-Maria Linden-Linna Those intending to participate in CLEF a Developers’ Day on day five. The E-mail: [email protected] 2000 are requested to send an e-mail to conference is also a unique opportunity Tel: +358 0 456 4501 Carol Peters ([email protected]), as soon to lear about the latest developments on Free subscription as possible, indicating in which task(s) Web standards, due to the extensive W3C You can subscribe to ERCIM News free of they intend to participate. track, organized by the W3C itself. charge by: • sending e-mail to your local editor; posting For further information and deadlines, see: For more information: http://www9.org/ paper mail to the address http://www.iei.pi.cnr.it/DELOS/CLEF/ • filling out the form at the ERCIM website at ht t p : / / w w w . e r c i m . o r g /

54 ERCIM News No. 41, April 2000 IN BRIEF

SINTEF Telecom and on computer animation. He currently The CNR Informatics Institutes Informatics has signed a chairs the BSI Committee responsible in Pisa have moved from their contract with BAE-Systems t o for computer graphics and image separate sites in the historical city develop a system for natural processing standardisation. More centre to the new CNR Research Area communication in rough environments recently he has been involved with the (NACRE). The concept for this system World Wide Web Consortium (W3C) was presented on the ERCIM 10th helping to develop standards for the anniversary CD-Rom. BAE-Systems is Web and setting up W3C European the second largest company in Europe regional offices including the W3C UK within aviation, space technology and Office at RAL. Bob is Professor of defence systems with more than 90.000 Computer Science at Brunel University employees. The Norwegian Air Force is where he has taught since 1968. He also The new Pisa research campus. currently negotiating a contract for has an Honorary Doctorate from the renewal of fighter aircraft. The Technical University of Darmstadt, on the outskirts of town. CNUCE, IAT, NACRE-system will be implemented in awarded for his trend setting activities IEI, ILC and IMC, whose research the Eurofighter which is a possible in the field of computer graphics. Bob activities encompass practically all candidate for this contract. Hopgood is retiring from CLRC in sectors of IT and Computer Science, April 2000, he is taking up the position are now located on the same site. The SINTEF has celebrated its 50th of Head of Offices for W3C and will be new research campus will provide them a n n i v e r s a r y . The research based at CLRC – so he won’t escape with a number of important facilities, foundation SINTEF was founded in entirely! Bob will also continue with such as a Conference centre and 1950 by The Norwegian Institute of his work and positions at UK printing services, which previously Technology in order to strengthen the universities. were lacking. When the Pisan Area of ties between the university and Research is fully operational, it will be Norwegian industry and to promote Professor Brian Eyre has been constituted by 15 CNR research applied research. SINTEF operates in appointed to the new post of institutes working in various fields of close collaboration with the Norwegian Chairman of CLRC (CCLRC). E y r e the applied sciences (including Biology University of Science and Technology who has been a member of CCLRC’s and Medicine, Physics, Chemistry, (NTNU). The two institutions Environmental and Earth Sciences) cooperate in projects and share with approximately 600 research, laboratories and equipment. Together, technical and administrative staff. they form a center of expertise of high international standard. SINTEF also ERCIM organizes Information cooperates with the University of Oslo. Days on the European With about 1800 employees and an Information Society Technologies annual turnover of NOK 1,5 billion, Programme in Latin America w i t h SINTEF is one of Europe’s largest support from the European independent research organizations. Commission. The objective is to raise Contracts from the public and private awareness among the Latin American sectors provide 92% of the operating S&T actors about the opportunities revenue. related to the calls for proposals issued Brian Eyre, new Chairman of CLRC. by this programme. Detailed The work of four members of information is given on the strategy of CLRC’s staff were recognised council since 1998, is appointed for one the programme, the content of its in the 2000 New Year Honours list, year from 1 April 2000. Dr Gordon annual workplan, the available amongst them Dr Gordon Walker and Walker, the present Director for modalities for participation, and the Professor Bob Hopgood, both of whom Research & Development at CCLRC, coming ‘roadmap’ for its activities. were awarded the Order of the British has been appointed Chief Operating ERCIM will produce a report to Empire (OBE) by the Queen. Dr Officer and will be Acting Chief identify subjects of common interest Gordon Walker has worked at Executive from 1 April 2000 until a and recommend actions or mechanisms Rutherford Appleton Laboratory for 37 new Chief Executive is appointed for funding bodies in order to increase years and was recently appointed Chief through open competition. Dr Albert the R&D co-operation between Europe Operating Officer for CLRC, Westwood completes his term of office and Latin America in Information responsible for the day-to-day as Chairman and Chief Executive of Technologies. After a very successful operations of RAL and Daresbury CCLRC on 31 March 2000. In future Information Day 16 March in Mexico Laboratory, its sister laboratory in CCLRC will have an independent part City, two further Info Days will be held Cheshire. Professor Bob Hopgood has time Chairman and a full time Chief 4 April in Rio de Janeiro and later on in been involved in computer graphics Executive, bringing it into line with the Buenos Aires. For details, see since 1959 including pioneering work other six Research Councils in the UK. http://www.ercim.org/EU-LA/.

ERCIM News No. 41, April 2000 55 ERCIM – The European Research Consortium for Informatics and Mathematics is an organisation dedicated to the advancement of European research and development, in information technology and applied mathematics. Its national member institutions aim to foster collaborative work within the European research community and to increase co-operation with European industry.

Central Laboratory of the Research Councils Rutherford Appleton Laboratory, Chilton, Didcot, GB-Oxon OX11 0QX, United Kingdom Tel: +44 123582 1900, Fax: +44 1235 44 5385 http://www.cclrc.ac.uk/

Consiglio Nazionale delle Ricerche, IEI-CNR Via Alfieri, 1, I-56010 Ghezzano (Pisa) Italy Tel: +39 050 315 2878, Fax: +39 050 315 2810 http://www.iei.pi.cnr.it/

Czech Research Consortium for Informatics and Mathematics FI MU, Botanicka 68a, CZ-602 00 Brno, Czech Republic Tel: +420 2 688 4669, Fax: +420 2 688 4903 http://www.utia.cas.cz/CRCIM/home.html

Centrum voor Wiskunde en Informatica Kruislaan 413, NL-1098 SJ Amsterdam, The Netherlands Tel: +31205929333, Fax: +31 20 592 4199 http://www.cwi.nl/

Danish Consortium for Information Technology DANIT co/CIT, Aabogade 34, DK - 8200 Aarhus N, Denmark Tel: +45 8942 2440, Fax: +45 8942 2443 DANIT http://www.cit.dk/ERCIM/

Foundation for Research and Technology – Hellas Institute of Computer Science, P.O. Box 1385, GR-71110 Heraklion, Crete, Greece Tel: +30 81 39 16 00, Fax: +30 81 39 16 01 FORTH http://www.ics.forth.gr/

GMD – Forschungszentrum Informationstechnik GmbH Schloß Birlinghoven, D-53754 Sankt Augustin, Germany Tel: +49 2241 14 0, Fax: +49 2241 14 2889 http://www.gmd.de/

Institut National de Recherche en Informatique et en Automatique B.P. 105, F-78153 Le Chesnay, France Tel: +33 1 3963 5511, Fax: +33 1 3963 5330 http://www.inria.fr/

Swedish Institute, of Computer Science Box 1263, S-164 29 Kista, Sweden Tel: +46 8 633 1500, Fax: +46 8 751 7230 http://www.sics.se/

Swiss Association for Research in Information Technology Dept. Informatik, ETH-Zentrum, CH-8092 Zürich, Switzerland Tel: +41 1 632 72 41, Fax: +41 1 632 11 72 http://www.sarit.ch/

Stiftelsen for Industriell og Teknisk Forskning ved Norges Tekniske Høgskole SINTEF Telecom & Informatics, N-7034 Trondheim, Norway Tel :+47 73 59 30 00, Fax :+47 73 59 43 02 http://www.informatics.sintef.no/

Slovak Research Consortium for Informatics and Mathematics Dept.of Computer Science, Comenius University, Mlynska Dolina M SK-84215 Bratislava, Slowakia Tel: +421 7 726635, Fax: +421 7 727041 http://www.srcim.sk

Magyar Tudományos Akadémia – Számítástechnikai és Automatizálási Kutató Intézete P.O. Box 63, H-1518 Budapest, Hungary Tel: +36 1 4665644, Fax: + 36 1 466 7503 http://www.sztaki.hu/

Technical Research Centre of Finland VTT Information Technology, P.O. Box 1200, FIN-02044 VTT, Finland Tel:+358 9 456 6041, Fax :+358 9 456 6027 http://www.vtt.fi/