Number 2, 2008, July

ing experts from TERENA and CLARIN providers, but in particular with the TERENA – CLARIN together to discuss opportunities of a closer European initiative that wants to harmonize collaboration. the various attempts led us to meet the sec- CLARIN wants to establish a federation of retary general, Karel Vietsch, and Licia Collaboration language resources and technologies cover- Florio who is one of TERENA's technical ing institutes from many countries in experts. Europe and come to agreements with the Due to CLARIN's European scope, TERE- Peter Wittenburg, Daan Broeder, NA is a natural and expected ally. CLARIN Dieter Van Uytvanck emerging national identity federations that hopefully will cover researchers from an seems to be one of the first initiatives that he Trans-European Research and increasing number of European academic comes to TERENA with a large user com- TEducation Networking Association (TERENA, http://www.terena.org/ ) was set up to offer a forum to collaborate, innovate and share knowledge in order to foster the development of Internet technology, infra- structure and services to be used by the research and education community. One of the tasks of TERENA is to harmonize between the different national approaches in the area of distributed authentication and authorization. Different attribute sets to cat- egorize and describe different users are being used and, when similar, sometimes they are not used the same way. All these differences hamper cross-national federations, and there is a need for harmonization. This is why TERENA set up, for example, the Schema Harmonization Committee (SCHAC, http://www.terena.org/activities/tf-emc2/ schac.html ) to: (1) build an internal kernel from existing local attributes and agree on syntax and semantics and let the kernel evolve via a collaborative approach; (2) define a method to accept new attributes/classes; (3) to promote the schemas within the NRENs (National Research and Education Network) con- stituency, making the results part of the local The home page of TERENA at http://www.terena.org schemas; (4) promote the schemas in other fields, such as GEANT2 (European data communication network), digital libraries, institutions. The core of such federations are munity and a distributed service provider EUNIS (European University Information trust agreements where, in particular, issues scenario. Therefore, it was agreed that Systems), (US research network of user management and the exchange of CLARIN will explain its goals and wishes at initiative), etc. user attributes need to be specified. one of the coming expert meetings, and that In July a first official meeting was organized The understanding that CLARIN needs to TERENA and CLARIN will keep in further at the TERENA office in bring- interact not only with the national identity close contact. C We are glad now to bring to your attention The developer's view is presented professor Editors’ our second issue. You will notice that it Bojana Dalbelo Ba{i} from respects the same main schema, that we tried University. Foreword to impose as a constant for the CLARIN On page 5 Thierry Declerck discusses the Newsletter, mainly including minimally a topic which brings about the essential issue parallel view of developers and consumers, a of standards in LR&T that will be of crucial presentation of major events related to importance to the CLARIN community. CLARIN, and reports describing recent We have selected for review in the middle developments in the CLARIN member pages three events that took place recently in states. Europe. LREC – the Language Resources We have invited Peter Wittenburg to open and Evaluation Conference, where at least the issue with a note on a recently set up col- two meetings with great relevance for Marko Tadi} laboration between CLARIN took place. & Dan Cristea TERENA, another Call for contributions They are described by CLARIN Newsletter editors important collaborat- Steven Krawer. Then, ing project that builds Dear readers the Digital Hum- ear readers, research and educa- of the CLARIN Newsletter, anities Conference in Dfor many people who attended LREC tional infrastructures, if you have ideas, thoughts, Oulu, Finland is this year, the magnificent Marrakesh is still and CLARIN. The comments, additions, corrections, reviewed by Martin extremely alive in memories: with its narrow reason why we have arguments, questions etc. which are Wynne, and the ESF and crowded streets full of colours and chosen this as the connected to the CLARIN project, workshop on the role smells, with the snake charmers performing cover story is that we even remotely, please feel free to of humanities in CEE their skills in the open, with the shop own- wanted thus to stress send them to us as your contribution countries in Sofia, ers almost grabbing you to visit their incred- the importance of the at [email protected]. Bulgaria, presented ible bazaar-like exhibitions, with the bar- interactivity with by Marko Tadi}. gaining habits reminding that we are already major pan-European initiatives intended to Peter Wittenburg and Tamás Váradi are in the middle of Orient although at the develop technological infrastructures that interviewed by us in an attempt to shed Greenwich longitude, with the arabesque of will be of help to researchers in the social sci- more light on the lively issue of the role of sculptured wood and stone heavily decorat- ences and humanities. commercial companies in LR&T research in ing the Muslim palaces, with the romantic Next to this page you can read in an article Europe, a topic which they addressed in an riads, the authentic old-style hotels built by Maria Gavrilidou and Stelios Piperidis article in the previous issue of this newslet- around inner patios where the tea arrives in about the efforts that are being performed ter. We felt that we should devote more the glass after executing a skilful and spec- presently in Greece to preserve the cultural space to the question of wheteher we are tacular vault in the air. heritage and the role that the language really in competition with the Internet and It was there that we decided to distribute a resources and technologies play there. software giants. historical, if you allow us to prefigure, first The next page, bring the parallel views of The following articles describe two CLAR- issue of our CLARIN Newsletter, as we consumer and the developer. This time IN national projects: the German one, as wanted to exploit the advantage of having @eljko Hodonj from the Croatian News authored by Erhard Hinrichs, Peter there more than a thousand of participants Agency has the contribution which describes Wittenburg, Alexander Geyken, Lothar working in Language Resources and how this company plans to use LR&T to Lemnitzer and Andreas Witt, and the Technologies for putting it directly in their further speed up their work-flow and also to Danish one, as authored by Hanne Fersøe. hands. make new information-broker products. Enjoy your reading! C

List of national correspondents Estonia Malta Tiit Roosmaa Mike Rosner Austria Finland Netherlands Gerhard Budin Kimmo Koskenniemi Peter Wittenburg Belgium – Flanders France Norway Inneke Schuurman William Del Mancino Koenraad De Smedt Bulgaria Bertrand Gaiffe Svetla Koeva Portugal Germany Antonio Branco Croatia Lothar Lemnitzer Poland Marko Tadi} Greece Maciej Piasecki Czech Republic Maria Gavrilidou Karel Pala Stelios Piperidis Romania Dan Cristea Denmark Hungary Hanne Fersøe Tamás Váradi Dan Tufis, ELRE/ELDA Italy Spain Bente Maegaard Valeria Quochi Nuria Bel UK Latvia Sweden Martin Wynne Andrejs Vasiljevs Sven Strömqvist

2 No 2 / Newsletter CLARIN of the organisations, including recommenda- dations, the material (basically the documen- Language tions and standards for multilingual content tation material, but in certain cases also pri- authoring and retrieval. mary data) had to be available in many lan- technology in the guages. For the translation of the material, Focus on LT translation memories and multilingual termi- nologies have been used, where they already Greek cultural Not unexpectedly, the studies concerning tex- existed, and created, when needed. For the tual archives and associated metadata focused translation aids, the use of TMX and TBX was on Language Technology methodologies, stan- enforced, catering for compatibility with stan- heritage sector dards, tools, applications and integrated sys- dards and ensuring exchange and reuse. tems, as regards all stages of creation, process- ing and maintenance of the digital content. Interoperability Consequently, NLP and LRs know-how has been sought after by the cultural content hold- As for interoperability with similar efforts in ers, and it has been successfully deployed in the cultural domain, the recommended model Maria Gavrilidou, the framework of this programme. was the CIDOC Conceptual Reference Model Stelios Piperidis Most of these cultural organisations have (CRM), which provides definitions and a for- ILSP – Athena RC sought support in indexing their digitised pri- mal structure for describing the implicit and mary or metadata material in an as intelligent explicit concepts and relationships used in cul- n Greece the largest programme related to manner as possible. For instance, shallow tural heritage documentation. Aiming to serve Ithe Social Sciences and the Humanities is semantic information extractors have been as a guide for good practice of conceptual the Operational Programme Information Society, used like term and keyword extractors, named modelling, it is now official standard ISO in the framework of which, more than 140 entity recognisers, etc., to automatically gener- 21127:2006 “intended to promote a shared projects are currently being completed. These ate meaningful index terms. Such extracted understanding of cultural heritage information projects aim at the digitisation, documenta- indexical data have in many cases been struc- by providing a common and extensible seman- tion and promotion of the Greek cultural her- tured in thesauri following ISO standards tic framework that any cultural heritage infor- itage, including all aspects of culture and civil- 2788 and 5964 for monolingual and multilin- mation can be mapped to.” isation. In this sense, museum objects, textual gual thesauri. Furthermore, and in an attempt Currently, and in the framework of a Call for as well as audiovisual archival items, collec- to provide unified, single-entry access to topi- Tenders, the Greek Ministry of Culture has tions and archives of great historical value, the- cally similar or related collections, meta-the- launched the implementation of a special atre archives, folklore museums' material, music and painting collections, ecclesiastical art collections, costume collections etc. have been digitised, documented and made avail- able to the public. The goal of this huge (for the country) endeavour was the preservation of the cultural heritage with the aid of infor- mation technologies, but also the exploitation of the cultural material, the development of related digital by-products and services, and last but not least, raising awareness about dig- ital content among the citizens of the country. To this Call many cultural material holders responded, such as museums, libraries, archives, cultural organisations, etc.

Reccomendations and best practices The use of standards and intelligent techniques at all levels of development has been explicitly stressed throughout the Programme. In the Preparatory Phase of the Programme several studies were completed which elaborated rec- ommendations as regards standards and best practices to be used. The studies concerned: technologies for the digitisation of 2-D mater- ial (text, image), 3-D material (objects, monu- ments, archaeological sites), video and moving Homer’s Illiad, VIII, 245-253, codex F205, late 5ct or early 6ct image, sound and music; technologies, stan- dards and metadata for the interoperability, saurus structures have been implemented for mediator service through which the Greek documentation, re-usability and management those collections for which indexing had been Cultural Assets which have been digitised and of digital content; technologies for the protec- performed on the basis of specific thesauri. documented so far are semantically unified to tion and management of IPR issues, and final- Multilingual access to the data was another key enable single point access to the wealth of ly, for the development of portals and websites issue in this framework: following recommen- these vast information sources. C

CLARIN Newsletter / No 2 3 Users’ view Developers’ view HINA – digitising and Which LR&T are available exploiting the news archives for supporting HINA?

@eljko Hodonj Bojana Dalbelo Ba{i} Development consultant in HINA University of Zagreb, Faculty of Electrical Engineering and Computing

n 2008 the Croatian News Agency (HINA) became the focal he tender for providing the services of semantic analysis of the Ipoint for the development of digital archives in Croatia. Only a TCroatian newspaper texts for Croatian News Agency (HINA) year earlier the Croatian parliament entrusted HINA with the pro- had put a serious challenge before us. Since the tasks were many tection of (news media) archives containing some 14 million docu- and covered different areas of expertise (language resources and ments originating from two databases, the archives of newspapers technologies, together with knowledge technologies and their com- articles in the Vjesnik publishing house (VND records) and the EVA putational realisation), Faculty of Humanities and Social Sciences database, created by HINA. These databases have the status of (Department of Linguistics) and Faculty of Electrical Engineering protected archive collections of Croatian culture. and Computing (Department of Electronics, Microelectronics, Being a news agency, HINA's prime mission is to provide credible Computer and Intelligent Systems) submitted a joint offer to HINA. The architecture of the system that we had to provide solutions for and unbiased information about events and people as soon as pos- was modular so different modules should be used for different sim- sible and deliver it to its subscribers. As a part of its business, HINA ple tasks and that more complicated task could be achieved by has planned the development of a multimedia database aimed to their recombination or usage in different workflows. The tasks that create new services for its subscribers. It has prearranged its work- we have to deal with are ranging from the simplest to the more flow for creating new business packages through media monitoring complicated ones. services. The project now includes all relevant printed news media sources in Croatia and all news production created by HINA. The basis of the system represent the LR&T tasks such as lemmati- Simultaneously, we are going to begin with digitization of the VND zation and POS/MSD tagging of Croatian texts which has only records, working backwards from the most recent ones. been done so far in academic and experimental environments and This is a simple, yet a demanding approach, to link all articles from was never tested in real, industrial strenght conditions where it has recent printed media and articles created in the past, because it is to achieve the highest possible rate of accuracy without sacrificing the only model that would enable a true understanding of the con- the speed, which can be crucial in newswire text flows. For this pur- tent and detecting the trends; a dynamic connection between time pose a new hybrid tagger for Croatian was developed which and topics, individuals and events. reached around 97% accuracy. The module for lemmatization and In this context a new general objective has also been defined; to support information access to all Croatian citizens and enable them full morphosyntactic tagging will also be used later for query pro- access to media information at the lowest cost possible. cessing during the user access to textual databases. The digital archive as an area of organized knowledge has been The next task, which builds on the previous one, is named entity designed as a system that connects information and events, people recognition and classification. The system for NERC in Croatian and their activities; a system which supports texts was developed three years ago (with an the organization of knowledge concerning f-measure of around 90%) but also as a the connection between people and events, Editors’ note research prototype with only few tests in real people and people, event and event – as a On this page we publish opinions, discus- business applications. At this first stage only system which recognizes and builds up a net- sions, view and arguments that will usually work of relations and correlations between come from two sides. One will illustrate the names of persons, organizations and loca- topics and the time of their occurrence. The standpoint of CLARIN users or “con- tions will be recognized and classified preconditions for achieving this objective and sumers”, while the other will try to present according to the existing gazetteers, while all of its segments was to find a partner capa- the ideas that are coming from the direc- later this initial classification will be used for ble of supporting the project with expert tion of LRT developers. more elaborated schemata. knowledge. The module for document classification has In April 2008, HINA opened a public tender to select the best provider for the services of semantic analysis of to deal with the daily flow of news documents that measures in tens the text in a modular fashion. A joint offer submitted by Faculty of of thousands. It will be able to classify new documents to a prede- Electric Engineering and Computing and the Faculty of Humanities fined classifying scheme. The initial tests without the thorough train- and Social Sciences in Zagreb was selected in this tender as the ing of the classifiers yielded 86% f-measure which was considered best provider. By late 2008, the laboratories of the two faculties will excellent. provide prototypes: a module for lemmatization and morphosyn- The module for keyword detection is yet to be defined in detail but tactic tagging of Croatian words, a module for recognizing and classifying named entities in the Croatian language, a module for the overall idea is to submit each document to statistical process- automatic document classification, a module for automatic detec- ing that will measure statistics of individual single- and/or multi- tion of key words from documents in Croatian, a shell that enables word units within the parts of the document to the whole document access to each module via a web service and a system for moni- or to the whole document collection. The significance of keywords toring the work of all modules. will be measured statistically and positionally (i.e. within the docu- By the end of April 2009, all versions of the system will be tested ment). thus bringing HINA to the first level of the realization of its objec- tives. The module for monitoring should provide opportunity to adapt the This approach illustrates how HINA understands relations between functioning of other modules (i.e. adding new named entities in pragmatic necessities of its business interests and the protection gazetteers, retraining the classifiers, error detection/correction etc.). cultural heritage and legacy knowledge, how it contributes to the This basic system will allow further development of more compli- protection of cultural entities and European values. cated knowledge-oriented services such as social networking. C

4 No 2 / Newsletter CLARIN “Language resource management”, created optional stage in the standardization process. The ISO standards in 2002 in cooperation with ELRA, the Normally, the FDIS stage is only reached if European Language Resource Association, changes to the DIS document before publi- for LR&T that are which is also a partner of CLARIN, and a cation as IS are necessary. main organiser of the LREC conference. 6. IS: International Standard. After being The general overview of TC 37 is now: approved, a (F)DIS finally becomes an being developed International Standard under ISO copy- – ISO/TC 37/SC 1 “Principles and meth- right. It thus becomes an accepted docu- ods” ment, which provides a reliable basis for – ISO/TC 37/SC 2 “Terminography and implementations and harmonised solutions. lexicography” Currently, the standards being discussed in – ISO/TC 37/SC 3 “Computer applications TC 37/SC 4 concern the markup of lexical for terminology”. resources and the annotation of linguistic Thierry Declerck – ISO/TC 37/SC 4 “Language resource data at various language processing levels DFKI GmbH management” (morpho-syntax, syntax and semantics). For

he production, processing, use and re- Tuse of multilingual linguistic data consti- tute a time-consuming and costly part of the daily work in the language industry. There is a critical need for established standards to enable the interoperability and re-use of multilingual data and linguistic processing tools in order to facilitate processes such as harmonisation, localisation, machine trans- lation and cross-lingual information re- trieval. The need to provide industry-validated stan- dards for language resource management has been recognized within the ISO (In- ternational Organization for Standard- ization) community for a long time. A spe- cial ISO Technical Committee was estab- lished more than 50 years ago, dedicated mostly to terminology and terminological resources. In a review of the first 50 years work of this committee, it is declared that: “…terminol- Visit the home page of TC37/SC4 at http://www.tc37sc4.org for more information on LR&T ISO standards that are being developed ogy plays a crucial role wherever and when- ever specialized information and knowledge There are six development levels for ISO this purpose a general framework (Linguistic is being prepared (e.g. in research and devel- standards, described below, based on the Annotation Framework, LAF) defining the opment), used (e.g. in specialized texts), definitions in the glossary of the ProSTEP representation of these phenomena is in recorded and processed (e.g. in data banks), iViP Association:2 development, with the purpose to enable passed on (via training and teaching), imple- 1. NWI: New Work Item. This document interoperability between the various types of mented (e.g. in technology and knowledge markup and annotation information contains the proposed scope and contents of 2 transfer), or translated and interpreted. In a newly suggested topic for standardization. involved. the age of globalization the need for If there is sufficient interest (of at least five Many partners of CLARIN can and will methodology standards concerning multi- nations i.e. national representatives) in this contribute to those standards, in areas such lingual digital content is increasing ”.1 subject, the document will be advanced to a as the citation of resources, interoperability of language data and tools. Recently, in 2001, the name of the Com- working draft. mittee has been changed into “Terminology 2. WD: Working Draft. The Working Draft But for CLARIN it is necessary that the LR and other language resources”, and this step document is derived from the NWI and part of the whole infrastructure stick as was crucial for many partners involved in contains the essential technical contents of much as possible to the existing standards. CLARIN since it opened the standardisation the future standard. This includes the coor- Otherwise it would be much harder to C work to all academic and industrial activities dination of different technical approaches. achieve our general goal. dealing with language in one form or an- 3. CD: Committee Draft. This document is other. the first version of a part of the future stan- 1 See http://www.infoterm.info/pdf/activities/ So, for example, the development of linguis- dard, which is internationally voted. Standing_document_02_50_years_ISO_TC_3 tically annotated corpora or computational 4. DIS: Draft International Standard. This 7.pdf lexicons can now be accompanied by ISO level emerges from the CD stage through 2 See http://www. prostep.org/nc/en/metanav standardization efforts. agreed changes and additions. /glossary.html Four Subcommittees of TC37 have been 5. FDIS: Final Draft International Standard. 3 See also http://www.tc37sc4.org/ for more established, the latest one being TC 37/ SC4 The Final Draft International Standard is an information.

CLARIN Newsletter / No 2 5 endangered and minority languages research to industrial development. Peter CLARIN at should play in this process? Wittenburg (Max Planck Institute 2) How can language resources and tools Nijmegen) discussed the typical pillars of the best be made available and made accessi- technical CLARIN infrastructure and in my LREC 2008 ble to the research communities at large? own presentation I identified a number of For example, resource providers may be fear- possible ways for CLARIN to fail. I don't ful of what happens to “their” resources and think that we managed to solve any of the tools. Will they be asked to hand them over major problems during the discussion that to some big “data bureaucracy”? How can followed, but it was interesting to see that Steven Krauwer accessibility and IPO issues be reconciled? there was a keen interest from the audience in CLARIN and in infrastructure issues in CLARIN coordinator What technical solutions are available for all of this? In this panel session, moderated by general. The slides of the panel presentations and of the presentation at the main confer- or those of you who attended LREC Erhard Hinrichs (University of Tübingen), five panellists highlighted different aspects. ence are available on the CLARIN website at F2008 in Marrakech it was hard to miss http://www.clarin. eu. CLARIN. No less than 12 papers at the Sadaoki Furui (Tokyo Institute of main conference contained references to Technology) gave an overview of what was On the whole I must say that many of the CLARIN, even if the project itself had only happening in Japan, especially in the pro- sessions and discussions in the main LREC recently started, well after the submission gram “Framework for Systematization and programme and in the workshop pro-

The CLARIN LREC2008 meeting

gramme showed that infrastructures are a deadline for papers for the main conference. Application of Large-scale Knowledge hot topic, not just in Europe but all over the For papers presenting results emerging from resources”. Sebastian Drude (Museu Goeldi the CLARIN project we will of course have Belem / Freie Universität Berlin) focused on world. This calls for intensive international to wait until the next LREC conference in digital language archives for minority and collaboration in order to make sure that we 2010. endangered languages and gave an overview all move in the same direction and, more importantly, that we move towards common In Marrakech the focus was on our plans of the major language documentation pro- standards to ensure maximal interoperabi- (and concerns) for the future, on mobilizing grams and the special characteristics of lity. During the conference we also managed the community and on identifying opportu- archives for endangered languages. Nicoletta Calzolari (CNR-ILC Pisa) looked at the lan- to organise an internal meeting for CLARIN nities and parties in humanities and social consortium partners attending the confer- sciences for future collaboration. guage resources landscape from the FlaReNet perspective, a new EC project, ence. As our project budget for travel is lim- Apart from a general presentation of the very closely related to CLARIN, that aims at ited we will try to organise CLARIN meet- CLARIN project at the main conference, providing recommendations for future lan- ings in the margin of some of our major con- which shouldn't contain too much new guage resources policies in Europe (see ferences and we hope that at the next LREC information for most of our Newsletter read- http://www.ilc.cnr.it/viewpage.php/sez=ricer conference there will be room for us to ers, we had a special panel session dedicated ca/id=904/vers=ita). The interesting differ- organise another meeting, but then not just to key issues in building an infrastructure for ence between CLARIN and FLaReNet is for consortium partners but for everybody language resources and tools. These issues that where CLARIN's target audience is the actively involved in CLARIN, both in the included the following social sciences and humanities research com- EC project and in the accompanying nation- 1) What is the scope of building a common munity at large, FLaReNet covers the whole al projects launched to prepare the construc- language resources and tools infrastruc- spectrum of parties interested in language tion of the CLARIN infrastructure at the ture and what is the particular role that resources and technology, ranging from basic national level. C

6 No 2 / Newsletter CLARIN of digital humanities centres, originating in DH2008 is effective as a gathering of a Digital North America, but now with participation group of scholars who are profitably sharing from around the globe. A panel session on ideas, techniques, resources and tools across the opening morning discussed aspects of traditional discipline boundaries. It was a Humanities 2008 'Defining an international humanities por- subject of ongoing debate among scholars at tal' in the context of CenterNet. It soon this conference whether they represent more Oulu, Finland, became clear however that what was needed than this, and whether they should see was not another portal to compete for space 'humanities computing' (or nowadays more June 25-28th on our desktops with those of our faculty, often 'digital humanities') as an academic institution, national community, academic discipline in its own right. Whatever the discipines, etc. What we need is to present answers to this question it is clear that it will our resources and tools as standards-confor- be very useful for CLARIN to engage with mant services so that they can be integrated the discussion. Martin Wynne into the researcher's environment. In the final plenary session, invited speaker CLARIN EB member Sylvia Adamson of the University of Also emanating from this community, and Sheffield demonstrated how the historical worthy of attention from CLARIN mem- linguistic research in which she is engaged igital Humanities is the joint interna- bers, are TaporWare, a set of online text would benefit from better electronic Dtional conference of the Association for analysis tools3, Project Bamboo, an initiative resources and tools, and posed a challenge to Literary and Linguistic Computing and the to develop new shared technology services “the geeks in the audience” to develop a Association for Computers and the for the Humanities4 and Heurist, a free more effective working environment for Humanities, and has been taking place online database service developed for schol- scholars like her. This is precisely the chal- annually since 1989, since 2006 under the ars in the Humanities5. In fact it has become 1 lenge that CLARIN has accepted. C banner of 'Digital Humanities' . It claims to clear that the priority is not for more tools, be the oldest established meeting of scholars more portals, or more initiatives, but a way working at the intersection of advanced of connecting together the existing frag- 1 information technologies and the humani- DH2008: http://www.ekl.oulu.fi/dh2008/ mented landscape into a ecosystem where ties. This year the conference was held at the 2 creators of resources, tools and services can CenterNet: http://www.digitalhumani- University of Oulu in Finland, and Tamás do so in such a way that they can interoper- ties.org/centernet/ Varadi and Martin Wynne presented a poster ate with the other resources relevant to their 3 outlining what CLARIN was all about. TaporWare: http://taporware.mcmaster.ca/ users. The CLARIN approach is to promote 4 Many presentations at DH2008 were of rel- this interoperability and to faciliate their Project Bamboo: http://projectbamboo. evance to CLARIN. These included org/ 2 deployment within a sustainable infrastruc- CenterNet, an initiative to build a network ture. 5 Heurist: http://heuristscholar.org/

necessary measures to further strengthen the – identification of pilot initiatives; ESF meeting on participation of CEE scholarship in the – involvement of Institutes of Advanced European Research Area. The workshop Studies and university-based centers of theme and philosophy was “Harnessing the excellence; humanities in Assets”; it was understood that efforts need – dialogue with neighboring sciences (e.g.: to be made to better promote the strengths of social sciences; environmental sciences); CEE Humanities scholars as research part- CEE countries ners. While different presentations also high- – consortium building (third-party-funding); lighted structural obstacles that still exist in – role of ESF member organizations in pro- this process, it was clearly demonstrated that moting CEE academic assets. Marko Tadi} the rhetoric of “catching up” has to be aban- The role of a research infrastructure for Editor doned since there is nothing to catch up Humanities was strongly put forward and in he special European Science Foundation between CEE countries and Western Europe this respect CLARIN gained the deserved Ttopic-oriented worksop organized by the countries. The CEE countries are simply in attention not just by its presentation but also Standing Committee for the Humanities some respect different and this difference has in discussions during the whole duration of (SCH) was held between 26th and 28th June to have its reflections in Humanities (and the workshop. The idea of LR&T as support- 2008 in Sofia. Social Sciences) research. The historical ing technology for text(-based) sciences that Since the exact title of the meeting was starting points are not the same, the financial would allow the development of new Stakeholder Workshop “Central and Eastern support is not the same, even the position of Humanities paradigm (e-Humanities) was European Scholarship in the Humanities – Humanities research is not the same in CEE appealing to the majority of participants harnessing the assets”, one could ask and Western Europe countries. This differ- while some of them still showed some reserve him/herself was it relevant for CLARIN in any ences were discussed. The workshop also to its final shape. The scepticism was not only respect. It certainly was since the CLARIN addressed the issues of international compa- based on the problems of technical nature research infrastructure FP7 project was pre- rability of Humanities scholarship in smaller (i.e. organizing the federation of archives, sented in an invited talk by Marko Tadi}. linguistic communities, and the related chal- succesful application of grid technology, The meeting assembled some 50 invited par- lenges of internationalization and globaliza- open and permanent access to research data ticipants from 12 countries and ESF officials tion. Nevertheless, the workshop was looking and results etc.) but was also based on the – representing some of the most relevant for an agreement on a joint strategy to move problem of empirical vs. rationalistic institutions in Central and Eastern European forward, comprising as possible elements of approach in Humanities research in general countries among the ESF membership EE, LT, study: (e.g. Chomsky’s philosophical standpoints) LV, PL, CS, SK, HU, HR, SI, RO, BG and – commissioned surveys (by country and/or which is a well known dichotomy that cer- important partners elsewhere – to debate the by discipline); tainly will not be solved soon. C

CLARIN Newsletter / No 2 7 we already tried to contact the big compa- amount of effort and money people are will- CLARIN’s role in nies to establish collaborations. This com- ing to spend. plementary nature has been pointed out by CNL: Do we understand correctly that you open and the British JISC (Joint Information Systems don't see any technical motivations for a dis- Committee) initiative already. However, tributed CLARIN future? permanent access everything will very much depend on how W&V: No, of course there are also technical the big companies will position themselves reasons: let us mention for example the need and on the policies they will follow: whether for distributed storage to ensure data sur- to research results they decide to pursue a more collaborative vival and distributed services to guarantee or a more ignorant policy. If CLARIN, for high availability and performance. Let us and data example, would be allowed to link up to the point out that the Amazoogles also need a emerging huge digital resource base of the distributed approach to storage and services companies, then this would be beneficial for for similar reasons. However, that does not the user community and certainly we would commit them to a model of distributed deci- not hesitate to allow them access to our sion making. There is still a single decision resources. What would be important is that taking board at the top of the hierarchy and Peter Wittenburg there are proper agreements that are benefi- this board can decide about optimal solu- & Tamás Váradi cial for all sides. CLARIN EB members tions to make good business. The situation is CNL: Aren't the Amazoogle services the different with CLARIN from this respect. best answer to overcome the huge fragmen- We are committed to other aspects than he topic that was introduced in the pre- tation we are suffering from in the humani- “business success” such as geographic distri- vious issue of the CLARIN Newsletter T ties? After all, as history has shown, they bution and balance of other kinds, which (CNL) raised quite a few questions. There- could ultimately enforce a standardization also means that no one should expect that fore CNL has organized an interview with with respect to the complex rights and for- we can be as cost efficient as the Amazoogles Peter Wittenburg and Tamás Váradi in order mats issues due their monopoly role. Wasn't are trying to be. to clarify and elaborate further their stand- it Microsoft who pushed us forward to a CNL: When you say that you could imag- point on the role of LR&T and its availabil- quasi standard at operating system level? ity for researchers and society in general. ine an atmosphere of collaboration, do you W&V: It could indeed happen that CLAR- have an example in mind? IN does not have sufficient power in our W&V: Yes, for sure. The publishers, for CNL: In your article in the previous CNL community to convince researchers to make example, have taken over the job to produce issue you sketched that there will be a com- use of certain standards and to simplify the and distribute books representing much of petition between some big companies, cur- rights situation. In that case the Amazoogles our cultural heritage. But, in addition, each rently abbreviated with the buzzword could succeed. They could simply rely on country has a national archive and a nation- Amazoogle, on the one hand and infrastruc- the fact that everyone who wants to be seen al library. Their role is beyond the one of ture initiatives, such as CLARIN, on the in the web needs to adhere to the constraints publishers. They need to take care of com- other hand. Why are you sceptical about or practices imposed by them. The pressure pleteness, archiving, curation, etc. and for their services? could become high for each researcher so these aspects there is no self-supporting W&V: Our skeptical attitude emerged from that all concerns are set aside. Basically this business model, but a national interest. With the experiences with the big publishers who implies that CLARIN also has to push its respect to the digitized heritage, including broke the “old” deal with the research com- own community in the direction of harmo- language resources, this necessity of splitting munities and began to see scientific content nization and simplification. On the other roles is even more important. Still we have a primarily as a market and not as a service. hand we need to remain sensitive enough to highly dynamic situation where the various Google's European vice-president stated cope with special requirements. roles and business models are not yet settled again very clearly that their primary concern CNL: So no reason for CLARIN, DARI- but are developing and influencing each is business, of course, and we should not be AH, ERIH, etc. to come up with an alterna- other and also shaping the landscape. naive in this respect. Prices are primarily not tive model? CNL: Could you imagine already now an dependent on the production costs, but on W&V: For sure we need to learn from the aspect of work where you think that the the market opportunities, i.e. the research Amazoogles in many ways. Again we count roles will be different and therefore comple- community needs to be careful not to find on a collaborative attitude. One of the mentary? themselves again in a situation where they aspects where we need to learn is how to effi- W&V: The Amzoogles are companies that need to pay too high prices to access the ciently establish and run a network of LRT content which they produce themselves. Too come and go, in particular, nowadays, while centres. For several reasons, centralization the lifetime of states or cultural regions in strong dependencies without alternatives could be very beneficial – think only of the don't seem to work. general seems to be much longer. States and potential saving in energy costs to run big regions have an interest in taking care of CNL: Isn't there also a chance of collabora- servers. But there are so many reasons that long-term preservation of their heritage and tion between complementary services? require a more distributed approach. Think culture. So, with respect to language W&V: Of course, we could imagine a com- of the relevance of “national” centres as resources, it will be the task of the linguistic plementary situation between the being responsible for “their” languages. We community, in collaboration with the states Amazoogles and, for example, CLARIN and know how important identification is for the or regions, to take care of their long-term

8 No 2 / Newsletter CLARIN preservation. The Amazoogles can't give business, i.e. when there is an option to CNL: It is a known fact that the Amazoo- guarantees beyond their business models. make money, a company would do so. So gles are making money mainly exploiting CNL: Are there other aspects where you see even if storage is free, it may and will be the advertisements – so this is another reason differences? case that access in whatever form will be not to change their business model. W&V: Good and efficient business will have associated with fees and as we said earlier: W&V: Well, we already discussed this problems in taking care of all sort of special monopolies tend to base their prices not on aspect. After all, it is business. Who can pre- needs such as minority languages, special the actual costs but on what the market can dict the precise circumstances for making formats, special rights, etc. These areas where give. And the Amazoogles are monopolies money in 5 years? special knowledge is required, where the that could define prices. CNL: So, you remain sceptical. number of users will be limited – usually CNL: So you think that the cost effective- W&V: Yes. We are all using Amazoogle ser- specialists – and where special concerns and ness of their services will not lead to low ser- vices and we like them. But let's be careful regulations will increase the costs. Take as an vice prices? with the scientific content and not make example the amount of curation that is nec- essary to bring existing lexicons into a gener- ic format that will allow even laymen to determine cognate sets, etc. It requires deep knowledge about the language, the encod- ings used and the structural characteristics to carry out useful transformations and this would only be useful for a small community of potential users in education and research. We don't see a business model for this. CNL: In all your argumentation you are indicating that there could be much gain in collaboration. Why did you put the “com- petition” aspect in the foreground of your argumentation in your statements? W&V: Indeed a good question. Of course, on our side the competition is a challenge to work at efficient cost levels and to reduce complexities where possible. The other part has to deal with sensitivity of various sorts. Funding agencies may easily come to the Could such impossible combinations like Amazon and Google taken together live at all? Gryphon illustration by Sir John Tenniel for Lewis Carroll's conclusion that the Amazoogles will take Alice in Wonderland care of all aspects of language resources and technology, so that they don't have to care. W&V: Correct – the cost effectiveness will research data and research results dependent We wanted to stress that this is not true. In only be offered as low prices to the cus- on the business models of big monopolies. contrast, there is a danger that funders may tomers if there is real competition. So, here Let the research communities determine want to reduce funds due to the “cheap” is an argument for having initiatives such as how they want to access and use scientific offers of the Amazoogles. Many researchers CLARIN to create a competitive domain. data. are charmed by the simplicity of the CNL: Some say that if the Amazoogles CNL: What if the Amazoogles established Amazoogle user interfaces. However, for would change their cost model, immediately scientific advisory boards where researchers deep linguistic or humanities research the commercial competitors would show up – so would be invited to determine the rules of options are not sufficient. And, as we indi- they cannot change their prices if they do the game at least for scientific usage? cated, the financial questions remain. not want to fail. W&V: If it were to be guaranteed that even CNL: OK let's discuss the correctness of W&V: As long as there is competition, as their Executive Boards would accept the your expectations in this respect. Offering you stated it. So research infrastructure ini- advice of the researchers, it would be a big reliable services will cost some money who- tiatives could play a role in this game. But step but unfortunately this is not the way ever is giving them. Why should a company the web opens up a new phase in centraliza- how business is functioning. not ask for some usage or access fee? tion, as the Amazoogles are demonstrating. CNL: A final statement please. W&V: That's absolutely correct. Whoever They offer excellent services and the whole W&V: Everybody agrees now that it is good will give services, they will cost some money. world is adapting to their style and/or con- to have the Amazoogles as they keep driving It is completely ok that a company asks a tent of services. Other service providers are us ahead in many ways. But we need certain price for a certain service. Currently reduced to marginalities or, if you want, research infrastructures such as CLARIN as the Amazoogle offer to store data, for exam- only a few will remain. If the major market well to support research, to take care of spe- ple, free of charge, they currently only house players agree on terms, we will be dependent cialties and establish at least some form of data that is open access and they have their on their services and it will become almost competition. license terms, so depositors have to adhere to next to impossible for newcomers to enter There is also much space for collaboration, these very simple rules. But, come on, it is the market. as well. C

CLARIN Newsletter / No 2 9 the ESFRI projects among all the relevant by starting a structured approach to a Danish The Danish and interested parties in Denmark, and it is BLARK. assumed that CLARIN must have received The Danish CLARIN project will follow CLARIN project good support; the Centre for Language standards and recommendations developed Technology is a well-known player in the in the preparatory phase of the European field; the technical as well as the CLARIN project, but it management part of the propos- is important to realize al were well accepted by the that this project has not Hanne Fersøe reviewers; the EU CLARIN pro- been granted as a Centre for Language Technology, ject had been invited to contract University of preparatory phase pro- negotiations by the time the full ject in parallel with the proposal was submitted. long with other governments, the European project. It ADanish government has decided to The partners include eight lead- involves an independent invest in research infrastructures for all fields, ing Danish humanities institu- Danish investment in including the humanities and social sciences tions: four universities, a univer- the construction of a as a part of the Danish Globalization sity library, a museum, and two national infrastructure Strategy. The infrastructure initiative has a government research institu- that will stand alone as a total budget of 600 million DKK (appr. 80 tions. The ten consortium mem- vital contribution to the million Euro) which will be distributed with bers are: Danish research enter- 200 million per year over 3 years, 2008- University of Copenhagen with prise. For this reason it 2010. Funding is not limited to projects three departments from the was vitally important which are included in the ESFRI Roadmap, Faculty of Humanities: for the consortium to design and plan the so infrastructure projects of all kinds may 1) Centre for Language Technology – co- activities in such a way as to be able to deliv- apply for funding. ordinator of the project, er not only the technical infrastructure as a A consortium headed by the University of 2) Danish National Research Foundation result, but also as many types of content as Copenhagen has been given a three year Centre for Language Change in Real possible. grant of 15 million DKK (appr. 2 million Time, Activities Euro) to construct a Danish research infra- 3) Department of Scandinavian Studies and structure for the humanities integrating writ- Linguistics, The work is organized in thematically ten, spoken, and visual records into a coher- 4) University of Southern Denmark, defined groups of activities, three of which ent and systematic digital repository. The 5) University of Aarhus, are dedicated to making content available, project runs from January 2008 until the end while one focuses on the technical infrastruc- 6) Copenhagen Business School, of 2010. The budget was granted following a ture. call for expressions of interest, where our 7) The Royal Library, The content that will be made available to consortium was invited to submit a full pro- 8) The National Museum of Denmark, the research communities comprises both posal. Following an international evaluation 9) Society for Danish Language and existing basic written language resources with three reviewers, the proposal was select- Literature, (contemporary and historic, general language ed for funding with a 50% budget cut. 10) Danish Language Council. and specialised sublanguages, literary and The Danish CLARIN project was accepted professional, as well as parallel corpora with as the only proposal solely for the humani- Mission and Vision Danish as one of the languages), three differ- ties, and the factors that supported it were ent spoken language corpora and associated probably numerous: the consortium is strong The vision is to create a researcher's toolbox tools, and data resources such as traditional and to the point with four universities and by establishing a number of digital Danish and electronic dictionaries, dictionaries and four cultural institu- text, speech and visual resources and associat- semantic word nets meant for computer sys- tions, thus providing a ed tools and to inte- tems, and the linking between different dic- very good basis for a grate these resources tionaries as well as between dictionaries and true humanities pro- into a web-based envi- ject, and at the same ronment for research corpora. time with the techni- thus creating a much The technical infrastructure will include a cal skills necessary. needed support for single web user interface to serve as the Furthermore, there is Danish humanities and Danish CLARIN platform. This platform an understanding in enhance its possibilities will give access to all the tools and text the Danish research for European collabo- resources of the infrastructure, as well as a administration that the humanities need ration. The Danish CLARIN project will personal workspace, communication facili- infrastructure support, in particular IT sup- also improve the conditions for Danish lan- ties, user authentication and rights manage- port; the Ministry conducted a hearing about guage technology research and development ment, and search and retrieval facilities. C

10 No 2 / Newsletter CLARIN and BABEL (“Better Analysis Based on guage technologists, but also include other D-SPIN – the Endangered Languages", http://www.esf. humanities disciplines. Since it is not possi- org/activities/eurocores/programmes/euro ble to address all relevant subfields of the German CLARIN babel.html). humanities at this stage, D-SPIN will try to The D-SPIN project is co-ordinated by the make tools and resources available that are of University of Tübingen. The partners are: interest to scholars from the field of history. Initiative the Berlin-Brandenburgische Akademie der In order to identify user needs, D-SPIN will Wissenschaften in Berlin, the DFKI in organize a workshop with board members of Saarbrücken, the Institut für Deutsche the largest on-line community of historians Erhard Hinrichs, Peter Wittenburg, Sprache in Mannheim, the MPI for ('H-Soz-Kult'). Since libraries are increas- Alexander Geyken, Lothar Psycholinguistics in Nijmegen and the uni- ingly playing the role of historical resource Lemnitzer, Andreas Witt versities of Frankfurt, Gießen, Leipzig, providers, D-SPIN will also seek the cooper- Stuttgart and Tübingen. The project will ation of the Staatsbibliothek zu Berlin s a support action for CLARIN the invite all language resource providing insti- (Berlin State Library). The “Sammlung AGerman Federal Ministry of Research tutions in Germany to become involved in Erster Weltkrieg" (text archive of World War and Education (BMBF) has launched D- the project at an early stage. To this end D- I), a project of the Berlin State Library, will SPIN (“Deutsche Sprachressourcen- SPIN will organize a national workshop serve as a test case for investigating the use- Infrastruktur" – Infrastructure for German with all interested institutions in the spring fulness of different levels of linguistic anno- Language Resources), an infrastructure pro- of 2009. tation, e.g. lemmatisation and named entity ject of language resources and tools. Apart D-SPIN will run concurrently with CLAR- annotation. Additional possibilities include from D-SPIN, the BMBF is also funding IN and will mirror its structure in order to annotations by the scholars themselves in two additional eHumanities projects: facilitate the cooperation between both pro- the form of stand-off annotation. Textgrid and eAQUA. jects and in order to generate synergies. The exploration of such usage scenarios for All these projects will closely collaborate and However, most of the work is devoted to the field of history will help to identify spe- will coordinate the necessary steps for the national tasks, in particular to the establish- cific requirements for the construction of an construction of an eScience infrastructure ment of service centers and the localisation, eScience infrastructure which will ideally for the humanities. On the international description and integration of key language can be applied to other humanities disci- level, D-SPIN will collaborate with DOBES resources and tools in Germany. plines as well. For further information about (“Documentation of Endangered The user communities targeted by D-SPIN D-SPIN, please consult: http://www.sfs.uni- Languages", http://www.mpi.nl/DOBES) are not only limited to linguists and lan- tuebingen. de/dspin. C

including textual research papers and other infrastructure to global research communities CLARIN-DRIVER scholarly publications. The recent D-NET in a vigorous awareness and advocacy pro- v1.0 open source software release forms the gramme fostering the development of digital basis of this distributed service-oriented repositories. The DRIVER network offers a Proposed architecture that enables enhanced interop- support service for repository managers, a erability of data and service-providers, with dynamic set of guidelines aimed at data functionality ranging from search, recom- interoperability, and the strategic support for Collaboration new forms of scholar- ly communication.

Dale Peters (DRIVER), Peter Collaboration Wittenburg (CLARIN) ecently at the DRIVER Summit in January CLARIN centres and R2008 it became obvious that DRIVER aggregated services and CLARIN are complementary infrastruc- will form nodes in the ture initiatives that will benefit from collabo- robust network of ration and take profit from each others work. DRIVER content While CLARIN is a research infrastructure providers. It will offer bringing together language resources and its metadata descrip- technology, DRIVER is oriented at a broader tions so that DRIVER scope to harvest data from across disciplines. can harvest them and Here we want to briefly present the DRIVER offer them in its ser- project and discuss possible collaboration vices. Also services of elements. DRIVER repositories mendation, collections, profiling to innova- should become nodes in the CLARIN network of language resource DRIVER tive tools for repository managers. By build- ing a robust network of content providers, providers to make the data available for researchers. The primary objective of DRIVER is to estab- enhanced with the set of services and soft- lish a flexible, robust, and scalable infra- ware tools that DRIVER offers, the DRIVER The proposed collaboration between CLAR- structure for all European and world-wide infrastructure also enables further collabora- IN and DRIVER promises a joint European digital repositories, managing scientific infor- tion to domain specific research communities research infrastructure for comprehensive mation in an Open Access model, increas- in a co-ordinated network of repositories service to the humanities disciplines with respect to language resources and technolo- ingly demanded by researchers, funding from a growing number of institutional gy and therefore plays a key role in the con- organisations and other stakeholders. repositories and from national institutions struction of an efficient research and innova- DRIVER's mission is to expand its content and data aggregators in Europe. Future tion environment. C base with high quality research output, developments will see the extension of the

CLARIN Newsletter / No 2 11 August 2008 November 2008 CLARIN calendar 2008-08-04 to 2008-08-15: ESSLI 2008, Hamburg, Germany 2008-11-01 to 2008-11-03: Chicago Digital Humanities/Computer 2008-08-16 to 2008-08-24: COLING conference with pre- and post-con- Science Colloquium, Chicago, USA ference tutorials and workshops, Manchester, UK 2008-11-10 to 2008-11-11: Web services architecture in CLARIN, Munich, of events September 2008 Germany 2008-09-08 to 2008-09-12: TSD2008 conference, Brno, Czech Republic 2008-09-11 to 2008-09-12: FSMNLP2008, JRC Ispra, Italy December 2008 Here is the list of CLARIN events and events 2008-09-17 to 2008-09-19: CLEF2008, Aarhus, Denmark 2008-12-09 to 2008-12-10: European Conference on Research Infrastructures, Versailles, France from the fields of language resources and lan- October 2008 guage tools that could be of an interest to 2008-10-14 to 2008-10-15: eScience Workshop: Metadata Principles, 2008-12-17 to 2008-12-12: IEEE e-Humanities Workshop, Indianapolis, CLARIN members. Berlin, Germany USA C

Evaluations and Language resources Distribution Agency (ELDA); Paris; University of Utrecht/Netherlands Graduate School of Linguistics; Utrecht; Khalid Choukri Jan Odijk Join CLARIN Université Paris 4 Sorbonne / CELTA ; Paris; Andre Wlodarczyk ILK Research Group ; Tilburg; Antal van den Bosch CLARIN project is a combination of LIF-CNRS ; Marseille; Michael Zock Huygens Instituut KNAW ; Den Haag; Karina van Dalen-Oskam Collaborative Projects and Coordination and Germany: Berlin-Brandenburg Academy of Sciences; Berlin; Alexander Norway: Dept. of Culture, Language and Information Technology; Bergen; Geyken Support Actions, registered at the EU under Koenraad de Smedt Deutsches Forschungszentrum für Künstliche Intelligentz; Saarbrücken; the number FRA-2007-2.2.1.2. It started Department of Linguistics and Nordic Studies, University of Oslo; Oslo; Thierry Declerck Janne Bondi Johannessen with the preparatory phase in 2008 that will Institut für Deutsche Sprache; Mannheim; Marc Kupietz make the grounds for the next phases and it Det humanistiske fakultet, Universitetet i Tromsø; Tromsø; Trond Trosterud Max Planck Institute for Evolutionary Anthropology; Leipzig; Hans-Joerg Norwegian University of Science and Technology; ; Torbjørn will cover the generic, language independent Bibiko activities. In order to do our work properly we Nordgård University of Frankfurt/Main Comparative Linguistics; Frankfurt/Main; Jost Poland: University of Wroclaw ; Wroclaw; Adam Pawlowski will have to rely on a much wider circle than Gippert Institute of Applied Informatics, Wroclaw University of Technology; just the formal consortium partners in the University of Leipzig; Leipzig; Codrina Lauth Wroclaw; Maciej Piasecki project. For this reason we have opened up University of Stuttgart; Stuttgart; Ulrich Heid Institute of Computer Science, Polish Academy of Sciences ; Warsaw; all our project working groups for participa- Universität Tübingen; Tübingen; Erhard Hinrichs Adam Przepiórkowski tion by organizations that are not part of the University of Giessen; Giessen; Henning Lobin Institute of English Language, Univeristy of Lodz; Lodz; Lukasz Drozdz consortium. Computational Linguistics Department, University of Heidelberg; Institute of Slavic Studies, Polish Academy of Sciences ; Warsaw; Violetta Heidelberg; Anette Frank Koseska-Toszewa Members University of Augsburg ; Augsburg; Ulrike Gut Portugal: University of , NLX-Natural Language and Speech Group; Country; Institution; Location; Contact person Greece: Institute for Language and Speech Processing; Athens; Stelios Lisbon; António Branco Austria: University of Vienna; Vienna; Gerhard Budin Piperidis Belgium: ALT (Acquiring Language through technology); Leuven – Hungary: Academy of Sciences; ; Tamás Váradi Romania: Al.I.Cuza; Iasi; Dan Cristea Kortrijk; Hans Paulussen Budapest University of Technology and Economics Media Research (BME Institute for Computer Science, Romanian Academy of Sciences; Iasi; Center for Computational Linguistics ; Leuven; Ineke Schuurman MOKK); Budapest; Peter Halacsy Horia-Nicolai Teodorescu Center for Dutch Language and Speech, University of Antwerp; Antwerp; University of Szeged, Department of Informatics, Human Language Research Institute for Artificial Intelligence, Romanian Academy of Walter Daelemans Technology Group; Szeged; Dóra Csendes Sciences; Bucharest; Dan Tufis ELIS-DSSP; Gent; Jean-Pierre Martens Iceland: Institute of Linguistics, University of Iceland; Reykjavík; Eiríkur University Babes-Bolyai; Cluj-Napoca; Doina Tatar Legal Informatics and Information Retrieval, Katholieke Universiteit Rögnvaldsson Serbia: Faculty of Mathematics, University of Belgrade; Belgrade; Du{ko Leuven; Leuven; Marie-Francine Moens Icelandic Centre for Language Technology; Reykjavík; Eiríkur Rögnvaldsson Vitas Laboratory for Digital Speech and Audio Processing – VUB – ETRO/DSSP Ireland: National University of Ireland; Galway; Sean Ryder Slovenia: Josef Stefan Institute; Ljubljana; Toma` Erjavec ; Brussels; Werner Verhelst Israel: Technion-Israel Institute of Technology; Haifa; Alon Itai Alpineon d.o.o. ; Ljubljana; Jerneja @ganec Gros ESAT-PSI/Speech; Leuven; Patrick Wambacq Italy: Dipartimento di Linguistica Teorica e Applicata, Università di Pavia; Spain: Institut Universitari de Lingüística Aplicada, Universitat Pompeu Bulgaria: Department of Computational Linguistics, Institute for Bulgarian Pavia; Andrea Sansò Fabra; Barcelona; Núria Bel Language, Bulgarian Academy of Sciences; Sofia; Svetla Koeva Istituto di Linguistica Computazionale; Pisa; Nicoletta Calzolari Universitat de Lleida ; Lleida; Gloria Vázquez Institute for Parallel Processing; Sofia; Kiril Simov Department of Computer Science, University of Rome “Tor Vergata" ; TALG Research Group (University of Vigo) ; Vigo; Xavier Gómez Guinovart Mathematical Linguistics Departement, Institute of Mathematics and Rome; Fabio Massimo Zanzotto Sweden: Lund University; Lund; Sven Strömqvist Informatics, Bulgarian Academy of Sciences; Sofia; Ludmila Dimitrova European Academy Bozen/Bolzano; Bolzano; Andrea Abel Språkbanken, Dept. of Swedish Language, Göteborg University; Croatia: University of Zagreb, Faculty of Humanities and Social Sciences; Latvia: Institute of Mathematics and Computer Science, University of Gothenburg; Lars Borin Zagreb; Marko Tadi} Latvia; Riga; Inguna Skadina Dept. Speech, Music and Hearing, CSC, KTH ; Stockholm; Kjell Elenius Institute of Croatian Language and Linguistics; Zagreb; Damir ]avar Tilde; Riga; Inguna Skadina Uppsala University, Department of Linguistics and Philosophy; Uppsala; Cyprus: Cyprus College / Research Center; Nicosia; Antonis Theocharous Lithuania: Institute of the Lithuanian Language; ; Daiva Vaisniene Joakim Nivre Czech Republic: Charles University; ; Eva Haji~ová Center of Computational Linguistics, Vytautas Magnus University ; Kaunas; Department of Linguistics; Göteborg; Anders Eriksson Faculty of Informatics, Masaryk University ; Brno; Ale{ Horák Ruta Marcinkeviciene Department of Computer and Information Sciences, Linköping University; The Institute of the Czech Language, Czech Academy of Sciences; Prague; Luxembourg: European Language Resources Association (ELRA); Linköping; Lars Ahrenberg Karel Oliva Luxembourg; Bente Maegaard Swedish Institute of Computer Science AB ; Stockholm; Björn Gambäck Denmark: Center for Sprogteknologi, University of Copenhagen; Malta: University of Malta, Dept. of computer science; Malta; Michael Copenhagen; Bente Maegaard Rosner Language council of Sweden ; Stockholm; Rickard Domeij Dansk Sprognævn – Danish Language Council; Copenhagen; Sabine Netherlands: Meertens Institute; Amsterdam; H.J. Bennis HUMlab, Umeå University ; Umeå; Patrik Svensson Kirchmeier-Andersen Data Archiving and Networked Services; Den Haag; Henk Harmsen Turkey: Sabanci University – Human Language and Speech Laboratory; Society for Danish Language and Literature; Copenhagen; Jørg Asmussen University of Twente, Human Media Interaction Group; Enschede; Roeland Istanbul; Kemal Oflazer Estonia: University of Tartu; Tartu; Tiit Roosmaa Ordelman UK: Department of Linguistics and English Langauge, Lancaster University; Finland: CSC – the Finnish IT Center for Science ; Espoo; Tero Aalto Center for Language and Cognition; Groningen; Wyke van der Meer Lancaster; Anna Siewierska University of Helsinki; Helsinki; Kimmo Koskenniemi Digital Library for Dutch Literature; Leiden; C.A. Klapwijk Oxford Text Archive; Oxford; Martin Wynne Department of Foreign Languages and Translation Studies, University of Instituut voor Nederlandse Lexicologie; Leiden; Remco van Veenendaal University of Sheffield; Sheffield; Wim Peters Joensuu; Joensuu; Jussi Niemi Leiden University Centre for Linguistics; Leiden; Jeroen van de Weijer University of Surrey; Guildford; Lee Gillam University of Tampere; Tampere; Eero Sormunen Centre for Language Studies, Radboud University; Nijmegen; Pieter Research Institute of Information and Language Processing at the The Research Institute for the Languages of Finland; Helsinki; Toni Suutari Muysken University of Wolverhampton ; Wolverhampton; Gina Sutherland France: ALTIF; Nancy; Bertrand Gaiffe Centre for Language and Speech Technology, Radboud University; Language Technologies Unit, Bangor University; Bangor; Briony Williams TELMA/DIS CNRS; Paris; Florence Clavaud Nijmegen; L. Boves / N. Oostdijk Department of English, The University of Birmingham; Birmingham; Oliver CNTRL; Nancy; Bertrand Gaiffe Max-Planck-Institute for Psycholinguistics; Nijmegen; Peter Wittenburg Mason

CLARIN newsletter • publisher CLARIN • editor-in-chief Marko Tadi} • editorial board Gerhard Budin, Dan Cristea, Koenraad De Smedt, Laurent Romary, Martin Wynne • http://www.clarin.eu

12 No 2 / Newsletter CLARIN