IPTC SpectrumNo 20 January 2006 IPTC - INFORMATION TECHNOLOGY FOR NEWS

Taking Shape - NewsML 2 Architecture IPTC Nominating Members

Agence France Presse (afp) - France - www.afp.com ANSA - Italy - www.ansa.it Associated Mediabase Limited -UK-www.mediabase.co.uk Austria Presse Agentur (APA) - Austria - www.apa.at BBC Monitoring - United Kingdom - www.monitor.bbc.co.uk Business Wire - USA - www.businesswire.com CCNMatthews - Canada - www.ccnmatthews.com CNW Group Ltd - Canada - www.newswire.ca Deutsche Presse-Agentur (dpa) - Germany - www.dpa.de Dow Jones & Company - USA - www.dowjones.com European Alliance of News Agencies - Europe - www.pressalliance.com Japan Newspaper Publishers & Editors Association (NSK) - Japan - www.pressnet.or.jp Keystone - Switzerland - www.keystone.ch KUNA Kuwait - Kuwait - www.kuna.net.kw - Japan - www.kyodo.co.jp Newspaper Association of America (NAA) - USA - www.naa.org ORF (Austrian Broadcasting Company) - Austria - www.orf.at PA News Ltd -UK-www.pa.press.net PR Newswire -UK-www.prnewswire.co.uk Limited -UK-www.reuters.com SDA/ATS - Switzerland - www.sda-ats.ch The (AP) - USA - www.ap.org The New York Times Company - USA - www.nytimes.com Tidningarnas Telegrambyrå (TT) - Sweden - www.tt.se TMNEWS-APCOM - Italy - www.apcom.it United Press International (UPI) - USA - www.upi.com World Association of Newspapers (WAN) - International - www.wan-press.org - China - www.xinhua.org IPTC Associate Members

AFX News Ltd -UK-www.afxnews.com Agence de Presse - - www.belga.be Agencia EFE - Spain - www..es Algemeen Nederlands Persbureau (ANP) - The Netherlands - www.anp.nl ANA, News Agency - - www.ana.gr AS Norsk Telegrambyrå - Norway - www.ntb.no Atex Media Command - Australia - www.atex.com Athens Techology Center - Greece - www.atc.gr BVPA - Germany - www.bvpa.org Canadian Press - Canada - www.cp.org CCI Europe - Denmark - www.ccieurope.com Cepic - Coordination of European Picture Agencies Press Stock Heritage - Europe - www.cepic.org EAST Co Ltd - Japan - www.est.co.jp/english/index.html EBU - European Broadcasting Union - Europe - www.ebu.ch Eidos Media Srl - Italy - /www.eidosmedia.com eRoket.com - USA - www.eroket.com Fingerpost Ltd -UK-www.fingerpost.co.uk Harris and Baseview - USA - www.harrisbaseview.com HINA - Croatia - www.hina.hr IFRA - Germany - www.ifra.com ITAR-TASS - Russia - www.itar-tass.com La Reppublica - Italy - www.repubblica.it Magyar Távirati Iroda Rt (MTI) - Hungary - www.mti.hu Mainstream Data Inc. - USA - www.mainstreamdata.com News Engin, Inc. - USA - www.newsengin.com NewsLink -UK-www.newslink.co.uk RelaxNews - France - www.relaxnews.com Ritzau Bureau I's - Denmark - www.ritzau.dk RivCom -UK-www.rivcom.com Suomen Tietotoimisto Oy - Finland - www.stt.fi XML Team Solutions Inc. - USA - www.xmlteam.com

2 IPTC Spectrum January 2006 International Press Telecommunications Contents Council

Chairman: Stéphane Guérillot Honorary Treasurer: Organisation Henrik Stadler Vice Chairmen: Development and Consolidation 4 Walter Baranger; Geoffrey Haynes; IPTC Membership 5 Rudi Horvath; John Iobst; IPTC Standards 6 Peter Müller; Hitoshi Saito. Management Committee 7 Managing Director: Klaus Sprick 7 Michael Steidl Editor: News Standards Summit 2005 8 Hugh Johnstone Challenge and Response 10

Published by the PR Committee International Press Making News 9 Telecommunications Council Royal Albert House Sheet Street Standards Windsor Berkshire SL4 1BE Starting from the Basics 12 England NewsML 2 Architecture Goals 13 Tel: +44(0)1753 705051 Fax: +44(0)1753 831541 NewsML 2 Architecture E-mail: [email protected] [email protected] Designing the Framework 14 NAR Metadata Handling 15 Web Portal: www.iptc.org Power Conformance 16 Contributors 17 Cover NAR Working Groups 17 NewsML 1 Maintenance A Proven Success 18 Package Item NewsML 1 Applications 19 for the NewsML 2 Architecture depicted as a NewsCodes stylised UML (Unified Modelling Language) Public Codes 20 diagram. See page 14 Automated Categorisation 21 for more details. This diagram is taken from Rev 12 of the NAR V1.0 News Content Model. Ready to Process 22 Working Groups 23

News Industry Text Format Well Proved 24

This issue of the IPTC Spectrum has hyperlinks for web addresses (in blue) and for page references (in red).

IPTC Spectrum January 2006 3 ORGANISATION ORGANISATION ORGANISATION Development and consolidation

Recent activities have helped a more fundamental approach consolidate IPTC’s position as would be appropriate. Further con- the leading source of news ex- sideration resulted in a set of pro- change standards and laid the posals that were considered by a foundations for the develop- special teleconference meeting of ment of a new family of stan- the Standards Committee, held in dards that will offer a high level early January 2005. This meeting of interoperability. resulted in a decision to change Aims of IPTC can be summa- the Working Party structure and re- rised as being “To establish and focus development efforts on a maintain an open, apolitical inter- new architecture for news stan- national forum to promote and en- dards. able the exchange of news information in an efficient manner, NewsML 2 Architecture while maintaining the highest tech- Now known as the NewsML 2 Ar- nical quality. At the same time tak- chitecture (NAR) this is intended to ing advantage of the advances in provide a single generic model for As an organisation IPTC telecommunication and computing exchanging all kinds of news infor- is run by a Management technology”. In applying these mation, and provide a common Committee and Chair aims the growing importance of the base for all new IPTC standards. elected annually at a World Wide Web has become a This model is based on a view of formal Annual General major consideration so there has the relationships between news Meeting by the been increasing interest in sys- (and topics) and the real world - nominating members. tems for multimedia news and on- further details on the concepts un- The Committee act as line publishing. derlying the NewsML 2 Architec- Directors of the ture are included in the Standards company and are Fundamental; approach pages. responsible for The decision to start work on a suc- Established standards - such as development of the cessor to NewsML 1 (which is now the NITF and SportsML are not di- organisation. Day to day over five years old) was taken early rectly involved in the NAR, so the operations are overseen in 2004 with production of a busi- established user bases will not be by the Managing ness requirements document. At affected, and appropriate develop- Director. the same time work was underway ment work will continue. However it Chair: on EventsML, while there was also is envisaged that future major re- Stéphane Guérillot (AFP) an increasing appreciation of developments taking Honorary Treasurer: place in the underlying Henrik Stadler (TT) XML technology. By the time of the Vice-Chairs: Walter Baranger (New York Autumn 2004 Meeting it Times Inc) had become apparent that Geoffrey Haynes (AP) Rudi Horvath (APA) John Iobst (NAA) Peter Müller (SDA/ATS) Hitoshi Saito (NSK) Retiring Chairman John Managing Director: Iobst (right) congratulates Michael Steidl the incoming Chairman Stéphane Guérillot (left) at Web portal: the Annual General www.iptc.org Meeting.

4 IPTC Spectrum January 2006 ORGANISATION ORGANISATION ORGANISATION

leases of these standards will be designed to integrate into the NewsML2 family of standards.

Common structure Use of the NAR will give IPTC stan- dards a common structure and style which will make them easier to implement and understand. The design is modular with common components, management rules and processes that can be reused as appropriate. Steps have also been taken to establish a common XML Schema Style for all IPTC standards, again helping to rein- force the fact that they are part of a family of standards.

Metadata handling The news industry is a major source of information, and proba- Progress Demonstration and discussion are bly one of the largest users - and Concentrating the available re- an important feature of Meetings, creators - of metadata, and an im- sources has resulted in rapid prog- helping to ensure that delegates portant feature of the NAR is a new ress, and by the end of 2005 the have a detailed understanding of way of expressing metadata. The NewsML 2 Architecture was suffi- how the work is proceeding. Here Mischa Wolf explains some NAR has been designed to be ciently advanced for a first Experi- aspects of the metadata approach compatible with the Semantic Web mental Phase to start. This is that is being adopted in the of the W3C and steps have been intended to help ensure that the NewsML 2 Architecture. taken to ensure it can be used with NAR will meet users practical re- the underlying Resource Descrip- quirements, and help refine the tion Framework (RDF). However it technical specification. nalised. It will then be possible to is important to note that the imple- It is hoped that results from the move on to detailed development mentation of metadata in the NAR first Experimental Phase will be of the first of the new standards - a does not require any knowledge of available for consideration at the general news markup standard as RDF. Spring 2006 Meeting (to be held at a successor to NewsML 1 and the end of March 2006) so the NAR EventsML - which should take Timescale Technical Specification can be fi- place during 2006. The decision to redirect efforts into the NewsML 2 Architecture has ex- tended the original timescale for development of the first members IPTC Membership of the new IPTC standards family - but it will also ensure that future embership of IPTC now includes a wide range of standards work will be easier and Morganisations representing different aspects of the news quicker. industry: news agencies, newspapers, photo agencies, Consultants have been ap- broadcasting organisations, syndicators, public relation pointed to help reduce the time organisations, system suppliers, software producers, and other needed for NAR development, and bodies involved in the news industry. to provide specialist expertise. As There are two main types of membership: a first step Jay Cousins (RivCom) Nominating members have voting rights at General Meetings and Ulf Wingstedt (Cnet) investi- and in Committee and Working Parties. They can have up to three gated ways of expressing IPTC delegates at a meeting for each subscription. This type of Standards using W3C XML membership is open to organisations and companies concerned Schemas and provided some Ar- with news collection, distribution and publishing. chitecture Implementation Guide- Associate members are eligible to vote in Working Parties and lines. can send one delegate to a meeting. Associate membership is The same consultants are work- available to organisations and companies concerned with news ing on XML Schemas to imple- collection, distribution and publishing and to system vendors ment the NewsML 2 Architecture. supporting the news industry. Special membership terms may be In addition Mark Birkbeck (x- offered to new members from economically distressed areas and port.net Ltd) has looked at ways in academic institutions. which the NAR metadata require- Further information on IPTC membership is available from ments can be reconciled with the www.iptc.org W3C RDF abstract model.

IPTC Spectrum January 2006 5 ORGANISATION ORGANISATION ORGANISATION

Development combination of emails - on dedi- week) teleconferences. Most of the work on the NAR has cated development discussion For example in the period be- been carried out by a small group groups that are restricted to IPTC tween the 2005 AGM and the 2005 of delegates (see page 17) using a members - and regular (two a Autumn Meeting work on the NAR involved nearly a thousand emails and around forty hours of telecon- ference time. As with other IPTC IPTC Standards work the success of this project owes much to the individuals who contributed their time, and to their NewsML 1 - a media-type-independent XML-based standard for organisations for enabling them to the representation of multimedia news throughout its lifecycle, make the commitment. including production, interchange and consumer use. www.newsml.org Established standards Although efforts have been con- News Industry Text Format (NITF) - XML-based markup centrated on the NewsML 2 Archi- scheme for individual news items. It can be used as a standalone tecture, work has continued to format for news distribution or as a content model for NewsML. ensure that the established stan- www.nitf.org dards continue to meet users’ needs. There have been new re- SportsML - standard for the interchange of sports data, leases of SportsML (V1.6) and the including scores, schedules, standings, and statistics. XML-based. NITF (V3.3), while NewsCodes de- www.sportsml.org velopment is a continuing process. A new Working Party was estab- NewsCodes - controlled vocabularies of terms widely used in lished to deal with any issues aris- the news industry. They include: an extensive Subject taxonomy; ing with NewsML 1, which has a roles and genres; and ratings for priority, urgency and relevance; substantial user base. The impor- www.newscodes.org tance of NewsML 1 has also been recognised by its adoption and re- IPTC Core - photo metadata primarily for use with Adobe’s lease as a Japanese Industrial Extensible Metadata Platform XMP. Successor to the widely used Standard - see page 18. “IPTC Headers” and part of the IPTC4XMP effort. www.iptc4xmp.org Standards co-operation Co-operation with other Standards Under Development bodies is an increasingly important aspect of IPTC’s activities and dis- NewsML 2 Architecture - this is not itself considered to be cussions on metadata issues have a formal standard for public release but the architecture on which been held between IPTC and W3C new standards will be based. Planned new standards include a representatives. Following the suc- general news markup (as a successor to NewsML 1) and cess of the first News Standards Summit, IPTC were co-sponsors of EventsML (intended for the exchange of information on event the second News Standards Sum- publishing, planning and coverage). mit - see page 8. A collaborative effort between Legacy Standards IPTC, the IDEAlliance and Adobe Systems resulted in an XMP Information Interchange Model (IIM) - container for news Schema and custom panels to al- information in any of the common news media (including text, low the use of the IPTC Core in photographs, graphics, audio and video). A subset of the IIM Adobe Photoshop. metadata sets was adapted to provide the “IPTC Headers”. In response to a request from Ifra www.iptc.org/IIM a Colour Space Task force was set up to investigate ways of retaining the EXIF data associated with digi- Digital Newsphoto Parameter Record (DNPR) - tal camera image files during container file format designed to carry digital news photograph agency processing. data within the IIM. www.iptc.org/IIM Participation IPTC7901 - original text message format developed by IPTC, Participation of non-members is and still in widespread use. The last revision of this standard was encouraged with a series of public carried out in 1995. discussion groups, covering the www.iptc.org/IPTC7901 NAR and individual standards - de- tails of these groups are given in IPTC standards and supporting documents are all available the sections dealing with individual for download from the links given above. standards. In addition relevant document are made available on

6 IPTC Spectrum January 2006 ORGANISATION ORGANISATION ORGANISATION

the IPTC Web site as work pro- lated many of the challenges faced ceeds. by the news industry - see page 10. Another aspect of the challenges Klaus Meetings faced by the news industry - this Sprick Following the established arrange- time the growth in volume and ments there were three Meetings availability of financial information in 2005: The Spring Meeting was was looked at by Ken MacKenzie, held in San Diego; the 40th Annual Reuters Global Head of News Ca- The 2005 General Meeting in ; and pabilities. A third Reuters speaker Spring the Autumn Meeting in Milan. All was Jeremy Lebrett who provided Meeting saw were well attended, and provided a an overview of the development the departure wider forum for discussion of the and application of XBRL (eXtensi- of Klaus Sprick, who stood work carried out by the develop- ble Business Reporting Lan- down from the Management ment groups guage), along with a Committee on his retirement In a new departure, as noted demonstration of how XBRL infor- from dpa. After joining the above, a special teleconference mation can be handled in NewsML. German News Agency (dpa) session of the Standards Commit- The way that metadata is dealt with in 1968 he quickly became tee was held in January. This ap- in XHTML 2.0 was explained by involved in IPTC activities - proved changes to the Working Steven Pemberton, W3C HTML which at that time were Party structure so they were in Working Group and Forms Work- mainly concerned with place for the Spring Meeting. ing Group Chair. safeguarding press Two speakers from the Press As- telecommunication services. Presentations sociation (PA) addressed different As IPTC started The Annual General Meeting was aspects of news content. Phil development of standards held in London at the invitation of O’Brien from EMPICS demon- for the news industry, Herr Reuters, who arranged a stimulat- strated the ShootLive system Sprick was appointed the ing series of speakers to comple- which provides real-time photogra- first chair of the new ment the technical sessions. phy for new-media applications, Technical Committee, which Delegates were welcomed by and Jennifer Campbell, Managing produced the first IPTC Geert Linnebank, Reuters Editor- Director Meteo Consult explained technical standard - IPTC in-Chief and Global Head of con- the Weather services offered by 7901 - in 1979. Election to tent with an address that encapsu- her organisation. the Management Committee followed in 1984, a position he held until his departure, including a term as IPTC Management Committee 2005-2006 Chairman from 1990 to 1993. here were a number of changes to the Management Committee at Klaus Sprick’s Tthe 2005 Annual General Meeting. John Iobst (NAA) had completed considerable contribution to his three-year term as IPTC Chair and Stéphane Guérillot (AFP) was IPTC was recognised by a returned unopposed as the new Chair. unanimous vote of the The departure of Klaus Sprick (see above right) left a vacancy on the Committee of the Whole Management Committee, and this remained open until the formal IPTC Council to make him the first Annual General Meeting - which deals with the running of IPTC as a Honorary Member of IPTC. company - when a new Committee was elected. Members are: In addition a dinner was held in his honour in San Diego (during the Spring Meeting) Chair Vice-Chair Vice- at which he - a keen sailor - Stéphane Walter Chair was presented with a model Guérillot Baranger Geoffrey of a J-class yacht as a (AFP) (New York Haynes memento. Times Inc) (AP)

Honorary IPTC Vice-Chair Vice-Chair Vice-Chair Treasurer: Managing Rudi Vice-Chair Peter Hitoshi Henrik Director Horvath John Iobst Müller Saito Stadler Michael (APA) (NAA) (SDA/ATS (NSK) (TT) Steidl

IPTC Spectrum January 2006 7 ORGANISATION ORGANISATION ORGANISATION

A regular feature of Meetings has Olympic briefing details of the press arrangements been presentations by members to The practice of arranging briefings and meet the management team. bring various aspects of their activi- for Olympic events continued with a This included Richard Palfreyman, ties to the notice of other members. visit to Torino - which is hosting the Head of the Main Press Centre in During 2005 these presentations in- XX Winter Olympic Games - at the Torino, who had previously met cluded the NewsEngin CatScan invitation of Mr Cristiano Carlutti, some of the delegates at the IPTC categorisation system, the services Torino Press Operations Managing briefing for the Olympics offered by Mainstream Data, and Director. 2000 (where he was head of Press the Eidos Méthode editorial system. Delegates were able to discuss Operations). News Standards Summit 2005

Intended to help users and developers gain a known as a Request for Comments (RFC). Bob better understanding of standards used by the Wyman, CTO, PubSub provided a quick news industry, and to show how the standards overview - with examples - of the standard and of work in practice, the second News Standard RSS - another standard that has a complex Summit was held in Amsterdam in May 2005, history. in conjunction with the Xtech 2005 PubSub - Publish and Subscribe - provide Conference. subscribers with rapid notice of new content, and The opening session provided an overview of techniques are being developed to deal with the the main standards in use (or under “Grey Web” - pages that search engines find development), with a second session looking at difficult to index. ways the standards are being used in practical An update on the PRISM (Publishing applications. Both sessions were chaired by Requirements for Industry Standard Mischa Wolf (Reuters). Metadata) and ICE (Information and Content Brief details of the presentations are given here Exchange) standards was provided by Dianne - the presentation slides, and audio files of the Kennedy, VP Publishing Technologies, presentations are available from IDEAlliance. www.newssummit.org/2005, providing a There was extensive coverage of current and continuing resource for interested users. planned IPTC standards: NITF features and its application by dpa-infocom Standards were considered by Hubertus Koehler, Chief Some aspects of the work being carried out by the Technical Officer, who also gave some W3C were covered by Steven Pemberton, W3C information on the EU-funded MINDS (Mobile HTML Working Group and Forms Working Group Information and News Data Services) project. Chair who provided an introduction to Metadata in NewsML and the IPTC News Architecture were XHTML 2.0. Aims of the work on XHTML 2.0 are explained by Laurent Le Meur, Technical to make it more accessible and usable. Steps Manager of Multimedia Development (AFP) IPTC have been taken to integrate XHTML with RDF in News Architecture WP Chair. a way that is accesible to the HTML community. Continued > Ivan Herman, Head of Offices W3C, explained that the Semantic Web is a metadata based web Future Attractions infrastructure, with RDF (Resource Description Framework) being a general model for metadata statements. The W3C Ontology Language Sponsors of the second News Standards (OWL) can be used to define knowledge concepts Summit were IPTC, Ifra, and the IDEAlliance, and relationships, with a series of increasing and it was organised organised by Diane levels of complexity. Kennedy (IDEAlliance), Harald Loeffler (ifra), The role of metadata in broadcasting was Michael Steidl (IPTC) and Misha Wolf considered by Jean-Pierre Evain, Senior (Reuters). Engineer, EBU with the main types being The first News Standards Summit was held administrative, technical and descriptive. There is in the USA in December 2003. It is intended a high level of complexity and efforts are being to hold further events but the date for the next made to define common semantics, while an EBU News Standards Summit has not yet been News Exchange Schema has been developed. fixed - it will probably take place after the TV-Anytime provides metadata so consumers release of the new IPTC NAR-based can search and select broadcast content. standards. Suggestions for the next News Atom has a complex history but is now an IETF Standards Summit Program can be sent to (Internet Engineering Task Force) standard, [email protected].

8 IPTC Spectrum January 2006 ORGANISATION ORGANISATION ORGANISATION

An outline of EventsML was provided by Arnaud tent, such as the TopNews system. Descamps, Chief Technical Officer, Relaxnews. AFP have moved towards the XML Development of the IPTC Core for XMP, with an representation of multimedia news with the overview of Adobe XMP, was explained by IPTC “Magazine Forum” being a pure NewsML platform, Managing Director Michael Steidl. and the approach taken was described by Laurent Le Meur. Applications As a major news host the APA Database Host Specific examples of the ways IPTC standards are makes use of standards for the main operations of being put to use were provided by a number of input categorisation and download. Rudolf Horvarth IPTC members: and Manfred Mitterholzer from APL-IT also demon- Business Wire use of a NewsML 1 profile for their strated the APS-IT Power Search full-text retrieval news distribution system - and the way it was intro- system. duced to their customers - was explained by Dean Some of the special features of mobile portals Large and Jayson Lorenzen. were explained by Kevin Smith, Architect, Dave Compton outlined the development of Vodafone Global Technology with reference to the Reuters News Systems - from IPTC7901, through Vodafone Live data service. Use of appropriate the IIM to NewsML 1, and went on to consider the standards allows improved search and marketing development of systems offering added-value con- facilities.

PR COMMITTEE PR COMMITTEE PR COMMITTEE Making News

As with other IPTC activities, larly high download figures for maintaining the organisation’s items like the IPTC NewsCodes public image - and encouraging and other standards information. membership - relies heavily on the efforts of members, coupled Press releases with continuing input from the Press releases are issued on a IPTC Managing Director. regular basis to cover activities at Presentations at appropriate the main meetings, and for major events have an important role in events such as the launch of the creating and developing interest in News Standards Summit and the All aspects of IPTC’s the standards - and in making con- start of the NewsML 2 Architecture public relations are dealt tact with systems suppliers who Experimental Phase. Releases are with by the Public work with the news industry. Dur- distributed by IPTC members, who Relations Committee. ing 2005 such presentations were also issue press releases covering This includes held at the NEXPO (USA) and Ifra their IPTC-related activities. encouraging better (Europe) events, while the Manag- understanding of the ing Director gave a talk on “IPTC Publications organisation and and Metadata” to the European As- Publications remain an important standards, and sociation of News Agencies. part of the publicity mix with the promoting the electronically distributed newslet- advantages of Web site ter - IPTC Mirror - complemented membership. The IPTC web site - and related by the annual IPTC Spectrum. This sites for individual standards - pro- is distributed in both printed and vide the public face of the organi- electronic form and is intended to Chair: Walter sation, with the aims and activities provide a less-technical overview Baranger (The clearly explained. Standards docu- of IPTC activities and achieve- New York Times) mentation and related information ments. is available for free download. Web Both the IPTC Mirror and the statistics show that the IPTC Web IPTC Spectrum are available on pages are well used with particu- the IPTC web site.

IPTC Spectrum January 2006 9 ORGANISATION ORGANISATION ORGANISATION Challenge and Response

At the IPTC Annual General Meeting, Geert Linnebank, Editor in Chief of Reuters, looked at the challenges facing the news industry and suggested some of the ways in which the industry can respond.

he IPTC is an important standards confidently predict that this new environment, of Torganisation and one which Reuters is bloggers, alternative news co-operatives, committed to, and very pleased to be, supporting. aggregators, search engines and the likes is set to That is not least because of the challenges the overwhelm the established industry of which we IPTC, and the news industry it represents, face. form part, and we can certainly argue about that, Why is this so, and what can we do about it? and at some length. Well, as I see it, the issue is this, in a nutshell. The news industry - our industry - has been in the Standards business of making B2B (business-to-business) But what is much more difficult to argue with is the standards for decades. And while there has been speed at which that environment is developing stiff competition within the industry, among and increasing its reach. And that, I contend, is ourselves, taken as a whole we didn't face any due because it is built entirely on standards. Not real competition from the outside. the kind of standards which we ourselves have So it did not matter much how difficult our traditionally developed and used. Not the kind that standards were to understand - or implement - or it takes our industry years to develop and to use. And it did not matter much how slowly those implement. standards of ours evolved. They were the only No, in this new news environment, this parallel game in town. So we became a bit complacent, universe, standards evolve like in warp time. perhaps a bit lazy, even. Generation follows generation many, many times faster than in our own world. Some survive and Change evolve, while others simply get left behind. Take And then, one day we all woke up and we found the growth of that blogosphere, which has been that the world had changed. So how did the world driven almost entirely by a standardised, low-cost change? What happened? publishing environment and also, to some extent, What happened is that a comet hit the earth, so by RSS. to speak. It sent, and continues to send, and will continue to send, shockwaves round the globe. Simplicity This comet is, yes, you guessed it, the Internet. It is probably a truism to say that RSS owes it’s We all know the Internet is much, much more success to what it says it is - Really Simple - easy than cheap connectivity. It dramatically changes to implement and use. Now I recognise that some the rules in the information space. It has brought of our needs are more extensive, but I think there immensely powerful natural selection into the is still a lot we can learn from RSS about making world of standards and software. This in turn has standards accessible. resulted in the emergence of entirely new I talked about the two parallel universes, and paradigms for information interchange. hinted at the risks that the two remain separate. I firmly believe our challenge is to bring the two Parallel environment together in a way at is both beneficial to the These new paradigms have been extremely thousands of millions of people who rely on the successful, resulting in the emergence of a very lnternet to get and exchange information, and also different kind of “news environment”, one which to our industry with its professionalism, its largely operates in parallel to ours, and one which creativity, its mission. is showing exponential growth where ours is static at best. Adopting change This new “news environment” is not based on I do see some encouraging signs that this can be the traditional one-way 'broadcast" model that our done, if we're prepared to adapt and adopt industry grew up around, but rather on a fully change. There is, for instance, the example of the networked model. Now some commentators BBC, a professional news organisation and

10 IPTC Spectrum January 2006 ORGANISATION ORGANISATION ORGANISATION

eminent representative of that industry of ours. has worn off, using propriety technology usually The BBC increasingly succeeds in using its equates to additional cost to the business. And if website, its online presence, and its brand, to we can make savings by adopting standards then solicit contributions from the public - both textual we can devote more to those areas where we do, and visual. These it uses to enrich its news and must, differentiate - primarily in the content service. Safeguards are in place to protect the itself. BBC from hoaxes or setups. And the contributors We must ensure that the standards allow us to feel they're “part” of the BBC, which helps secure play to our strengths, and also deliver the best their loyalty. possible service to our customers and consumers. What the BBC is doing here, if on a small scale From that we will all gain. still, is bridging the two parallel universes. In the process it is expanding its network of eyewitness Search reporters by millions. Compare that with Reuters Let's not forget search. Recent developments in 2,300 journalists worldwide - the very best and online search have revolutionised the way people brightest, highly trained and very professional, but consume information. How often do you land on a also potentially utterly outnumbered. Web site via search rather than typing in a URL for the home page? We all use search to get Vital role roughly what we want. I passionately believe that despite the emergence Although competition is hotting up, Google of the “parallel universe”, the industry represented remains the benchmark here. It has created ever around this room has a vital role to play in the more sophisticated search techniques bringing in future. Our journalistic standards, our video, mobile services and much more, and it professionalism, our processes to validate are continues to innovate. At the same time, it keeps arguably more relevant today, not less, as the tide users happy with a simple look-and-feel interface of raw, unverified, information, rumour, gossip and to all its services. propaganda keeps rising like never before. But unless we dare to venture outside our Metadata “walled gardens”, challenge our “broadcast” Metadata - a hot topic everywhere, and not least approach and look for workable ways to become here at the IPTC - is at the heart of effective part of, and of harnessing, the Internet, there is a search. Google knows this: their search algorithm real risk that we lose our voice, and with it our is based on the implicit metadata of “who links to relevance. what”. But it is not always perfect, and it resists the top-down metadata view. Co-operation We can ask ourselves, against the might of a Back to standards. I think we all accept the need Google is there really a role for us in metadata? for co-operation - at least as a principle. But the Well, consider this ... when a news items breaks, reality is sometimes quite different. Our business no one can have linked to it! I believe this is where is often fiercely competitive. We all really want that the real experience we all bring comes in. Working scoop, that exclusive. This means that we often together on how we define, code and then search work in isolation from each other. That's for content is core to delivering the high quality, understandable and good, but it also makes it multi-media services many of us here are difficult to create industry-wide standards. committed to providing. Our customers want us to standardise on the means by which they can link and bind with us. This address has been edited for publication That doesn't mean they want us to standardise on everything - far from it, They want us to compete, and aggressively, on the quality of Public participation our content, on the coverage we provide, on its timeliness, its reliability, its angles, its voice. An indication of the success that the BBC is having in obtaining These are core strengths by contributions from the public is given by the response to the which we will continue to London bombings - the first of which took place on the 7 July, just differentiate ourselves. over a month after the IPTC AGM. According to the BBC, within an hour of the first bomb going off they had received fifty pictures Cost-effective from the public, with around twenty-two thousand emails and text Now I think I speak for everyone messages being received in the day. here when I say cost does In their terms and conditions BBC News state that contributors matter. We all need to be as agree to grant them “a royalty-free, non-exclusive licence to cost-effective as we can - that's publish and otherwise use the material in any way that we want, a simple fact of life. My belief is, and in any media worldwide”. They also make the point that the standards really help us here. copyright remains with the contributor and that they will endeavour After the first-starter advantage to give contributors appropriate credit.

IPTC Spectrum January 2006 11 STANDARDS STANDARDS STANDARDS Starting from the Basics

Overall aim of the new family of Work on a successor to NewsML IPTC standards is to establish a 1 started in spring 2004 and initial single generic model for ex- consideration of the requirements changing all kinds of newswor- was combined with efforts on other thy information. This will be planned standards such as achieved by the development of EventsML, and existing standards the NewsML 2 Architecture like SportsML. As work proceeded (NAR) which will provide the it was decided that a more funda- framework for future IPTC ex- mental approach would be better, change standards. so work was started on the Although NewsML 1 has been NewsML 2 Architecture. widely adopted by the news indus- try - see page 18 - feedback from Working structure users - and non-users - indicated A revised working structure was in- that there were a number of signifi- troduced to support this work with cant areas in which improvements the creation of a new NewsML 2 could be made. Specific points that Architecture Working Party (page needed to be addressed included 14) responsible for producing the the level of complexity, a lack of framework, and a renamed and re- consistency in applications, and structured News Content Working limited metadata handling capabili- Party (page 22) dealing with the in- ties. dividual content standards based Responsibility for the on the framework. technical activities of XML developments Existing Working Parties con- IPTC belongs to the To some extent these reflect the in- cerned with the NITF (page 24) Standards Committee, creasing experience in the practi- which sets priorities, cal application of XML standards A new working structure was assesses available and developments in XML technol- established to simplify resources, establishes ogy. It has to be remembered that development of the NewsML 2 the Working Parties and NewsML 1 was an early XML appli- Architecture. Activities related to cation - it is now more than five the NAR are shown within the bold oversees their work. blue frame. There are two Working Formal approval of new years old, which is a long time in Parties, each of which has standards - and related such a rapidly developing technol- subsidiary Working Groups to deal material - has to be ogy like XML. with different aspects of the work. given by the Standards Committee before they are released.

Chair: Henrik Stadler (TT)

Web portal: www.iptc.org

12 IPTC Spectrum January 2006 STANDARDS STANDARDS STANDARDS

and the IPTC NewsCodes (page 20) continue unchanged, while a new Working Party was created for NewsML 2 Architecture Goals NewsML 1 Maintenance (page 18). • To simplify and unify the overall design for representing The decision to restructure the newsworthy information. working arrangements and con- • To be flexible, thus allowing lightweight “no bells and whistles” centrate efforts on the NewsML 2 Architecture was taken at a special feeds and highly complex news feeds, based on the same teleconference of the Standards model. Committee in January 2005. • To specify more details, leaving less space for interpretation. • To streamline the processing model, providing only a single way Rapid progress to express specific structures and functionalities. With the aims firmly established, • To develop a new model for expressing metadata from the work has proceeded at a fast pace, ground up. resulting in development of the • To provide an abstract model to be implemented by specific NewsML 2 Architecture Model and news exchange standards. Technical Specification drafts along with a draft XML Schema • To maintain, at the functional rather than syntactic level, a high files by the end of the year. This level of backward compatibility with NewsML 1. made it possible to start an initial • To simplify the implementation of IPTC news exchange Experimental Phase (EP1) in De- standards as a whole. cember 2005. Results from EP1 • To align IPTC news exchange standards with requirements from (expected during March 2006) will the “Information Highway”. help refine the concepts and the way they are implemented, to en- sure that the aims are being met. as being “what is a matter of fact with standard tools. Underlying and is good to know”. In this case syntax is XML (Extensible Markup Abstract relationships the real world is represented by Language from the W3C) while the Underlying concept behind the factual knowledge of concepts in design makes use of W3C XML NewsML 2 Architecture is the basic the real world. These may be real Schema and complies with the relationship between the abstrac- objects, such as people, organisa- W3C RDF (Resource Description tions of news and topics and the tions, locations and other material Framework). This should allow real world they report on and de- items, or abstract concepts repre- easy transfer of information to scribe. sented by the codes used to iden- other XML-based standards and In this context news is consid- tify the subjects of content. integration into the Semantic Web. ered as being “what has happened and is fit to be published” and oc- Common framework New standards curs in the real world. The content Basing new IPTC standards on a As the name suggests the of news - as written by a journalist, common framework - the NAR - NewsML 2 Architecture is intended captured by a photographer or means that operations will be car- as the underlying structure for new cameraman, or in the form of struc- ried out in a consistent way, mak- news-exchange standards. All new tured content - represents a more ing them faster to understand and IPTC news exchange standards abstract level, though the actual easier to implement. The same will be based on the NAR and it is technology used to report the news function will always be handled in anticipated that future versions of does not make any difference to the same way, in all standards existing standards will be aligned the abstract “news content”. The based on the architecture. with the NAR where possible and NewsML 2 Architecture is de- So far as possible the new struc- appropriate. signed to provide a wrapper for this ture will use generic building Present plans include: “news content” which can be ex- blocks, this means that develop- A new standard for general news, pressed in any media. ment of new IPTC standards to succeed NewsML 1. This will Dealing with the news needs should be much simpler. In addi- carry general news but will not pro- more than a technical representa- tion, implementers will be able to vide packaging features; tion of the content, so a second produce software components that Events ML, which will carry struc- level of abstraction is used to pro- can used in different applications - tured events information; vide information about the content, such as for handling various types A major revision of SportsML with the metadata being applied to of news content. which will allow use of the sports give a formal description, summary A generic model has been data structures as content of a or categorisation. adopted that will allow for future NAR News Item; Applying a management layer to extensions. This means that it will Revision of the IPTC NewsCodes the content and metadata results in be possible to extend standards to to use NAR technology to express a “News Item” which can be readily meet future - as yet unknown - re- a set of topics and their relation- identified and processed within quirements of the news industry. ships; news systems. Use of industry standards means Integration of ProgramGuideML as In a similar way, Topics are taken that processing can be carried out a NAR-based standard.

IPTC Spectrum January 2006 13 NAR NAR NAR NAR NAR NAR Designing the Framework

Central element of the new IPTC tended Audience and Editorial standards family, the NewsML 2 Note. Architecture (NAR) will provide Examples of other Common the underlying structure for new Components include Descriptive content exchange standards. Metadata (again made up from a Important features of the ap- set of other components) Person, proach adopted are the develop- Organisation, Location, Label, Sig- ment of a set of reusable Common nature, and Rights Metadata. Components along with an under- All Common Components have a lying structure - based on man- consistent design and format, and aged Items - that can be used for are maintained for use in IPTC processing all types of news con- standards development. It is antici- tent, and a system for the repre- pated that some standards will also Task of the NewsML 2 sentation of metadata. need dedicated components, and Architecture (NAR) these will follow the same design Working Party is to Common Components principles. develop an architectural Common Components can be con- framework for the sidered as building blocks that can Any Item be used in the construction of stan- Content, both news and topics, is management and dards. At the lowest level, the data handled in the form of Items, which distribution of all types types represents a group of values are manageable objects with a per- of news-related content. and operations that can be carried sistent, unique identity and a set of This framework will be out on the values. These include management properties that allow the basis of all new IPTC basic data types and more com- processing. An abstract Any Item news exchange plex data types where the basic acts as the model for all managed standards. data type is extended by attributes to include additional information to more fully represent the value. Chair: Laurent Le Typical examples include dates, NAR status Meur (AFP) strings and numbers. A basic component represents a The overview of the single piece of business informa- NewsML 2 tion and uses one of the data types Architecture given on to provide the model for the con- these pages is based tent. Basic components may be on the version used on their own or as part of an- released in December Vice-chair: Misha other component. Wolf (Reuters) 2005 (for use in the Aggregate components repre- first Experimental sent composite pieces of informa- Phase). tion with a specific business Development of the meaning, they can be made up NAR is continuing and from basic components and other although the main aggregate components. For exam- features have been ple Administrative Metadata is a established it is Discussion groups: Common Component which is important to remember http://groups.yahoo.com/group/ made up of a series of other com- that design changes newsml-2 may be made. http://groups.yahoo.com/group/ ponents - including such items as iptc-metadata Date Content Created, Creator, In-

14 IPTC Spectrum January 2006 NAR NAR NAR NAR NAR NAR

Representation of the News Item. Portions shown in blue are the Any Item, which is the model for the News Item, while specific News Item components are shown in yellow. The Signature Component (in green) is part of the Any Item but can only be used in applications that conform to the “power” compliance level. Connections indicate that components are included in the individual items. It is also possible to have a generalisation relationship where one part is a specialised version of the other. This type of relationship is shown in the Package Item diagram on the front cover. items and three generic Items have the item), a unique Item Identifier Management Component been developed for use in the (which has a form similar to that of In turn the Item Metadata Compo- NewsML 2 Architecture. These are a NewsML URN - RFC-3085 - but nent contains a Management the News Item, Topic Item, and the without version information) a Component and a collection of Package Item. If any new content Catalog (for scheme alias declara- Links. standard has specific re- tions, (see “NAR Metadata Han- Information directly related to the quirements not met by the generic dling” below) and an Item management of an Item is carried items a new Item will be generated Metadata Component. It may also by the Management Component, - again this will be based on the have an Item Version,aLanguage This includes such information as Any Item. indicator (the default for the con- indications of the nature of the item Mandatory parts of the abstract tent of the item) and a Signature and of the content; individual dates Any Item are a Schema Version Component (see “Power Confor- for Item Created/Modified/Re- (for the XML Schema that specifies mance” on page 16). leased/Embargo Ends/Item Re-

NAR Metadata Handling by an abbreviated form known as a scheme alias, and the code is then added after a colon. The Efficient, flexible and extensible mechanisms resulting {scheme:code} pair is called a CURIE. for handling metadata are an essential feature Any given item may use references to a large of the NewsML 2 Architecture. number of different taxonomies, while different Simple forms of metadata that been identified users will want to use different taxonomies for use in the NAR are String, Date, Integer and (including ones developed for their own use). To IRI (Internationalised Resource Identifier. This is a identify specific taxonomies the scheme alias complement to the URI - Uniform Resource parts of the CURIEs used in an item are included Indicator - that allows the use of a wider range of in the “Catalog” structure which is at the top of characters, such as non-Latin scripts). each of the NAR Items.This provides a mapping In many cases metadata will take the form of between the scheme alias and the URI of the coded values taken from a controlled vocabulary scheme. Since this “Catalog” could be fairly large (or taxonomy), with a typical example being the it can be stored in a central repository and simply IPTC Subject NewsCodes. To comply with the referenced from the Item. Semantic Web each term is identified by a URI Provision has also been made for Labels, which (Uniform Resource Indicator). This URI is made are a specific forms of metadata that are intended up from a (separate) URI for the scheme and the to be human readable such as notes and code value. instructions However, URIs can be fairly long - for example The NAR is designed to be compatible with the the URI for the IPTC Subject NewsCodes could Semantic Web of the W3C and the structure be “urn:newsml:iptc.org:20001006:topicset.iptc- adopted is compatible with the Resource subjectcode:”. So to avoid having to include the Description Framework (RDF). However, it is complete URI each time the scheme is used a important to note that the implementation of compact syntax is needed to handle them. In the metadata in the NAR does not require any NAR the intention is that scheme will be replaced knowledge of RDF.

IPTC Spectrum January 2006 15 NAR NAR NAR NAR NAR NAR

Content Metadata diagram showing the Administrative and Descriptive metadata. The Publication Component and Rights Component (shown in green) can only be used in applications that are compliant to the “power” level.

an article about a person to their bi- ography, for example. Derivation Links establish par- ent/child relationships. Typical ap- plications include linking a translated article to the original one, or an edited picture to the original.

News Item News content is carried in a News Item. The content may be in any media type or format, and may be tired; the Publish Status, be published and accessed but no structured. Typical examples are a identification of the provider; and new references should be created. news report, a picture or a video the conformance level the Item Taken together application of clip, and structured sports or complies with. these values allows flexible man- events data. Specific values for the Publish agement of items in an active news Characteristics of a News Item Status are; Usable - the Item may environment. include: short term interest (as be published without restriction; news is volatile); it will usually be Witheld - until further notice the Links updated for only a short period, Item must not be published or Links provide a named relationship and may then be archived; it is ex- used, and must be withdrawn or re- between the Item and another tar- pressed through a set of alterna- tracted if already published; and get item (or web resource) - links tive renditions of some media Cancelled - must not be published do not have be included so the col- content; it refers to an arbitrary set or used and must be withdrawn on lection may contain zero to several of concepts and entities; and it may retracted if already published. links. be associated with other News These status values can be Navigation Links provide a con- Items or Web resources. qualified by: Embargoed - must not nection between an Item and an- Since the News Item is derived be published until the specified other related Item or a Web from the Any Item it has all the as- date and time; and Retired - may resource. This can be used to link sociated management properties.

Power Conformance integrity and message authentication for any type Two conformance levels have been defined for of data - this is provided by supporting the model the NewsML 2 Architecture - “core” and and syntax defined by the W3C. Similarly there “power”, with the “core” level designed to will be a Rights and Publication Components in provide simplicity and interoperability between the Contents Metadata. It will be possible to applications. associate both Signature and Rights Components The “power” conformance level includes all the with individual content parts, as well as to a “core” functionality plus extra features. It offers a complete item. higher degree of flexibility, but at the cost of more Another area where “power” compliance levels complex programming, and some loss of will be an important feature is in metadata interoperability (since not all applications will handling and as a first stage provision has been implement this level). However, “core” processors made to apply identification, details of the creator can be designed to deal with “power” level and the creation date to all metadata. With features by ignoring the information that cannot be complex metadata types it will also be possible, processed. for example, to indicate the equivalence of So far only some explicit “power” features have different codes in different schemes, or to give a been identified - these include a Signature degree of confidence in the accuracy of the Component, which makes it possible to ensure metadata.

16 IPTC Spectrum January 2006 NAR NAR NAR NAR NAR NAR

It also has a Content Metadata items, while the items remain inde- a News Message, though use of Component which carries the Ad- pendent of the package which only this is optional as any other syndi- ministrative and Descriptive Com- keeps references to them. Typical cation protocol can be used for ex- ponents with the metadata items are the “top ten” list of news change if required. A simple associated with the Item. Descrip- stories or the representation of a structure is used with exchange tive metadata can generally be in- newspaper page section. properties located in a message ferred when the content is used Again, since it is derived from the header and the items being ex- (such as language, genre, sub- Any Item, the Package Item has changed contained in an Item set. jects, titles) while the Administra- the associated management prop- tive metadata provides information erties, while it also has a Content that cannot be inferred, such as Metadata Component and a Group when and where it was created, Set which represents a tree of Contributors and who by - see the diagram on items. Root of the Group Set is a page 16 for further information. Group Component, while the tree The NewsML 2 Architecture The News Content Set wraps al- may contain other Group Compo- Model and Technical ternative renditions of the content, nents along with links to News Specification documents which can be included directly (as Items and Topic Items (and possi- incorporate work by the XML content or plain text), in en- ble future items) in any order. following : coded form (such as an image) or Each Group Component has a Mark Birbeck (x-port.net), be remote and referenced by a hy- Group Role (to show the part it Dave Compton (Reuters), perlink. plays), a Group Mode, and a col- Jay Cousins (RIVCOM), lection of Group Components and Honor Craig-Bennett (Press Topic Item Composition Links to include ex- Association), Arnaud Intended to carry information about ternal items or Web resources in Descamps (RelaxNews), concepts (both named entities like the package (by reference only). Takahiro Fujiwara (East), organisations and abstract like the There are three Group Modes: Darko Gulija (HINA), Paul news subject), the Topic Item will Complementary and Unordered Harman (PA), Alan Karben generally only contain short, struc- (this is the default); Complemen- (XML Team), Jeremy Lebrett tures information about the topic tary and Ordered (where items (Reuters), Laurent Le Meur and its relationships with other have to be use or displayed in a (AFP), Johan Lindgren (TT), concepts. specific sequence), and Alterna- Jayson Lorenzen Typically there will be long term tives (where there are equivalent (BusinessWire), Stuart interest for the content; the content pieces of content - such as transla- Myles (Dow Jones), Michael will generally be updated infre- tions - with the user selecting one Steidl (IPTC), Miles quently, but over a long period; it of them). Whitehead (Reuters), Ulf will be focused on single concept Wingstedt (CNET) and or entity; and it will be referred to by News message Misha Wolf (Reuters). a large number of other resources, Support for the exchange of items particularly News Items and other in the news workflow is provided by Topic Items. As with the News Item the Topic Item is derived from the Any Item so has the identification and man- NAR Working Groups agement components, it also has a Content Metadata Component for Work on the NewsML 2 Architecture is carried out by a set of descriptive and administrative me- Working Groups. Their efforts are combined by the NewsML 2 tadata. A Topic Content Compo- Architecture Working Party to produce the NAR Model nent supports properties specific to and Technical Specification, and to guide development the concept being handled, along of the XML Schemas. with links to associated topics. Pro- Chair of the NAR Working Party is Laurent Le Meur vision is made for Specialised (AFP) and Vice-Chair Mischa Wolf (Reuters) - see page Topic Content which may be a 14. XML structure. This is a flexible Common Components - production of components (in structure that allow inclusion of Johan the NAR framework) that will be used in more than one structures that represent, for ex- Lindgren content markup standard. Lead: Johan Lindgren (TT). ample, information about a person News Management - establishing processing models for or a location or an organisation. all types of news content covered by recent IPTC news exchange standards. Lead: Stuart Myles (Dow Jones). Package Item News Metadata Framework - specifying how metadata Third of the generic Items, the will be expressed, referenced and managed in all new Package Item provides some major versions of IPTC standards. Lead: Misha Wolf structure to news related informa- (Reuters). tion - typically of medium-term in- Stuart News Structure - developing an abstract news model for terest. Structure is expressed Myles IPTC standards. Lead: Laurent Le Meur (AFP). through a hierarchy of groups and

IPTC Spectrum January 2006 17 NEWSML 1 NEWSML 1 NEWSML 1 NEWSML 1 A Proven Success

Released in 2000, NewsML 1 is URN specification (RFC 3085), designed as a multimedia struc- which had not been established tural framework for news, that when the original documentation can be applied at all stages of was produced. the news life cycle. When released it was envisaged Application efforts that typical uses would be for the Specific efforts have been made to transfer of news between news promote the use of NewsML 1 in agencies and their customers, Japan with NSK (Nihon Shinbun publishers and news aggregators Kyokai - The Japan Newspaper and between news service provid- Publishers & Editors Association) ers and end users, as well as within setting up a series of task forces, editorial systems. initially to introduce NewsML 1 to An indication of the success of the Japanese news industry, and NewsML 1 in finding such uses is then to provide support and further given in the table on page 19. This encourage its adoption. gives brief details of some of the These efforts mean that the stan- The NewsML 1 applications that have been devel- dard is now widely used - for exam- Maintenance Working oped - most of this information has ple Kyodo News have been Party deals with the been supplied to IPTC by the or- delivering NewsML to newspapers functional specifications ganisations concerned, and it and Web sites since 2002. As part and requirements for seems likely that there are many of their application they developed NewsML 1, and specifies more applications. an extensive series of DTDs by implementation analysing legacy text formats so guidelines with New standard the information can be carried in proposals for best The NewsML 2 Architecture draws NewsML ContentItems. practice, with the aim of on the experience with NewsML 1 Kyodo News have adopted the and makes use of recent XML de- NITF and the IPTC Subject News- promoting the use of velopments and it is intended that Codes for their services. Now that NewsML 1 for packaging there will be a new “general news work on the new IPTC standards and syndication of content” standard to take over from family is well under way, Kyodo are multimedia news. NewsML 1. also working on the migration of Because of this it is not envis- their applications from NewsML 1 aged that any major modifications, to the NewsML 2 Architecture. Chair: Dean Large (Business Wire) or additions, to NewsML 1 will be undertaken. An important aspect JIS X7201:2005 of current work is establishing a An indication of the importance of clear and easy to implement tran- NewsML 1 to the Japanese news- sition path to the new standard. paper industry is given by its re- lease as a Japanese Industrial Update Standard - JIS X7201:2005 - in Vice-Chair: The last major revision of the stan- July 2005. Takahiro Fujiwara (EAST Co. Ltd) dard was V1.2, which was Work to achieve this was initially released in October 2003. How- carried out by NSK teams, followed ever, work undertaken to preparing by a JIS Working Party which in- NewsML 1 for submission as a cluded users, system providers, Japanese Industrial Standard (see outside experts and academics. A below) identified a number of minor Japanese version of the NewsML Web site: errors in the NewsML 1.2 Func- Technical Specification was pro- www.newsml.org tional Specification and these were duced, while the Implementation corrected during 2005. At the same Guidelines are also available in Discussion group: time the opportunity was taken to Japanese, along with other sup- http://groups.yahoo.com/group/ include a reference to the NewsML porting documents. newsml

18 IPTC Spectrum January 2006 NEWSML 1 NEWSML 1 NEWSML 1 NEWSML 1

Palace system for content management is based NewsML 1 Applications on NewsML. MarketWire (www.marketwire.com) - press release distribution of corporate news and media Agence France Presse (afp) (www.afp.com)- advisories. NewsML compliant multimedia news feeds. The “Magazine Forum” is a pure NewsML platform. NewsPress (www.newspress.de) releases available in NewsML format. ANA (www.ana.gr) - news feeds in NewsML. PA News (www.pa.press.net) “Top 10” story ANSA (www.ansa.it) - “ipernews” real-time news collections in NewsML, with associated pictures, collections are produced using NewsML. to a wide range of traditional media and corporate APA (www.apa.at) - text feeds and multimedia organisations, covering a variety of topics such as feeds formatted as NewsML available. news, sport, and entertainment. Asia Corporate News Network (ACN) PR Newswire UK (www.prnewswire.co.uk)A (www.asiacorpnet.com) Financial and corporate NewsML compliant editorial system. press releases. Reuters (www.reuters.com) NewsML compliant Belga (www.belga.be) - Multimedia feed (text, multimedia news feeds. picture, audio and video) to media clients and SA/ATS (www.sda-ats.ch) - NewsML compliant telecom operators. multimedia news feed. Business Wire (www.businesswire.com)- The Irish Times (www.ireland.com) NewsML NewsML compliant news feed. compliant text news feed of the daily newspaper. Chunichi Shimbun Group (www.chunichi.co.jp)- The Tokushima (www.topics.or.jp) Newspaper NewsML-based photo database for a group of six production system with full multimedia capability daily newspapers. based on NewsML. CTK (www.ctk.cz) - multimedia feed in NewsML Thomas Publishing Company format. CTK’s new multimedia editorial system is (www.thomasnet.com) ThomasNet Industrial based on a XML format derived from NewsML. Newsroom provides industrial news articles. Most Japan Newspaper Publishers & Editors news distribution is now done in the NewsML Association NSK (www.pressnet.or.jp) Many format. A daily snapshot of the news is available members offer NewsML compliant news feeds as NewsML - and editorial systems for newspapers based on http://news.thomasnet.com/NewsML/inr-newsml- NewsML. daily.xml. JCN Newswire (www.japancorp.net) Regulatory United Press International (UPI) (www.upi.com) News Service for publication to financial terminals - NewsML compliant text news feed. and the Internet. Wall Street Journal Online (www.wsj.com)- Mainichi Newspapers (www.mainichi.co.jp)- news feeds packages delivered in NewsML.

There is also a substantial group of suppliers based products for the news industry including a offering NewsML 1 compliant systems. These NewsML format checker. include: Eurocortex (www.eurocortex.fr) - pure web solutions to manage (edit, search, export) Athens Technology Centre (www.atc.gr) - News newswire information and multimedia content for Asset Agency Edition is NewsML capable. the Press and Media industry. CCI Europe (www.ccieurope.com) - Editorial Netwyse (www.netwyse.fr) - Koorou content Systems use IPTC NewsML with IPTC NITF for management system supports up and down two-way integration of third party systems for NewsML content streams and allows updates to archiving, web content management, e-paper, NewsML content. wires, events, syndication etc. Promind Systems Inc (www.xslmaker.com)- Delta Solutions (www.deltasolutions.nl)- XSL Maker is a visual design tool for Web pages Software for the production and distribution of that can apply CSS styling to NewsML content. (audio)-news. NewsML is used for the communication with the editorial intranet and the Transtel (www.transtel.com) - “nm_fusion” suite distribution to end users. for software applications for news content. NewsML feed up and down stream can be Digital Collections (www.digicol.de) - Asset handled. Managment solution uses IPTC NewsML for integration of wires, editorial, web CMS, e-paper, Xtenit (www.xtenit.com) - communications syndication etc. platform for managing content and delivering e- mails. Content management integration can East Co (www.est.co.jp) - range of NewsML handle NewsML feeds.

IPTC Spectrum January 2006 19 NEWSCODES NEWSCODES NEWSCODES Public Codes

Metadata is the key to efficient there has been more gradual use of information, and its value growth with additions in selected is much greater when common areas as users develop their clas- values are used within an indus- sification systems. There is a for- try. IPTC have developed, and mal mechanism for the submission maintain, an extensive set of and approval of new entries, which metadata for use by the news in- have to be proposed by a IPTC dustry, known as the IPTC member - details are available on NewsCodes. Although the the NewsCodes Web site. NewsCodes represent consider- able intellectual capital they are Replacement entries freely available, and can be One of the changes made to the downloaded from the IPTC web Subject NewsCodes involved the site. replacement of an existing Sub- Most widely used of the News- jectMatter heading by two new Codes are the Subject News- ones. Consideration of the implica- Codes, which also appear to be tions of this decision resulted in the finding applications beyond the introduction of a formal policy for Responsibility for news industry. There is a three deprecating entries, which also maintenance and level structure with seventeen deals with cases where it is consid- development of the IPTC main Subjects, a second level se- ered appropriate to move or re- ries of SubjectMatters under each place an entry. sets of metadata - Subject, and a third level of Sub- Under this policy the deprecated specifically designed for jectDetails under the SubjectMat- heading is left in the Subject News- news applications - lies ters. Over 1300 terms have been Codes but marked up as depre- with the NewsCodes defined, and where appropriate cated with a recommendation that Working Party, which several codes can be assigned to there should be no further use. The also undertakes the give a detailed description of the numeric code remains associated creation of new code content. with the original entry so existing sets when appropriate. uses will not be affected and will Language independent continue to return the original Each of the Subject NewsCodes meaning. has a “formal name” (a numeric Chair: John code), a “name” and an “explana- Versatile Minting (UPI) tion”. The names and explanations Initially developed for use with are developed and maintained in IPTC standards such as NewsML English (British), but can be trans- and the IIM (Information Inter- lated into other languages for use. Although the words change the (numeric) code remains the same Vice-Chair: Honor and is language independent. For Craig-Bennett (PA some languages translations sup- News) plied by IPTC members have been made available. A comprehensive overhaul of the Subject NewsCodes was carried out in 2004, with the addition of a substantial set of new entries, Web site: along with the provision of expla- www.iptc.org/newscodes nations for all entries. Since then

Discussion group: Updates to the NewsCodes are http://groups.yahoo.com/group/ available as NewsML 1 and RSS newscodes feeds

20 IPTC Spectrum January 2006 NEWSCODES NEWSCODES NEWSCODES

change Model - see page 6), there are twenty-six sets of NewsCodes in addition to the Subject News- codes (and the associated Subject Qualifiers). These sets cover a wide range and examples of the way they can be used include: Describing content (including Of Interest To, Genre, Role, Rele- vance, and Scene - along with the Subject NewsCodes); Handling news items (including NewsItem type, Status, Priority); Detailing technical parameters (in- cluding Encoding, Colorspace, Characteristics property).

Review and update The NewsCodes have been devel- oped over an extended period, drawing on IPTC members experi- them because they were entities The NewsCodes Viewer provides ence of the news industry. Follow- rather than subjects. an easy way to browse the codes ing the overhaul of the Subject However, the value of a set of and compare translations where NewsCodes, an extensive review Entity NewsCodes was appreci- available. Shown above are the of the other NewsCodes sets is ated and the possibility of generat- Subject NewsCodes (the viewer works with all NewsCode sets) with now under way, with the aim of en- ing one is under consideration - the code tree to the left and the suring that the entries meet current though it is appreciated that putting selected term at the top to the needs, and that all terms have it together, and maintaining it, right. The Spanish equivalent of proper explanations. would be a major undertaking. the chosen term, and its New sets of NewsCodes are de- Updates of the IPTC NewsCodes explanation are shown at the veloped where there is a clear re- are issued when changes or addi- bottom. quirement, and a recent tions have been formally ap- submission for addition to the Sub- proved, and users can register to ject NewsCodes was for a set of receive immediate notification of New approach headings to cover International Or- any new versions by NewsML or Metadata handling has been an ganisations. Although it was recog- RSS feeds. Latest versions of all important feature of work on the nised that there were practical the NewsCodes are available for NewsML 2 Architecture, with de- advantages in having identifiers for download in the form of TopicSets velopment of a new model for ex- such organisations it was not con- and there is a dedicated viewer for pressing the metadata. This has sidered appropriate to include browsing and using them. been designed to allow easier inte- gration of the metadata with the W3C’s Semantic Web and is com- patible with RDF (Resource De- Automated categorisation scription Framework). Introduction of NAR-based standards offers A major driving force behind the increasing use of the IPTC new opportunities - and raises Subject NewsCodes is the increasing availability of automatic fresh challenges - for development systems for applying the codes. and application of news-specific The CatScan system developed by NewsEngin metadata sets. (www.NewsEngin.com) is based on the IPTC Subject NewsCodes At the same time there have and also scans for organisations, people and places of particular been suggestions that the three- interest to the user. It is said to be capable of scanning twenty level hierarchical structure of the typical news stories, or press releases, a second - or more than a Subject NewsCodes is starting to million documents a day. be a limiting factor in their develop- The system uses an array of marker terms related to each ment and there is a requirement for category and looks for matches between the marker terms and both internal and external links. terms in the story. Each marker term has an associated score, and Taking these factors into account a when there is a match the term’s score is added to the total for the major initiative has been launched relevant category. If the overall score for the category exceeds a to look at the requirements for, and threshold, the category is applied to the story. application, of new taxonomy stan- Specific advantages in using the IPTC Subject NewsCodes dards for news applications. Al- include the way that they allow a “bottom-up” approach with each though this will be a IPTC category contributing to its parents, and that they are a familiar development effort, steps are be- standard. ing taken to seek input from other interested parties.

IPTC Spectrum January 2006 21 NEWS CONTENT NEWS CONTENT NEWS CONTENT Ready to Process

When news content of a specific cover events such as Formula 1 type is marked up in a standard and the Indy-car series. Exten- way it can be much easier to de- sions may be needed to cover liver and process. The News other types of motor sport, but this Content Working Party investi- will depend on user demand as gates news areas that would SportsML developments are benefit from the use of a com- strongly user-driven. Similar con- mon mark-up and works to- siderations apply to plug-ins for ad- wards the introduction of ditional sports. appropriate mark-up systems Current release of SportsML is (based on XML). V1.6 which includes a number of Where an area is considered of enhancements to the core. The sufficient interest - in terms of the new motor plug-in relies on some amount of news coverage involved of these enhancements made for The News Content - one of the first steps is to estab- V1.6, so is not backwards compati- Working Party oversees lish the business case. Assuming ble with older versions of the development and that this is favourable more de- SportsML. support of standards for tailed work on a standard may be There is also a draft XML the mark up of all sorts undertaken, depending on the Schema version of SportsML. A of news-related material, available resources. Where possi- download package containing the ble work can be carried out in col- V1.6 DTD, draft XML Schema, re- with most of the work laboration with other standards source files, documentation and being carried out by bodies, though the particular re- examples is available from Working Groups quirements of the news industry www.SportsML.org. dedicated to specific means that this can be difficult. interest areas. IPTC4XMP SportsML For some time a series of datasets Longest established of the IPTC generally known as the “IPTC specialised content programs, headers” have been used to attach Chair: Geoffrey SportsML was first released in metadata to images, using photo Haynes March 2003. It has a modular editing packages such as Adobe (Associated Press - AP) structure with a core DTD to han- Photoshop. These datasets origi- dle information that is common to nally came from an early version of many sports - including scores, the IIM (Information Interchange schedules, standing and statistics Model - see page 6) and IPTC (including wagering statistics). have since developed a much This core is supplemented by wider range of metadata. Web sites: plug-ins dealing with information Developed by Adobe Inc, XMP is www.eventsML.org that is specific to individual sports. a general way of embedding XML www.sportsML.org www.iptc4xmp.org For example, for baseball this in- data in a file, and uses the W3C www.ProgramguideML.org cludes the types of pitch and hit RDF (Resource Description and ways a player can get out. Framework). Available as an open Standard descriptions for the ac- standard for use by developers, Discussion groups: tions, and for teams, are held in re- XMP has been widely integrated in http://groups.yahoo.com/group/ source files. the Adobe product range. eventsml Current plug-ins cover American A collaborative effort between http://groups.yahoo.com/group/ football, baseball, basketball, golf, IPTC, Adobe Inc, and the IDEAlli- sportsml ice hockey, motor racing, soccer ance has produced an “IPTC Core” http://groups.yahoo.com/group/ and tennis. Most recent addition is Schema for XMP that allows ProgramGuideML/ http://groups.yahoo.com/group/ the motor racing plug-in - at the straightforward transfer of data iptc4xmp moment this is mainly intended to from “IPTC Headers” to the XMP.

22 IPTC Spectrum January 2006 NEWS CONTENT NEWS CONTENT NEWS CONTENT

In addition custom panels have been developed to provide an in- terface to the IPTC Core within Adobe Photoshop. A package including the IPTC Core Specification, technical docu- mentation and a set of custom pan- els for use with Adobe Photoshop CS can be downloaded from http://www.iptc4xmp.org, where the RDF XML Schema for the IPTC Core is also available. In addition video tutorials on use of the panels - provided by David Riecks and the Stock Artists Alliance - can be found at http://www.stockartistsal- liance.org /tutorials. information about coverage of There are four custom panels to EventsML events by news organisations (of- allow the use of the IPTC Core with As the name suggests, EventsML ten referred to as a “Daybook”). Adobe Photoshop CS. Shown is intended to be used for exchang- Typical uses of EventsML are above is the Image Panel which ing information about events. For seen as being: within news organi- contains information about the this purpose events are defined as sations, where the information visual content of the image. The other panels allow for contact “something that happens”, while might be used for a daybook or for information, description of the for news applications the events the management and planning of visual content of the image, and also have to be subject to news news coverage; for event organis- workflow and copyright details. coverage. Events may be planned ers to provide information to the or unplanned, with the special case public - either as individual events common features with the efforts of “breaking news” being capable or as part of a series; and for ag- being made to develop a replace- of developing into a major area of gregators (and on-line portals) to ment for NewsML 1, and this was a activity. gather information about various significant factor in the decision to Main interest areas are for: events and publish the information concentrate efforts on the NewsML Event publishing – communicating in a consolidated format - or make Architecture. information about events, including it available to users from a central Following this decision the associated news items, source. Initial work on the standard EventsML Business Requirements Event planning – managing the resulted in production of an formed part of the input to the coverage of breaking or upcoming EventsML Business Requirements NewsML 2 Architecture (NAR) and newsworthy events, including sup- document (this is available on direct development of EventsML port for gathering associated news www.eventsml.org). was held back until the NAR was items; However, it was realised that the available. It is anticipated that Event coverage – communicating work being carried out had many EventsML will be one of the first members of the new IPTC stan- dards family.

Working General news content General news content Lead: Laurent Le Meur Another standard that will be Groups (AFP) based on the NAR is the new gen- eral news content format that will follow on from NewsML 1. Development work on the news In contrast to standards like content standards is undertaken SportsML and EventsML, the new by individual Working Groups standard will carry journalistic con- within the News Content tent that is typically unstructured - Working Party. Current groups this would include text, images, include: Alan Johan audio and video. Where appropri- SportsML Karben Lindgren ate features of NewsML 1 will be Lead: Alan Karben (XML Team carried forward to the new stan- dard, but it will not be concerned Solutions) Laurent Le Dominic Johan Lindgren (TT) Meur Chan with packaging and exchange functions (as was the case with EventsML NewsML 1). Leads: Johan Lindgren (TT), One of the first steps was to es- Dominic Chan (Canadian tablish the NewsItem as a central Newswire) concept of the new standard, but it became apparent that this had

IPTC Spectrum January 2006 23 NEWS CONTENT NEWS CONTENT NEWS CONTENT

more general applications and so candidate standard in 2004, but provides feedback to the develop- its development was moved to the has not yet been brought into gen- ers of these external standards - in NewsML 2 Architecture Working eral use. addition some members are ac- Party. The possibility of reworking tively involved in the work of other The NewsItem will probably in- ProgramGuideML into a member standards organisations. clude a management component, of the NAR family is being consid- One area of particular interest is metadata constructs, a content ered - this might be as a way of financial data - this is a major metadata component, and a news handling data from TV-Anytime (a source of information for the news content set that will contain alter- standard supported by the EBU industry - and it appears that XBRL nate renditions of the content. that provides programme informa- (Extensible Business Reporting tion for consumer use). Language) is being more widely Programme data adopted. The approach being ProgramGuideML (www.program- Other standards adopted by the NewsML 2 Archi- guideml.org) is an interchange The News Content Working Party tecture should make it much easier standard for radio and television also maintains a watching brief on for future IPTC content standards programme information, based on relevant external developments. to handle data produced using NewsML 1. It was adopted as a Where appropriate the group also other MLs like XBRL.

NITF NITF NITF NITF NITF NITF Well Proved

The NITF is a XML based stan- indication is given by the fact that dard that can be used to struc- the NITF discussion group has ture individual news articles, some 440 members, and contin- Development and and provides information about ues to grow. maintenance of the well- the content and about the docu- Many of the points raised in the established News ment itself. group appear to be from new users Industry Text Format Content information may include - with other members of the group (NITF) is carried out by text formatting, specific identifica- generally being able to provide an- the NITF Working Party tion of the subject of the content swers to questions raised and pro- (using IPTC Subject NewsCodes), vide other guidance. and enriched text elements that Chair: Alan Karben can be used to identify people, Changes (XML Team places, organisations, hyperlinks Similarly, the mature nature of the Solutions) and so on. Content is not restricted standard is reflected in the fact that to text but may also include tables, there tend to be few requests for lists and images, for example. additions or modifications. These Document information can in- seem to arise as users develop clude details such as when and more complex applications and where it was produced, copyright generally tend to be minor. Vice-Chair: Charles Tichenor information and news manage- Version 3.3 was released in (The Associated ment attributes. 2005, and a typical change was Press - AP) modification of the

element, Widely used which is used to hold the definition Longest established, of the IPTC of a term - the change was simply XML-based standards (released in to extent the content model for the in 1999), the NITF is probably the element to include rich text. most widely used XML vocabulary Web site: in the news industry. XML Schema www.nitf.org As a well documented and stable Work on a XML Schema for the standard it is relatively straightfor- NITF is under way - this is a major Discussion group: ward to implement, so it is hard to development that does seem to be http://groups.yahoo.com/group/ nitf assess just how popular it is. Some wanted by some users.

24 IPTC Spectrum January 2006