lemeur

Slide1: 

News Standards Summit IPTC NewsML Laurent Le Meur (AFP) NewsML support WP chairman December 2003

Implicitly structured news: 

Implicitly structured news ANA0059 4 S 0182 GBR /AFP-CLF29 Golf-EPGA-IRL,profile Irish golfers hoping for rare home success DUBLIN, July 2 (AFP) - Irish Ryder Cup star Padraig Harrington will be the favourite to win the European Open at the K Club on Thursday despite never having tasted European Tour success in his home country.Harrington won the 1998 Irish PGA Championship but while he has had European Tour success in almost every continent in the world he has yet to win an event on home turf. The 31-year-old Dubliner will also be keen to emulate compatriot Darren Clarke who won at the K Club in 2001. Harrington, who has already enjoyed success this year with a 19 under par win at the German Open in May, will also be using the K Club as a warm up event for the British Open on July 17. bm/pi03 Categories (3 max) Keywords Attributes place, date, credit Priority

Binary photo (IIM): 

Binary photo (IIM) FAS46 - PARIS FRANCE : A model presents a creation by British designer Julien MacDonald for Givenchy 11 October 2003 in Paris during the ready-to-wear Spring/Summer 2004 collections. Reference: FAS46 Created: oct 11, 2003 Author: Pierre Verdy (STF) Category:ACE, Fashion Origin:FRA, PARIS Urgency:5 Credit: AFP Geometry: 1648x2464 AFP

NewsML – what is it?: 

NewsML – what is it? « Media-independent, structural framework for multi-media news“ (IPTC) Created by the IPTC, supported by NAA Roots in IPTC7901, IIM, NITF … & XMLNews It is about News metadata News packaging News management News exchange & syndication Tech: Monolithic huge DTD, no RDF

NewsML - Dates: 

NewsML - Dates October 2000; V1.0 October 2002: V1.1 Compatible October 2003: V1.2 Minor release, compatible; documentation October 2004? V2.0 Major release, new XML technologies

NewsML – Some providers: 

NewsML – Some providers News agencies AFP, ANSA, ATS, PA, Reuters, UPI … Web sites Dow Jones’ WSJ.com News distributors Business Wire Newspapers NSK members (Japan) Studied by the chinese news community

NewsML Concepts: 

NewsML Concepts Media independent Structure only Multi-support 4 levels Content (ContentItem) Structure (NewsComponent) Management (NewsItem) Exchange & syndication Concepts more important than syntax Content Structure Management Exchange

Content level: 

Content level ContentItem Extraction of characteristics Inclusion of content Reference to content Media neutral ContentItem Caracteristics Content payload NITF or XHTML

Structure level: 

Structure level NewsComponent Extensible metadata See later Collection of objects Complements Alternatives, with basis for choice Contains different types of objects Complex trees, role for each constituent NewsComponent Metadata

Management level: 

Management level NewsItem Identification State (usable, embargoed, withheld, canceled) Partial update (before, after, replace, delete) NewsItem Management parameters; e.g. identification: urn:newsml:afp.com:20031208:Hkg2456:1 Before A After B Replace C Delete D

Collections & associations: 

Collections & associations NewsItem Inclusion of items Reference to items Named relationships NewsItem Relationships Inclusion of items Reference to items

Exchange level: 

Exchange level NewsML envelope 1+ NewsItems Exchange parameters NewsML envelope Exchange parameters: destination date product priority NewsItems

Metadata : 

Metadata Media-independent Administrative About the creation of the news (who, when, where) Descriptive About the news (language, genre, subject, target audience, topic occurrence) Copyright and Usage rights About the ownership and use of the news NewsLines = metadata for humans, card index Quick comparison with Dublin Core, PRISM and XMP

Controlled vocabularies: 

Controlled vocabularies = taxonomies (dictionaries) language and media independent Supported by a news oriented set of metadata: Type, Genre, Subject (see next slide), Topic: Place, event, organization, person …

IPTC Subject codes: 

IPTC Subject codes Categorization / editorial subject More detailed than legacy news categories (INT, ECO …) More than 1.200 terms A document may have several subjects Three levels : name (sport), matter (swimming), detail (50m butterfly), + qualifier (final, men) Numeric codes (ex 15062016 + [15000001, 15000024]) Wide use in the news community Check IPTC web site to get TopicSets

Use Case: Agence France Presse: 

Use Case: Agence France Presse AFP: worldwide multilingual & multimedia news agency Multimedia News (-> 2000) around 150 NewsML media customers 400 NewsML stories a day, 5 languages Structure++, associations++, ease of integration++ US: news.yahoo.com Fr: www.tv5.org Wireless news (-> 2002) Text wires (->2002) >7000 NewsML stories a day, 6 languages Via internet & satellite, + archives Metadata++ (subjects …) Maps internal XML (production) to NewsML (distribution) Uses internally the concepts of NewsML

Multimedia products: 

Multimedia products Collections Structured text Multi-resolution illustrations

Illustrations & Links: 

Illustrations & Links Illustration with caption and metadata Links to other articles Links to other collections

Mobiles sites (i-mode, WAP2): 

Mobiles sites (i-mode, WAP2) Mon Info 1. La Une 2. Incroyable! 3. Sport 4. Cinéma - Culture ------------------- 5. Plus d’info 6. Photos 7. Sport Direct 8. Alertes 9. Recherche La Une Reprise des contacts entre Israël et les Palestiniens La Une La Une Un soldat israélien fait asseoir deux Palestiniens près d'un stock d'armes ... 0. i-mode Dernière heure: Arrestation d'un Serbe de Bosnie accusé d'une campagne de viols à Foca JERUSALEM (AFP) - Israéliens et Palestiniens devaient tenir mardi une seconde rencontre, après une interruption des contacts de deux mois… 0. Accueil 1. Voir photo 1. Reprise des contacts entre Israël et les Palestiniens 2. La justice désigne des responsables dans l'hormone de croissance 3. La chancellerie s'implique dans le sang contaminé 4. Bush veut punir les PDG indélicats 5. Le Tour arrive enfin en France 0. Accueil 0. Accueil

NewsML quick summary: 

NewsML quick summary 3 years old standard Supported by a standards body - IPTC News B2B exchange / syndication Media independent 4 conceptual levels Rich features (but perceived as complex) Multimedia packages (trees) Includes or points to content No API, few open software tools

NewsML v2: 

NewsML v2 October 2004 (?); same concepts. Simpler to implement Simple and modular conceptual model Standardized processing model, ? API Use of namespaces, XSD schema Compact (direct consumption) Usable through whole news lifecycle (e.g. storage & processing) Clear documentation Along with the evolution of all IPTC standards Rocket / payload model Some convergence/integration with other initiatives (AdsML, JDF, PRISM, Atom ...) ?

Links: 

Links IPTC web site (www.iptc.org, www.newsml.org) Newsgroups NewsML & NITF (groups.yahoo.com/group/newsml) (groups.yahoo.com/group/nitf) Japanese community (www.newsml.jp) Chinese community (www.cnewsml.org) Toolkit java (sourceforge.net/projects/newsml-toolkit) CD-Rom « Learn NewsML » (www.widelearning.com)