From CD-ROM to Web 3 0 News and

From CD-ROM to Web 3. 0, News and business information is a constant challenge SLA Presentation September 2008

History • Canadian News Disc with Southam (copublishing) • Third generation of • Second generation of • First generation of online platform Online platform online platform • Périodisc La Presse, 1 st CD of news-reel in Canada product for libraries (Virtual News Library) Offices On-Line Era CD Era 1989 • Launching of new • Launching of Biblio-Branché • Opening of Ottawa 1990 1991 1992 1993 1994 1995 1996 1997 • CD Launching of Le Monde à Paris • 1 st contract: CRTC, production of a CD-ROM • SI 3, 1 st system of daily electronic press reviews in Canada 1998 1999 2000 • Foundation Ixiasoft (spin-off) 2001 2002 2003 • Addition of international sources • Exceed 100 sources 2004 2005 2006 2007. . . • Exceed 1 000 sources • Opening of Paris Offices • Actualité Québec, 1 st multi-title CD in Canada • Opening of Offices in Toronto Product/Service Achievement “A First”/ Launching of Technology Solution Geographical Expansion - Offices 2

Overview of CEDROM-SNi Offices in Montreal, Toronto, Ottawa and Paris 100 employees 1200+ publications – 800 traditional publications (daily, weekly, magazines, . . . ) – 400 social media (web sites & blogs) 80 000 news stories per day 85 millions in the archive We still do CD-ROM. . . 3

Content was king But it’s changing. . . There is more and more newspapers that give their content freely on their websites – Ads versus subscription model – Google puts newspaper archives for free over the internet Social media is growing – At last, the public will have a channel to talk There is too much content now and that’s the new challenge 4

What do you expect from a CMS? Relevant Content – – – Validated content Selectable content External content but also internal content Powerful search engine – – Full text but also multi-criteria query Boolean and proximity operators (no better way yet) A system that will be able to launch queries to others systems (federated search) Relevancy mechanism (to whom? ) Monitoring & Alert system – To alert you as soon something arrived or once a week/day to give you an overview of the week/day Document Management system that helps you to organise the documents – – To keep important or relevant documents To prepare a folder on a topic for you or someone else Shareable system to publish (share) the content with colleagues, clients, . . . – – 1 to 1, 1 to N, N to N Intranet Posting, RSS feed, e-mail, . . . You need a system that will be available Anytime and Anywhere – Web access but more and more on your phone A DRM system – If there is any rights over the content Ok! but what’s next? 5

Next step will be. . . Analysis of the search results (Summary) – Taxonomy will be very helpful to cluster the results – Semantic analysis will be very useful to identify strong concepts and entities inside of the full text 6

7

This is a big challenge Finding documents that fit a query within a large information haystack was a challenge 20 years ago but we’ve found a pretty good solution for that Obtaining very quickly information on each of those documents and “crushing” that information to give a summary of the results is a new challenge – Retrieving the name of the authors from a list of 2 million documents and calculating the numbers of news stories for each of them in less than 2 seconds. . . That’s a challenge 8

Next step will be. . . Analysis of the search results (Summary) – Taxonomy will be very helpful to cluster the results – Semantic analysis will be very useful to identify strong concepts and entities inside of the full text Creation of content with the content – Web 3. 0 – The Semantic Web 9

The WEB 3. 0 10

11

12

wikipedia Official website blogs Social media Social network IPOD –MP 3 Player launch in 2001 by Apple Buzz rate : 87 % Satisfaction rate: 74 % (détails…) Price : 99$ - 499$ (détails…) Competitors : Microsoft Zune, Créative Zen Cons : Battery’s life, Hard disk reliability (20%), screen is fragile. 13

14

Next step will be. . . Analysis of the search results (Summary) – Taxonomy will be very helpful to cluster the results – Semantic analysis will be very useful to identify strong concepts and entities inside of the full text Creation of content with the content – Web 3. 0 The sentiment (tone analysis) – This feature will help the system to red flag content that is very negative or very positive – Very useful in identifying trends – It is NOT perfect and need a lots of fine tuning 15

16

Sentiment Most systems are currently doing only sentiment at the document level – Not really interesting We are working with NStein to have a tonality by sentence, paragraph, document but also by entities. – This document is talking about “Nortel” and it is 85% negative “Cisco” and it is 55% positive “Linksys” and it is neutral. . . – This kind of information is very useful to create charts and to do some brand monitoring 17

Conclusion With the first generation of our products, we gave you access to the content (Fast search, fast delivery) With the second generation, we gave you 24/7 monitoring system (Alert) With the current generation, we gave you the platform to “animate” your community. (Share folders, intranet posting, RSS, Blackberry, . . . ) In the next step, we will give better tools to view the forest, not only the trees. – You will be able create content from the content (Dash board, trend analysis, . . . ) and you will be able to share that new information with your community. 18

Questions 19