Search World Forum – Trust Relevence and Rights

After introductions by Jim Kennedy of the AP, Mary Hodder of Dabble describes that the Semantic Systems needed have to have relevence and context. She gives the example of the World Cup – where people care about the “head butt” and not the play by play of the game that media provided. Would media be able to know the meaning that people discover?

Mary also mentions that standards bodies aren’t necessarily going to be able meet demand. Better, Mary explains, bottom up efforts like microformats fills in the need for standards and sharing. Moreover, top-down standards don’t provide the users the ease of use and accessibility they need.
Fabrice Florin of NewsTrust responds that their product provides trust, a qualitative evaluation of content. Their metadata product is the trust network, and the value of that meta data.

“How to scale?” – Jim asks. The goal of the AP is to standardize entity extraction and categories. Asking Josh, of the”forever” dominant search company Google, “integrate metadata and trust standards?”.

Google is concerned about scaling also, not only for centralized news wires, it also works with broader community of publishers in the blog world. They are working on analyzing news, from an algorithmic perspective, they have to find the best of breed approach at scale. Google will not be the only approach to relevance. Newstrust’s Fabrice looks at how labeling and reviews to also give journalists ranking.

An audience member asks about using http://everything2.com, which has a self policing online community system. Asking about things that don’t need editors.

Jim asks Googles Cohen if the Digg model of community review helps quality information. Josh responds that not many things have solved the problem of quality improvement. Jim furthers if Google is looking at improving quality. Google might not be using user voting, but is open.

Mary suggests that voting systems for Digg inherently lowers quality – she suggests that rating systems aren’t so important – rather that the network is important, the ability to share. Fabrice suggests that trust and verification of the sources are important.

Jim explains it’s not about search engine optimization, but to deliver better information. Mary of Dabble, uses MRSS for media meta data extraction to aggregate media, but even with standards, people use it differently. Josh furthers that geocoding is a great first step in getting relevancy.

Identity becomes the challenge explains Mary, because it’s very hard for developers to track users across sites and stakeholders. Bloggers are amateurs, and they work hard, and it’s too much, likely, to ask them to create meta data. The debate continues with many panelists commenting on how to provide valuable metadata, that isn’t gamed and spammed information.

Amy, of the audience, asks what search and metadata can do to find specific news stories – how can it help findability. Also, what can journalists do help contribute to good metadata. Jim responds that there needs to be just a basic structure to the content, like headline, byline, dateline, body, etc.

Questions also covered monetization, since quality isnt always what is profitable. Mary also explains that the problems aren’t about standards, its about how standards are used. She sees adoption as positive, and what ever is used becomes the standard. Mary also mentions http://dataportability.org, to create standards which are needed.