Social Search – what did the tutorial at Compute and HeadStart reveal ?

I attended the tutorial on ‘Search and Social platforms’ at the recently concluded HeadStart and Compute 2009. The speakers were Priyank Garg from Yahoo who manages their search engine products, Rajdeep Dua and Vijaya Machavolu from Google who work on OpenSocial implementations, and Jai Ganesh of Infosys who works on how enterprise may put social networks into good use.

All three talks – on BOSS, OpenSocial, and Social Network analysis – were individually good but somewhere, I guess, the connect between the topics was missing. While Yahoo’s Priyank spoke about BOSS APIs and about Vertical Lens that allows site and content owners to submit data and meta data to Yahoo for it to crunch the data and spew out relevancy rankings, what obviously was more intriguing are two search engines he mentioned – Delver and 123People – both ‘social search engines’ that use the BOSS APIs (?). Somehow, it was not too clear how Delver and 123People benefited from BOSS, as both of them seem to extracting data from a social graph that BOSS does not expose. Well, nor does BOSS expose a link graph and it seems it only returns ranked results.

Vijaya spoke about OpenSocial and its APIs – people, activities and persistence related – and Rajdeep spoke about Shindig, the Apache license open source OpenSocial container implementations. Again, what interested me most was Sun’s work on SocialSite, an extension of OpenSocial. SocialSite also seems to have added some search related handlers which probably searches through the social graph that SocialSite accumulates over time.

Jai from Infosys then went on to talk about Social Network analysis which I was expecting to reveal some more about work that goes behind features like Facebook’s ‘Friends you may know’ or better LinkedIn’s very intriguing ‘People you may know’ feature or Amazon’s recommedation engine which deals with issues like authority, popularity and relevance associated with nodes in a social graph. But somehow, Jai focussed more on theory and on all sorts of tools ranging from manual questionnaires to recommendation engines to data and text mining, stuff that somehow did gel well with Yahoo’s and Google’s talks.

We are going to put up the videos of all these talks up on the site very soon and will also try to put up the presentations made in all the tutorials.

Leave a comment

Your email address will not be published.