Stephen Arnold gave a fascinating presentation, showing some new interfaces for search engines. He described three stages of the evolution of search, the brute force era of 1976-1998, the social+analytics era from 1998-2006, above all the period of Google and measuring inbound clicks, and the future, to 2010, of compound processes, that is to say putting queries in context, using algorithms, knowledge-bases and intelligence He suggested that products like Autonomy which use neural networking will be important . Voice and video (conversations etc) are the richest area of information and are currently not indexed or organised at all.
He expected there to be much more non-objective searching, as people try to "steer" hits. There will be a big period of wheel-reinvention as the new search tools discover the limitations that Dialog and others found in 1980. He mentioned promiscuous searching systems as holding some nasty surprises, but I'm a bit vague as to what a promiscuous searching system might be.
New things for 2006: tag, comment. Microsoft and Yahoo have both "bet the farm" on social software. Words are no longer enough: images, data. Faceted search is becoming more common. Metrics: data analysis and maths will become more important, and librarians need to be able to understand them.
Complexity increases, concurrently with development of Google. Computational cost also increasing (bandwidth, memory disk space, computational time) but precision/recall are increasing much more slowly.
Amazon had to discontinue personalised search as their technology costs go up like a rocket. Search is expensive, so the first to be sacrificed. Yahoo's costs are also out of control.
Then he showed us a series of search services:
Podzinger: indexes spoken podcasts in English, funded by the US Defense Advanced Research Projects Agency, DARPA, one imagines for surveillance purposes.
Exalead: offers a baby view of screen, and related terms
Mondosoft, who developed a search engine for the Vatican site hard wired to the Catholic catechism.
There's http://rwsm.directtaps.net/ a faceted search tool developed by Microsoft,
He also showed us a new interface to Google, which allowed one to refine searches, which I can't now find, and Windows Live and Alltheweb
His conclusions were that the new search comes with enormous costs, that technology is driving search again, that new systems are considerably more complex, that multiprocessor chips are coming and have many implications for search, that there is huge confusion, hubris and fear uncertainty and doubt in search sales.
When one company controls 65% of search traffic, there is a burning need for objectivity. He called on professional associations (he said the SLA, but it is equally valid for CILIP) to get involved with these issues.
Technorati Tags: ILI2006, internetlibrarianinternational, search


Recent Comments