Getting vertigo retrieving information

Last week I went along to the ISKO UK seminar/event on Information Retrieval (IR) held at University College London.

Brian Vickery gave a talk about the first fifty years or so of IR.

Like any good event, I came away with loads to ponder. I’m still pondering some of my notes (I wish my handwriting as neater…)

Stephen Robertson of the Microsoft Research lab talked about where search was beginning to go and what was being explored by companies such as Microsoft and Google.

During the Q&A session, there seemed to be a theme of users using Google and Yahoo et al for quick references, such as three word searches. To some extent this is probably for the ease of the interface but…

What I really got out of this was an intriguing thought.

Google et al are good at general searches. They can find vast amounts of data quickly and easily and presnnt them via the algorithm to the user. Their search is horizontal.

Yet repositories can contain vast amounts of better search data than the search engines can create using controlled vocabularies, RDF, RDFa and so on. They have people writing the classifications for them who know the subject well (we hope) and can make more rational judgments than a search engine. So if repositories and data stores could come together and leverage their inherantly more detailed vertical search via XML and RDF interfaces to link to each other and allow the search engines rapid access to the relevant data. Their search is vertical.

The experts in the field will probably already know where the relevant stores are but casual and non-acadmic users will not. Nor are they likely to take them time to delve through advanced searches. We are time pressured. The vertical search engines may well not have the resources as our large search friends but a few adaptations should allow better access and also lever the knowledge into a more public sphere.

No Comments

Leave a Reply

Your email is never shared.Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.