The Wi-Fi seems to have recovered from yesterdays problems, so back online for the second day of the conference.

Yesterday afternoon I saw a couple of very interesting poster sessions:

From Germany, using the FAST Data Search engine to create an alternative search and retrieve interface for their 11 million bib records (to be expanded by adding other similar sized collections). I was struck with the user experience – much more like Google of course – but dispensing with the idea of the ‘OPAC’ – perhaps a concept that is well past it’s sell-by date? The other thing that I was really impressed by was the ability to do ‘drill down’ into the data set retrieved by various facets, including authors and subject headings. The speed at which it was able to create this faceted browsing, on the fly, for large datasets (e.g. 16,000 records retreived in 0.42 seconds, with facets ready to use). The other thing to note here is that FAST is a well established tool which is out there, and ready to use. Let’s stop trying to develop library based search, and use these excellent tools that already exist. An ideal candidate for this approach would be COPAC – the interface is not brilliant and the data set it large enough to benefit.

The second session was from Austria. The session itself was showing how they were integrating digitised material with their catalogue using open source software (notably swish-e, an open source search and retrieval tool, not unlike FAST above). However, what I was more interested in was that this demonstration suddenly brought home to me the possiblity of scanning in TOC from books to enhance the metadata available to the searcher. So far we have thought about digitising items to make them available electronically. However, scanning ToC to make them more retrievable wasn’t something I’d considered. If we were to scan and OCR all contents pages from our teaching collection, I think we would see a real benefit for the users (we regularly have examples of people failing to find material because they are searching for the chapter title or author rather than the monograph details. Of course, using ToC is not a new idea, I just hadn’t thought of the use of digisation as a way of obtaining this data.