SLA2009: Unified Discovery Services

(came in late because the speaker I initially chose to see failed to show up)

A speaker from Serials Solutionâs Summon reviewed various pieces of research done recently both by LIS researchers and by big libraries. Summed up pretty well by Tenopir (he copied her graph), increasing costs, decreasing importance as a gateway for research creates a value gap for the library.  Scholars view the library and its resources as reliable and authoritative, but painful to use. So they start with google, because itâs easy and then link out to our subscriptions and my never know theyâre using library resources. Thereâs a misunderstanding of the library as a gateway.

Second speaker â about Summon. The difference between this and, say, MetaLib, is that this spiders and caches whereas MetaLib is a federated search.  That is, it goes out to the various targets as the search is executed.  This can take OPACs, (real time availability) courseware, entitlements management (like if you have an ERM), plus all sorts of A&I databases and electronic journals.  They have 100+ providers including 400+databases, etc. (seems like this is very dependent on on going cooperation with the content providers, but theyâre owned by one of them, so you have to wonder if this could be a weakness?). It has open APIs. Itâs hosted. It will be available next month.  Itâs live at Dartmouth if you want to play. Popular refine â full text available. Title provides OpenURL link.  They might have indexed the f/t at Proquest, but that library gets it through Ebsco or L-N â it links according to your âentitlementsâ so you go to the appropriate copy⦠All kinds of advanced search with proximity operators and all. The point is that is looks easy to the user but it does a lot of fancy things on the backend.  (no doubt!)

Someone asked about relevance, in general â they have their own method (proprietary, tweaked, based on TF-IDF)⦠I asked if relevance can be customized by institutions based on their collections/needs/users â he gave a little crap on how complicated it is, and basically said no, not planned, but they might come up with several different schemes or settings for different types of libraries. 

Another question was about the licenses with the content providers to participate. Sort of a wishy washy answer but they have had a lot of takers, I guess.

Q: I work in a pharmaceutical company. What are your plans for integrating local documents? A: They do that for universities with unique special collections or IRs, that arenât shared with other universities.

Q: I work in a law firm â we have billable and non-billable resources. Can it work with that? A: that can happen at the linking to full text basis. There are other things that are statistical or data set type things, so those are integrated a bit differently and they handle them individually.

Q: usage stats? A: we can and do offer counter compliant stats, but they wonât really give you enough information so we have more detailed user analytics

Q: can you drill down to a specific db? A: may work on that

Q: pull out consortial holdings? A: if thereâs a consortial union catalog â easy, when itâsâ more complicated, theyâll have to address that depending on the specific questions.