Having found a seat near an electrical plugin, I’m able to once again take some notes.
I’m just waiting for the session to start. This will be the first I’ve heard Andrew (and Roy for that matter) speak since NCSU launched their Endeca enhanced catalog. Like junkies desperately waiting for their next fix, the room is absolutely packed, no doubt, because there is so much interest in this project. The California Digital Library report (released not long ago) is well worth reading if you’re at all interested in what your catalog and systems ought to be doing.
Here we go. Roy kicks things off with the plea:
The word OPAC should die.
As Roy states, catalogs do a few things well. Inventory control and know item searching within 4 walls are done OK. Where the catalog comes up short is searching for items that are not “known items.” They are not good at unfolding complexity and displaying logical results by groupings (faceted browsing). Relevance rankings aren’t there or useful and its hard for people to intuitively find similar items.
Roy blames “automation in the backroom” for this mess. Back room processes like circulation, acquisitions, and cataloging got moved out into the public view as an afterthought. The systems were optimized for the librarian and not for the user. We’ve failed to be able to exploit the data and resources that we have in our systems.
“We’ve ignored discovery” [this is critical to understand]. We’ve abdicated responsibility to vendors and thrown up our arms. NCSU is one example of a library system that has taken this responsibility back.
Assertions:
A library catalog is one finding tool among many
Some things it does well, some things it does poorly
Users want what is useful – not just what is in your building or collections
Users want a variety of information and information types
Catalogs must become one system that interoperates with others. Let the
catalog focus on inventory instead of smashing in everything. Let other systems
become the primary finding and discovery tools. And yes, there is hope! See:
“The Changing Nature of the Catalog and Its Integration…”
Georgia Pines- good work – looking to replace a commercial system. RLG is another example of interesting work using new tools such as FRBR (I like the way Roy describes FRBR… FRBR helps to cut through the noise of a standard catalog and actually expose the complexity only when the user is ready to see it.) OCLC’s doing some cool things as well [as Lorcan described in his talk]. See also: librarycsum.edu/search/books (no weird cooperative name for a consortium but just a clear message about how to find more items available to them [yes, that makes so much sense]).
Andrew Pace’s turn…
The state of catalogs: “One rarely gets what one deserves; One almost never gets what one never asks for.”
NextGen OPACs [doh… there’s that term]: Vivisimo, Aquabrowser, Endeca faceted search, Innovative’s OPAC Pro, ExLibris Primo, Polaris, SirsiDynix is teaming with FAST (see press release… [if i had Internet I would look this up right now]), and the list goes on…
To illustrate how the catalog can be useful, Andrew did a search of NCSU’s catalog using “vietnam war.” I could literally see people drooling. I’ve played with it quite a bit, what did strike me during his demo, was just how fast the results are returned. It was a live demo and it did live up to the Google speed benchmark. So what does Endeca give them? Speed, relevance ranking, automatic stemming, true browsing, spell checking, “did you mean,” and a number of other things that make me green with envy.
Purchase Decision: very rapid development cycle. Endeca Profind coexists with SirsiDynix Unicorn ILS and Web2 catalog. The Endeca indexing (which doesn’t require down time) will be cut down with tweaking to about 3 hours by next month (currently takes 7 hours).
Endeca doesn’t understand MARC records. Endeca discouraged them from changing records into XML in favor of flat files which are easier to index. 10 dimensions were taken from the MARC record and displayed through Endeca.
NCSU experimented with different interfaces, but ended up going with a very simple interface. they borrowed alot of ideas from other Endeca clients, Amazon, and other similar search services. NCSU is working on FRBRizing it more to make viewing editions of works easy to understand. Talking with OCLC about faceted subject terms. Looking at allowing patrons to add their own dimensions to records. They’ll add some bling (like book jackets). Continual tweaking. And more.
Usability Testing & Stats: Andrew ran through some interesting stats which I didn’t record, but he thought they would be publishing some of this in some fashion, sometime soon.
“The catalog is just the first step…”
