Larry,
Thanks for your interest in our project and for this clearly very pertinent question. My first response would in fact be another question: improves it compared to what? One of our motives for adopting FAST was the wish to use a controlled vocabulary in place of the keywords we had previously used in our minimal-level cataloguing. It's generally accepted that controlled vocabularies improve retrieval compared to uncontrolled terms, and FAST seemed well suited to our need for a subject vocabulary that was compatible with LCSH while being simpler to assign.
A second motive was to solve some of the well-known problems associated with breaking up LCSH strings for use in facets, such as split entries for geographical headings and subdivision terms that lose meaning out of context. We felt those were problems worth solving. We don't in fact have any research to quantify the impact of FAST facets on end user retrieval relative to LCSH, but I think there's a good reason for that: to my knowledge there hasn't previously been a full-scale implementation of FAST in a faceted catalogue, so a real world comparison hasn't been possible until now. This could be a good topic for a research study.
I've already mentioned a third reason for adopting FAST, one that is closely related to the first. As a fully enumerated scheme it provides us with IDs for all of our subject terms. The theoretical case for using those identifiers is very compelling, but the empirical case for it has yet to be made. A number of linked data projects are under way, such as LD4L, that will try to make that case.
To answer Amy's question, we solved the problem of cluttered displays by configuring our system to show FAST alone in the facets, and LCSH alone in the individual record views. We show FAST in individual records only if no LCSH are present.
--
Chew Chiat Naun
Director, Cataloging & Metadata Services
110D Olin Library
Cornell University
607 254 8031
________________________________________
From: Program for Cooperative Cataloging <[log in to unmask]> on behalf of Laurence S. Creider <[log in to unmask]>
Sent: 24 June 2015 18:38
To: [log in to unmask]
Subject: Re: [PCCLIST] FAST implemented in Cornell library catalogue
Very interesting. Do you know whether there is data that indicates that
FAST headings improve the ability of users to retrieve the material they
want or need?
Thanks,
Larry
--
Laurence S. Creider
Head, Archives and Special Collections Dept.
University Library
New Mexico State University
Las Cruces, NM 88003
Work: 575-646-4756
Fax: 575-646-7477
[log in to unmask]
On Tue, June 23, 2015 2:35 pm, Chiat Naun Chew wrote:
> Since Cornell began its FAST project early last year we've been gratified
> by the interest shown by colleagues at other institutions. I'm pleased to
> announce that we now have FAST implemented as the faceting vocabulary in
> our production catalogue. Please try it out for yourselves:
>
> https://newcatalog.library.cornell.edu/
>
> We feel that FAST offers noticeably cleaner and more intuitive faceting
> than LCSH, something that is especially noticeable in the geographical
> facet. Equally important, each FAST heading is associated with an
> identifier and thus helps lay the foundation for our ongoing linked data
> work.
>
> In addition to FAST the new catalogue has a number of other notable
> features, including what we think is a particularly well-implemented
> browse feature.
>
> It's been hugely rewarding to work with the OCLC Research team on this
> project. A special vote of thanks to them, and to Cornell's Discovery and
> Access team.
>
> If you'd like to know more, our presentation from ALA last year is a good
> place to start: http://connect.ala.org/node/226498.
>
>
>
> Chew Chiat Naun
> (on behalf of the Cornell FAST team: Gary Branch, Steven Folsom, Sarah
> Ross and Ardeen White)
>
|