Cataloging and User Experience

-What users want

-info changing at umprecedented rate, but we’re not changing fast enough to keep up

-OCLC 2003 enviro scan; info conusumer trends: self-sufficiency, seamlessness (liesure/work overlap) – libraries are not seamless!

-2005 OCLC college student survey; satisfied with what they find through our search

-2009 OCLC what libs/users want; users=seamless flow, immediacy, enhanced content (TOCs, etc), advanced search options for narrowing, but w/ one box, expectations are based on popular sites

-Info now, in one place, and don’t want to have to go to lib to get it (but we’re not entirely digitized, so not there yet)

-Challenge=make catalog wrk for all

-

 

-Netx gen

-features:keyword search, relevance ranking, faceted search, search limits, did-you-mean, item recommendations, RSS feeds, user feedback, cover art, TOCs, reviews, tagging

-Endeca: works w/ existing ILS, simple, relevance and faceted, NCSU using

-VuFind: Open source (Villanova); faceted, author bios (browse search?)

-WorldCat Local: Does not sit on top of ILS, searches WC; single search, can link to ER, multiple languages (U IL-Chi)

-eXtensible catalog: still in development at U of Rochester (rollout in 2010); open source; will be discovery layer as well as converting existing MARC to XML to help prep for RDA’s scenario 1

(she didn’t mention Koha)

-If it metadata isn’t right, none of these will work right!

 

-RDA

-1997: AACR3 began to be talked about due to all the new e-formats, but work didn’t begin until 2004, so even more formats!; quickly realized that print-based ideas wouldn’t work; RDA started in 2005

-online product scheduled for release in 2010 w/ testing, eval, and training by national libs after release

-principles-based w/ more left to catalogers’ judgement (less ridgid rules); new formats shouldn’t flummox catalogers any more;

-content standard, not display standard (MARC, XML, whatever) w/ focus on user

-strusctured v. diff from AACR; still describing items, just adding other things; subjects, concepts; how to connect works, manifestations, etc., with subjects, people, and concepts

-www.rda-jsc.org/docs/5sec7rev.pdf

-

 

-FRBR

-user tasks: find, identify, select, obtain

-RDA is shaped by conceptual framework of FRBR; way of showing relationships to a greater extent than we have in the past

-Work (story, song, etc.), Expression, Manifestation, Item

- challenge is to take use from any expression or manifestation to all the related ones

-writers of FRBR and RDA didn’t dictate how to do it, so it’s up to the softwarre developers (only one company, in Sweden, that has developed a FRBR-based program)

-we currently catalog at manifestation level—wil be moving to datasets for each work, then another for expressions, and another for manifestation, but they’ll link in such a way that they’ll feed in info so things won’t have to be redone each time.

 

Three RDA scenarios

3: “flat file” = OPAC w/ no linking (card catalog)

2: linked bib and authority records (like we have now)

1: relational/object oriented db structure; linked datasets; new infrastructure w/ info based online (future); will we have the infrastructure, will linnked data always be available, “will someone change it like wikipedia?” (URGH! Haven’t we gotten over this yet?)

 

Prepare

-catalog w/ RDA in mind; note rlationships; use authority records

 

Q&A

-we currently spend a lot of time doing authority work; will it be more or less important in scenario 1? More important! But less work because the effort will be more spread out. (LC’s lack of subfield d codes, which they had quit doing to save time but are now going to want to get RDA to work properly)

-Is FRBR more a program than a database? She doesn’t know, doesn’t understand semantic web very well, nor metadata. But the metadata people think that this will work. Q’er is worried that will have to redo work; P’er says programs should be able to do the work to convert MARC to XML

-More on user tags in nextgen catalogs? Can you search by them? Depends on the catalog, but they should be set up that way.

[example throughtout was P&P and various off-shoots]