ALA Annual New Orleans 2011
FRIDAY—June 24
8:30 AMMorial Convention Center – Room 263
ALCTS Preconference
What IS it, anyway? Library of Congress Genre/Form Terms for Library and Archival Materials
Friday, June 24, 2011, 8:30 a.m. – 5:00 p.m. Morial Convention Center, Room 263
This preconference will introduce participants to Library of Congress Genre/Form Terms for Library and Archival Materials (LCGFT). It will focus on the correct application of LC genre/form terms to works, with an emphasis on the differences between genre/form terms and subject headings. Participants will also be briefed on the principles underlying the genre/form projects being undertaken by LC’s Policy and Standards Division, and plans for further development of the thesaurus.
Speakers: Janis L. Young, Senior Cataloging Policy Specialist, Policy and Standards Division, Library of Congress; Yael Mandelstam, Head of Cataloging, Fordham University Law Library, Fordham University; Beth Iseminger, Music and Media Cataloger, Harvard College Library, Harvard University
NOTES
SEE HANDOUTS
Genre: category of works that is characterized by a similar plot, them, setting, situation, and characters
Form: work with a particular format and or purpose
$v are form (not genre) subdivision
One reason for creating the new file is that authority records are being created for validation during cataloging. They are also trying to be consistent in syntax. They want a mechanism that provides for regular updating, e.g. GSAFD hasn’t changed 10 years. They are incorporating terms from RADFG, MIGFG, MIM, a cartographic list. LC wants fewer files to need to consult. If there are too many used, it is harder to make sure which thesaurus you used each season; if you are consistent, it inhibits collocations, causes user confusion, and slows catalogers down. Many terms will stay in LCSH because there are books about materials in a genre. But others, like those in music just need to be changed.
LCGFT has no subdivisions.
Form is often used in facet displays – it’s better to define terms to be used (rather than leaving it up to vendors to fabricate them inconsistently)
Law will have about 80 new terms.
Music will be a wholesale change. Medium of performance doesn’t belong in LCGFT.
H1790 – individual works of fiction; it will change drastically
Assumptions and policies: Terms describe the expression, not the manifestation. a) catalog the intellectual or artistic expression, not its physical carrier (VHS, DVD, electronic, cassette, etc.) b) allow for collocation of all manifestations of a work within and across catalogs
Motion pictures, American is not valid for American films because the intention was to use it on film with foreign speakers made in America but intended to be sold in the foreign country.
The map project was approached with “post-coordinated” as a philosophy. E.g. Topographic maps is a heading and the place has $v Maps.
In LCGFT each term represents one facet – don’t overlap with each other. They are not subdivided. They are trying not to repeat data already existing elsewhere in the bib record. Therefore, with 041, 546 etc containing language, the genre shouldn’t have the language. Ed O’Neill suggested that the list ought to be able to stand on its own and not depend on the bib record (I agree). Genre/form terms and subject headings are symbionic; each elaborates on the other.
Feature films is anything more than 40 minutes. Shorter are Short films.
LC hasn’t decided how to approach current LCSH headings that have 360 to use them as a genre – e.g. Artists books.
Broadest term is authorized but generally not be assigned – the 360 says for broad collections of…e.g. Motion pictures.
Apply as many LCGFT terms as necessary to bring what a work is; apply LCSH as usual.
Most genre terms are for fiction; not usually for non-fiction.
Feature films – always apply to fiction or non-fiction. Fiction films or Nonfiction films.
Consider original production/showing to decide if to include “television programs”
Reporting or broadcasting of an event is nonfiction.
LC is using Internet videos for TV shows redistributed online – probably be changed.
3:30-5:15Morial Convention Center– Room 243
CCS Forum
Turning Catalogers into Semantic Web Engineers, or, Out of the Catalog Drawer and onto the Internet Highway.
The Semantic Web often enters into discussions about the future of cataloging and the purpose of RDA.To meet users’ expectationscatalogers must become Semantic Web engineers so the library catalog will get out of thedrawer and onto the Internet highway. Speakers will address issues to consider as we approach the entrance ramp.
First Session:
Bibliographic data in the Semantic Web – what issues do we face in getting it there?
Gordon Dunsire -
Library catalogues use the bibliographic record as the unit of metadata, while the Semantic Web focuses on the metadata statement of a single bibliographic attribute. The presentation will discuss some of the issues arising from this paradigm shift, as well as the technical, legal, and economic barriers likely to be encountered in transforming legacy records into library linked data.
Second session:
Encoding bibliographic data - The Things and Strings of the Semantic Web
Karen Coyle -
She will discuss moving bibliographic data toward linked data on the semantic web. In “Understanding the Semantic Web: Bibliographic Data and Metadata” she noted libraries’ future “must include the transformation of the library’s public catalog from a stand-alone database of bibliographic records to a highly hyperlinked data set that can interact with informationresources on the World Wide Web.” In RDA Vocabularies for a Twenty-First-Century Data Environment, she noted “With Web-based data, we can use the vast information resources there to enhance our data by creating relationships between library data and information resources. This will increase not only opportunities for users to discover the library and its resources, but also the value of the data by allowing its use in a wide variety of contexts”
Third session:
RDA and the Semantic Web.
Ed Jones -
Does RDA play nice with the Semantic Web? If so, how? What RDA elements should be linked, and at what level of granularity? Some MARC fields might not lose information if they were treated as single elements, while others might benefit from greater analysis and control at a more granular level. If something can't be controlled, is it necessary to identify it separately?
-=-=-=-
David Reser from LC came to me to congratulate UND on following the RDA test as they expected (which apparently others had not done)
4:00-5:15PMHilton Riverside– Grand Salon C
FAST Subcommittee
As previously reported, the development of FAST is completed and our focus is now on documentation, maintenance, enhancement and supporting new applications. FAST: Faceted Application of Subject Terminology: Principles and Application was published by Libraries Unlimited in July 2010.The book’s 354 pages includes four parts; (1) Introduction, (2) FAST Headings, (3) Application of FAST, and (4) Authority Control and Implementation. Previously, the lack of detailed documentation had been an impediment to FAST implementers.
The major maintenance work being done on the FAST authority file in 2011 included:
- Adding of additional references to the FAST authority records,
- Removal of duplicate authority records,
- Identification of incomplete and poorly constructed authority records,
- Validation of linking fields,
- General error identification and correction.
An online service was introduced which will add FAST headings to bibliographic records with LCSH headings or convert a set of LCSH headings into FAST headings. A file of MARC records can be uploaded, FAST headings added, and the update file downloaded. This service is intended to demonstrate the conversion process and to illustrate the relationship between LCSH and FAST headings. It was not designed to process large files which can be processed much more efficiently in a batch process. The conversion service will also convert either a single LCSH heading or a set of LCSH headings into FAST headings. There is a link to the conversion service on the FAST home page:
The interface for mapFAST was substantially revised, improving both usability and performance. In addition to identifying materials in OCLC's WorldCat, materials in Google books can also be accessed. The underlying functionality of providing FAST headings with geographic coordinates is also available as a Web Service. is available at: There is a link to mapFAST on the FAST home page.
A major development effort focused on enhancing access to the FAST authority records. The database engine employed since the inception of the project is dated and needs to be replaced. Linked data has become an essential means of providing access to bibliographic data and we committed to making FAST available as linked data. Rather than view these as distinct efforts, we decided to build a new search engine built on linked data. That effort is nearly complete and is being tested internally. The new search engine has a much improved interface with added functionality including auto completion and other enhancements. If testing continue to go well, a beta version should be released in late June.
The total number of FAST authority records has changed only slightly. The FAST authority file contains about 1.65 million authority records. Some new headings were added and a few duplicate and erroneous headings were deleted but most of the activity focused on adding (1) geographic information, (2) references, and (3) links to LCSH authorities. While is still a time lag between FAST and LCSH, the lag has been reduced. The full FAST authority file can be licensed for non-commercial use.
The FAST team continues to explore collaborative effects to improve FAST. We are very interested in working with the community to further enhance FAST and continue exploring ways to allow the community to add and/or correct information in FAST records. The FAST database now includes a link in geographic authorities display to allow user input of the coordinates. We are considering additional options for augmenting and/or correcting FAST authorities.
Please address any questions or comments to Ed O’Neill at .
7:30-9:30PMMarriott Convention Center– Room Blaine Kern E
CCS Executive committee
NOTES
Kristin Lindlan, ALCTS Planning wanted us to be aware of the Critical issues document forming the ALCTS Strategic Plan. It was discussed briefly and returned to later in the meeting.June Abbas, RDA Update planning, reported that pre-conferences have been successful. There are 4 forums planned at Annual involving vendors, educators, update, etc. The Task Force should continue, at least the update forum for Midwinter and Annual. Maybe it would make sense to cooperate with the IG’s – Dale Swenson. Steven wondered if a session on authorities and FRAD and/or NACO. None of the NACO coordinators will have had any training on how to do authorities with RDA. Shelby noted that Kate Harcourt of the other RDA task force thought the task force should continue. Shelby noted that the appointments all expire. The Chair-elect will contact the chairs for advice on continued membership and appoint new people as needed. De-mystefying FRAD is Qiang’s book soon to be published by ABC CLIO. She has one chapter on RDA authorities and the entity relationship model. We discussed the strategic plan. We thought we could address IIB Identify and implement best practices for encouraging virtual participation and IIC Utilized technology to incorporate virtual members into live, in-person meetings. We discussed changing our name. We first decided we should, next we decided “cataloging” should remain, and finally decided on metadata management. Therefore Cataloging and Classification Section (CCS) will be Cataloging and Metadata Management Section (CaMMS). We discussed several items to work on and George will send his notes to Mary and me.
.
SATURDAY—June 25
8:00-10:00Hilton Riverside – Kabakoff
Technical Services Managers in Academic Libraries IG
Table discussions – notes will be on CONNECT. Our table led by Bruce Evans (Baylor) talked about: Core competencies for all Cataloging hires. Are catalogers and metadata specialists the same? Do catalogers need programming skills? How are we really able to catalog hidden/unique collections? Minimal level cataloging? Acquisitions merged with ILL? Or with campus computing?
10:30-12:00Morial Convention center– 393-4
RDA Porgram Update Forum
Transitions, transformations, shifting sands
Gordon Dunsire
Triples can link to other triples. Each must have an identifier for subject (subject predicate object). We are moving to individual metadata statements. You don’t need to copy a “record” to add an additional piece of information. However we need to be able to reassemble the statements as needed, usually via an identifier. We have statements instead of records. E,g the record is b1234.
Note: anything with a number is a URI and anything with quotes is a text string
b1234 title “cataloging is fun”b1234 author n8765
b1234 content type t986
t986 definition “content expressed” – this could be metedataregistry
t986 preferred “text” – this is a display label
n8765 place of birth p9876
p9876 name “Edinburgh”
b1234 title “cataloging is fun” -- b1234 author n8765 -- b1234 contenttype t986. Then another triple is: t986 Definition “content expressed” The display label is -- t9876 preferred “text” -- n8765 Place of birthp9876 – then -- p9876 Name “Edinburgh”. Then user generated tags might get added that are either useless or erroneous.
Metadata ecosystem: links to statements created outside of the prof library community - other communities archives, publishers; machine-generated statements can be created. The test of truth is provenance – who said that statement. We can use statistical inferencing to throw away the garbage (weird user tag) 0 sort of like Olympic skating scoring throwing out top and bottomand keeping the middle – wisdom of the crowd. We have lots of legacy records that need de-duplicate which can be done with “same-as” links. We have lots of players in the ecosystem – ISBD, IFLA…. Then we need to make choices on types of statements and lets re-use what we have, e.g. dct:title frbr:titleproper. NOTE: from Karen’s yesterday – strings of text are not computer actionable. The computer needs data as in the triples. 500 MARC fields are particularly horrible and very ambiguous so no computer can convert them to anything else.
Glenn Patton
OCLC is working on mapping from ONIX to MARC with ver. 3. OCLC currently can convert records from one type to another but they processes through OCLC MARC. They can bring dc-qualified, ONIX, MODS, Dublin Core, MARC to any of the others. They used MARC since most users wanted it and it is more detailed than the others. The problem is concepts are not equivalent, e.g. 100 MARC definition is not the same as ONIX creator. MARC 100 has lots of other subfields that don’t go anywhere in the others. MARC records have redundant information and data in some fields is ambiguous. So now what with RDA? RDA doesn’t fit well in MARC and cramming it in may be more difficult than we want to deal with.Example: ONIX records that an item is electronic deliverable by download in Kindle format, but the MARC just says “electronic resource” with some notes (that aren’t actionable). RDA relationship designators don’t fit well in MARC though we could say $i electronic reproduction of (work) in a 700, 533, or 776. OCLC is looking at RDA or other structured metadata vocabulary as the hub for crosswalks. They think they can remove MARC dependency and utilize other mappings and not have to keep having to “refresh” legacy data.
Jennifer Bowen
Extensible catalog
She believes we will be able to get beyond MARC by 2013 for some work in libraries using a FRBR-based metadata structure that incorporates some RDA elements, but we will be in limbo.The Extensible Catalog (XC) will be able to use both. It is open source: it has a discovery interface and a set of tools to manage metadata and build applications. Discovery interface has a search box and facets with subject, authors, format, language, date, etc. Record displayed at same level as ODIN standard. Behind it is XC schema with 3 separate linked records in XML using Dublin core based elements – XC work, XC expression, XC manifestation. They can transform and MARC record in their ILS into the 3 linked records automatically. They take the MARC Holdings record which is linked to the MARC bib and create an XC holdings record. Everything is linked. The XC schema is using the RDA element sets and vocabularies in the metadataregistry, which are to be completed in 6 months. They have other elements which they have also registered.
“Transforming our bibliographic framework” – XC has unique identifiers [missed]. XC provides: MARC data which can be handled in bulk, has discovery interface, uses subsets of RDA elements and roles, and can show primary relationships between work, expression and relationships. Hope XC can provide a bridge from MARC to RDA.
Jenn Riley
Speculating on the future of the metadata standards landscape. Data structure (MARC, EAD) data content (CCO, DACS, AACRS), data format (XML), data exchange (OAI, Z39.50 SRU/SRW). RDA is somewhere in between data structure and data content. DCMI Singapore framework is a new application profile model (CHECK) or Our discussion are made difficult by different understanding words. Vocabularies is “properties” to DCMI but a controlled vocabulary in OCLC is a vocabulary.