Summary Comments from the
Bath Profile Meeting
Boston Spa, UK - - - October 3, 2001
Drafted October 17, 2001
These comments are based on the notes taken during the meeting by Janifer Gatenby (official minute taker), Dana Dietz, and Larry Dixson. The purpose of this document is to summarize and highlight the key agreements and clarifications made.
The meeting was well attended and we worked through much of the ambitious agenda. As chair, Bill Moen made some unilateral decisions as to what agenda items got discussed as the meeting was running out of time.
1. Use of Bib-2 and New Attribute Architecture in Bath Profile (BP)
CENL has a need to mix attributes from multiple attribute sets in a query and for the definition of some searches not defined in Bath (e.g., a National Bibliographic Record Number, National Subject Headings). First, this requires Version 3. This is handled currently in BP at Level 1. Bib-2 gives the best answer for this. But mandating Bib-2 in current Functional Area A is not acceptable.
The agreement was to:
- Put in a statement in Functional Area A, Level 1, "clients and servers should actively support multiple attribute set values in queries within 18 months [or give a specific date such as June 2003]."
- Create a non-normative appendix that shows how to express current Functional Area A searches in Bib-2 attributes. Would be considered part of BP Version 1. There will be a stated cutoff date for current Functional Area A when it will no longer be part of BP. This provides a migration path to support of the new attribute architecture, Bib-2, and the use of multiple attribute sets.
- Create a new Functional Area that explicitly uses Bib-2 and new attribute architecture for expressing all searches in Functional Area A. Would be considered a part of BP Version 2. This new BP Version would be ready by June 2002, at which time the current Functional Area A would be deprecated, and ultimately removed from BP Version 3 (Summer 2003?).
- The use of Bib-2 would provide a way to express searches for National Identifier Number and National Subject Headings.
Lennie Stovel did a quick mapping of the searches to Bib-2 during the meeting.
2. Required Record Syntaxes in BP Functional Area A
There has been continuing concern about the record syntaxes required in Functional Area A. The requirement for more than one MARC syntax means conversion by either the client or server. This is not an optimal solution since not all servers have business cases to support more than one MARC syntax.
The agreement was that:
- In Functional Area A, all levels will require only the support of MARC 21 and make a strong recommendation that clients and servers support UNIMARC.
- The requirement for XML in Functional Area A for interchanging bibliographic records will be removed.
- Servers will be required to support MARC 21 and SUTRS.
- There is some sense that the support of SUTRS by servers is only necessary when the server will not provide records in MARC 21 because of business reasons. If that is not the case, then a server shouldn't have to support SUTRS. We can't say MARC 21 OR SUTRS because that could mean a server could only do SUTRS, and that is not what we want. All servers must be capable of delivering MARC 21 records, and in the case where there is a business case not to return MARC 21 records, servers can return SUTRS. The wording on this needs to be really clear.
- Clients will be required to support MARC 21 and SUTRS.
3. Define New Functional Area for Search and Retrieval of Authorities
There have been calls for including in BP for search and retrieval of authority records (CENL, Barbara Tillett as part of the IFLA efforts). Larry Dixson provided a strawman draft of a new functional area to address these requirements.
The agreement was to:
- Continue the development of a new Functional Area for Search and Retrieval of Authority Records.
- This would be included in a new version of BP.
- Larry Dixson will continue working with input and help from the following:
- Poul Henrik Jorgensen (or his designated contact)
- Juha Hakala
- Rob Bull
Revised drafts of this will be developed and posted for review to the Bath list, with the expectation that an approved version will be ready for June 2002.
4. Author Searches in Functional Area A, Levels 0 and 1
The US National Profile committee proposed changes to the Author Searches as defined in BP. Most of them related to the Author Search - Precision Match,
The agreement was to:
- Move Author Search - Keyword from Level 1 to Level 0.
- Move Author Search - Precision Match on Established Name Heading from Level 0 to Level 1.
- Define a new Level 1 search: Author Search - First Characters in Field (defined as a left-anchored phrase search with right truncation.
Further discussion is needed to clarify whether the Author Search - Precision Match on Established Name Heading and Author Search - Precision Match on Established Name Heading with Right Truncation with the structure "normalized name" go away or exist alongside the newly defined Author Search - First Characters in Field.
5. Type and Format Searches
The US National Profile committee proposed that the current BP searches for Type/Format searches in Level 2:
- 5.A.2.6 Format/Type of Material Search - Keyword
- 5.A.2.7 Format/Type of Material Search - Phrase
could be better expressed as two separate keyword searches using coded values for the query term. This is what the US National Profile will be including. The Type search would be a Level 1 search and the Format search would be a Level 2 search. This proposal was not accepted because of the concern over the codes that are included in the US National Profile searches. The suggestion was to look at ISO 8459-5 and the US National Profile committee will do this and submit a proposal to Bath for inclusion in Version 2.
6. Functional Area B, Holdings
Both ONE-2 Profile and the US National Profile have or will specify holdings retrieval in the profiles. It is likely that these two groups, plus interested parties can develop the specifications acceptable for BP.
There was agreement to:
- In Version 1 of BP, focus on retrieval of holdings records.
- Define one search using the Use attribute 1044, Processing Institution for searching holdings information.
- Do not embed a bibliographic record in the XML record but instead assume there will be two presents; the scenario is a search on bibliographic records, then a present for the bibliographic record and a present for the holdings record. To do this, there will be ESNs defined that excludes the bibliographic record (AcutalBibItem) to do the present for holdings information.
- This does not preclude in future doing searches using bibliographic and holdings attribute (with Version 3) together and returning a single holding record that includes the bibliographic record in one present.
- Members of the ONE-2 (Poul Henrik, Ole Husby) and the US National Profile group (Mark Needleman and Christina Perkins-Meyer), and J. Zeeman and J. Gatenby will prepare a proposal for a more fully developed Functional Area B, incorporating these recommendations, for inclusion in Version 2.
7. Character Sets
The US National Profile committee proposed changes to the BP on the requirements for character set negotiation, and the use of Latin-1 for coding the query. While there were various points of view expressed, there was not clear consensus on BP remaining silent on the Latin-1 requirement.
There was agreement to:
- Develop a proposal for how to do character set negotiation (Ray Denenberg and Bill Moen).
- Leave other requirements for Latin-1 as is.
8. Specifying XML Schemas in BP
Because more than one XML schema is used in BP (e.g., Holdings, Cross-Domain), there is a need to be able to specify which XML schema to use. A proposal to the ZIG for handling this was addressed by the ZIG, and was approved.
As A Result of the ZIG Meeting
One final thing: During the ZIG plenary session on Friday, Bill Moen raised the issue of negotiation of profiles. Ray said that there is a way to do this through the negotiation model that has been approved by the ZIG. Moen suggested that we need to include this requirement in the Bath Profile at some point. And further, we need to label functional areas and levels of conformance with OIDs and it is these, rather than the entire profile, that are negotiated. Moen will be working on this (with Ray) for the US National Profile. Once we figure out how to do this, Moen will propose that the BP take the same approach.
Attendees at the Bath Meeting held October 3, 2001
|
