Z39.50 Implementors Group meeting
April 4 and 5, 2002
OCLC, Dublin, Ohio, USA
compliments of Lennie Stovel
Updated May 27 (See below)
3.1.1. Overview – Ray. see PowerPoint document
Ray also described MODS (Metadata Object Description Schema). Rebecca Guenther at LC is the link between MODS and Dublin Core Libraries (but there is no discernible relationship).
3.1.2. Applications – Pat – see PowerPoint document
3.1.4. Schemas/wsdl – Poul-Henrik – see PDF document
3.1.3. CQL – Ralph – see html document
3.1.5. PICA Implementation – Jan – see PowerPoint document
3.1.6. OCLC Applications – Ralph -- see powerpoint document
OCLC is deploying an srw server for an instructional management system digital repository, which is not real yet. They will do updating as well. The metadata is being harvested using OAI to an srw server accessed through an MIT interface.
Ray re ZING in general: ZING has components in three categories.
Bill: there are new markets looking for search and retrieval, for example, OAI; srw might be useful.
Discussion: how to get the word out through a variety of avenues.
3.3. ZOOM – Rob Sanderson – see html pages
3.2 Explain [Zeerex] – Rob Sanderson – see html pages
This looks like an app for OAI – don't bother with search, just put an XML blob in a file system. But do you want all the explain records at once? Maybe. Also, OAI software is open source.
Someone else could harvest the records and make them available for searching.
Should the records contain geographic information about the location of the server? They could.
The collection is what is interesting, not necessarily its Z-ness; collection-level information is missing from Zeerex.
You could use EAD to describe collections or use the description element in Zeerex. You could include an EAD in an explain record because XML is extensible.
You need to have the thing you're searching on in the record. For example, "authoritative" should be present in the record.
A record creation tool would be useful.
Who will use Zeerex and what for? Zeerex exposes the hard problem of helping a user pick a database. It could be used for automatically configuring a search for local holdings. Integrators could collect the records (this has worked so well in the OAI context – not).
You need to be able to generate HTML for human consumption as well; this could be helped by style sheets.
Rob also described Distributed Virtual Databases. You scan an index from beginning to end to get all the terms, then you put them into a second database which is searchable and linked to the first database. This is nifty but not scalable.
4. BIB-2 – Barbara Shuh – see PowerPoint document and agenda documents
There need to be complete examples. There also need to be complete examples in the Bath profile.
The expansion/interpretation attributes need to be recombined with the format/structure attribute "set of words" and moved to the Utility attribute set. Ralph will draft this.
A suggestion to get rid of right truncation on word and character boundaries was withdrawn.
In the mapping of Bath profile searches to the Bib-2 attribute set, the Bath keyword searches raised the question whether we need a "word" attribute to denote a single word as opposed to a set of words that could be one or more words. The consensus was that we do not need it.
The presence of both "set of words" and "right truncation on term" in the same search was a topic of discussion. Ralph's changes will include the ability to specify right truncation on each word in a set of words.
There is an outstanding work item to look at Scan in the Bath profile in terms of Bib-2.
The Bath meeting following the ZIG meeting will discuss migration to Bib-2, and the standard number searches in the Bib-2 mapping.
2. ZIG/Maintenance Agency Business
The NISO committee working on the US national profile for Z39.50 identified some perceived weaknesses in the Holdings Abstract Record Schema, and one error. Because the correction to the error could disrupt existing implementations, it was agreed that the new fields would be added but the old element would not be dropped.
The second proposal also recommended deleting an element; this part of the proposal was withdrawn. A more formal proposal for these two changes will be presented to the ZIG.
The remaining changes to the Element Set Names were acceptable and will be implemented by the Maintenance Agency. At the same time, it will be clarified that the reporting levels listed in the document are intended for use as element set names.
2.3. Element Set Names – Slavko – see agenda document
Since there was confusion about the element set names "f" and "b", which appear in Z39.50, and since profiles were using these names but defining them differently, some guidelines for their use were proposed. Discussion also covered the use of element set names to specify a schema, which could be a namespace identifier. It was agreed that profiles should not mention "f" and "b". The ZIG could also recommend that schemas be carried as element set names.
The topic of a registry of elements set names was discussed and dismissed. It was suggested that the Maintenance Agency could list element set names in a central place if needed.
2.6. Proposal for a Z39.50 Present Request parameter: compSpec-2 – see agenda document
Based on the discussion of item 2.3, this proposal was withdrawn.
2.2. Proposal to revise the Z39.50 Character Set and Language Negotiation Definition – see agenda document
This proposal was accepted.
6. Reports and Updates
6.3. NCIP – Pat – see PowerPoint document (technical difficulties prevented the ZIG from seeing this)
The final document will go out for final review by the committee on Monday. It is hoped that it will go to NISO ballot by May 1. The committee will be holding Placeware meetings for NISO members.
NCIP is a framework with three components. Only the first two are being voted on.
Each application profile would have to specify how to use Z39.50 and NCIP together, if needed.
Holdings are an area of overlap between NCIP and Z39.50 profiles. The area needs more discussion. NISO may be able to bring together a group for this discussion.
2. ZIG/Maintenance Agency Business, continued
2.4. Z39.50-2002 Revision status – Ray
A draft is available. Ray expects it to go out for ballot soon. The draft includes three items requested by ZIG members at the ZIG's October 2001 meeting.
ZIG members who are NISO voting members can submit additional editorial changes with their ballots. Any issues raised in the balloting might be brought to the ZIG.
2.7. Proposed change to Truncation attribute 104 – Ralph LeVan – see agenda document
A counter to this proposal appeared on the www-zig list. It is more complicated, and it resolves the issue of searching on a question mark character. Ralph's proposed change was approved.
Following the meeting this decision was reconsidered. See http://lists.w3.org/Archives/Public/www-zig/2002May/0042.html
2.5. Proposed Implementor Agreements
Ray led a discussion of implementor agreements in general. All existing agreements have been incorporated into the new revision. Ray proposed a new definition of implementor agreements: a spec to which a profile could point. This would help avoid having similar specs in different profiles. The suggestion to call these "reference documents" instead of "implementor agreements" was approved. Thus, there are three categories: contributions, reference documents approved by the ZIG, and implementor agreements approved by the ZIG as the best way to implement something.
Diagnostics: Suggested User Display and Message Logging Using Bib-1 Diagnostic Set
This is a useful resource for implementors, to help them compose messages for users rather than using the wording inside the standard, which is intended for implementors. It is not a document the ZIG agrees to as an implementor agreement or reference document. It will go on the Maintenance Agency site as a contribution.
Creating a Search from Scan Results
This document will be posted as a contribution for now. It may come back to the ZIG for approval as a reference document.
2.8. Future meetings
A meeting has been tentatively scheduled for October in Helsinki. It is not clear if that location will draw more people from Europe. The major anticipated agenda item is possible resolution of ballot comments. A tutorial would draw people from many Nordic and Baltic countries. The suggestion to not schedule another meeting right now, but to revisit the schedule in 5 or 6 months, was considered favorably.
An appealing alternative could be a conference on library protocols, covering other protocols along with Z39.50, and including a tutorial and a ZIG meeting. (People remembered EFILA, the European Forum for Implementers of Library Applications; perhaps it is appropriate to re-invent a similar organization.) Such a conference could be held in Helsinki in the March or April timeframe.
6. Reports and Updates, continued
6.1. Bath Profile – Carrol
The group is working toward a new version that will include significant new functions, for example, authorities and holdings. The group's meeting followed the ZIG meeting
6.2. NISO SC AV Profile – Bill
Functional Area A is close to ballot, following one last review of the changes made at a meeting that preceded the ZIG meeting. Functional Area B, Holdings, may become a Draft Standard for Trial Use by late summer.
CNL Profile – Juha
CNL has decided not to continue to develop its profile separately. It has endorsed the Bath Profile. It has successfully proposed changes to Bath to support some needed extensions.