Z39.50 Implementors Group Meeting Library of Congress Washington, D.C September 13-14, 1990 1) Introductions Introductions were made. 2) Purpose of Group Mark Hinnebusch restated the purpose of the group for new attendees. 3) New Elements Sets Issue was deferred since item came from Thinking Machines Corporation and no one from there was attending. 4) Z39.50 Version 2 Ray Denenberg distributed a draft of Z39.50 Version 2. He said that the main purpose of version 2 was to bring Z39.50 into compatibility with ISO 10162/10163 and that it does not address other extensions. Its main difference from version 1 is that it now uses ASN.1 descriptions. It should be ready ballot in 3-6 months. Ray promised to make an electronic version of the ASN.1 material available. Document became ZIG 90-0010. Ray asked that serious concerns about version 2 should be sent in by 10/13 and should be sent to the mailing list. He asked that comments not address proposed changes but rather the process by which version 2 was brought into compatibility with SR. A discussion was then held on how to progress version 2 to version 3. 5) ZIG Document and Secretariat Discussion A discussion was held on what to give ZIG numbers. A decision was taken that revised documents get new ZIG numbers rather than replace older numbers. The outdated document will indicate what ZIG number revises or supersedes it. ZIG 90-0011 was assigned to the revised PICS Pro Forma document. This revises ZIG 90-0002. DRA offered to act as the secretariat for the group and will mail out documents that are not available electronically. The decision was taken for everyone to post their US mail address and fax number to the mailing list so that DRA could collect them. Mark Hinnebusch will send out a reminder to do this. 6) Maintenance Agency Ray reviewed the role of the maintenance agency (to be administered by the LC Network Development and MARC Standards Office) and its status. The agency is established and NISO will issue a press release announcing it. Ray also distributed 2 forms, a defect report form and a change request form. The change request was assigned number ZIG 90-0012 and the defect report was assigned ZIG 90-0013. It was decided to modify the change report by dropping "US library community" from the benefits statement. It was also decided that new services will come under the category of changes. Version 2 will be the 1st change form. The maintenance agency plans to have an implementers guide prepared. 7) Stoplists Clifford Lynch recalled issues of stoplisted words and how to find out what was stoplisted in a database. This is a possibility for the explain service. There was a discussion of the possibility of finding out what terms were stoplisted in a query and whether this should be done with an error message or a user information field. This lead to a general discussion of other issues of intermediate results and finding out postings on terms. Ralph LeVan proposed a ZIGE document to discuss issues of intermediate results. Mark Hinnebusch proposed a 2nd ZIGE document to discuss how a client can use resource control to indicate to a server not to do a search if one component of the search causes too many postings. Clifford will work with Ralph to incorporate his ideas on stoplisting into Ralph's document. 8) Resource Control Standardization/Registration Clifford observed that formats of resource reports should be a registry item. Ray mentioned that Z39.50 version 2 already includes this. An ASN.1 description of the resource control APDU was reviewed. Clifford proposed a new APDU that could be sent any time a connection was idle to establish resource control limits and to get back the status of resources spent. Mark Hinnebusch proposed adding a threshold at which the target will start sending resource controls. Clifford will draft a document on these issues. Clifford also brought up the issue of values for elements in resource control and will draft a document on this. There is the possibility of using ISO standard definitions for currency. Mark Hinnebusch proposed a mechanism a client could use to find out the quantity of system resources used on the target. The conclusion was that this is not feasible in Z39.50 due to problems of units of measurement of resources. It was also decided to investigate whether this feature could be provided by using network management protocols. 9) Attribute and Error Message sets Error message sets is still an outstanding item since Ralph didn't produce the promised work. A discussion was held about adding attribute sets. The decision was taken to replace the attribute set in Z39.50 with the one from SR but with the provision that fields in the Z39.50 set that were not in SR would be included. Also a use attribute for name will be added. There was a discussion of what should be in BIB1 and how large it should be and at what point another set should be defined. There was also a discussion of how to define the semantics of attributes. A decision was taken that people with proposed attributes should send a description of them to the mailing list and Ray will consolidate and coordinate them. It was also decided that the no truncation attribute in the truncation value codes should be put back as meaning do not truncate for fields that default to truncation. 10) Status Reports FCLA - Waiting on Title II grant but doing design work. Plan is to use IBM OSI/CS code. OCLC - will also use OSI/CS but not dependent on it. Is also building upper protocol stack on top of TCP. Should have something by 4th quarter 1990. Will support type 1 as well as type 0 queries. OCLC is giving code to Duke and Case Western for user interface design experiments so code from those places can be expected next spring. OCLC also announced that BER record handling code given to UC will be put in the public domain. Ralph will announce availability on mailing list. PSI - Has prototype origin that uses Z39.50 version 1 and does type 0 query against the OCLC database. Wants to generalize and go to version 2. Code is publicly available from ssd-info@psi.com. Also has a white pages component available. CNRI - Has experimental gateway to NLM databases. Trying to develop standard user interface and vocabulary assistance function. Interested in Z39.50 as transport mechanism between user workstation and gateway. UC-DLA - Current schedule is to have something running by late 1st quarter 1991. CMU - Moving ahead with prototype. Intend to be running limited campus distribution by January 1991. Maybe type 0 query. Using Motif layered over TCP. Hope to be in conformance with version 2 by spring. Will be going ahead with images and SGML text. UC-Berkeley - Moving along more slowly than expected. Looking at ISODE and moving ahead with OCLC code. National Library of Canada - Joint effort with Utlas. Completed feasibility study and now reviewing it with Utlas. Next step will be implementation. VPI - working on Smalltalk simulation. Almost done but nothing before early next year. NOTIS - Rewriting public access module. Once that is complete will work on Z39.50. First use will be for NOTIS to NOTIS interchange. Also dealing with user interface issues. DRA - Got ISODE working under VMS. Trying to get ASN.1 to compile. Had been using code from Thinking Machines Corporation but now writing own code using version 2. Dartmouth - Interested in doing something in next 3-6 months. LC - Installed IBM OSI/CS and played with ASN.1 compiler. Working with IBM to fix lack of recursion in IBM product. 11) TESLA Program Mark Hinnebusch discussed the possibility of demonstrating prototype implementations and interoperability at a TESLA program at summer ALA. ALA will provide equipment and phone lines. Andy Cohill said VPI might be willing to participate. OCLC will have a commercial product by December and will be available for interoperability testing. This will include Z39.50 version 2 over TCP and maybe over an OSI stack using ISODE version 4 over X.25. Ralph will announce availability of OCLC for interoperability testing on mailing list. Clifford said DLA may be able to participate as well. Issue of whether to do this and who will be ready to participate will be decided by phone and on the list before the next meeting. 12) Explain Facility Clifford hasn't gotten any feedback on the initial document and said he still needed to flesh out the proposal but hasn't done it yet. He will do so by next meeting. He solicited any comments people might have. 13) After hours discussion A discussion was held on where to go drinking after the meeting broke up for the night. Bullfeathers was agreed on. 14) Glossary of Terms Discussion was held on the need for having a glossary of terms used with Z39.50 and on what format that should be. Ray will take responsibility for getting this done and solicited contributions on what should be defined and proposed definitions. These may become an appendix to the standard or be published as a technical report. Ray will be the editor for the submitted terms and suggestions should be posted to the mailing list within the next 30 days so they can be included in US ballot on SR. Wayne Davison also will make OSI definitions developed by SC21 available. 15) ASN.1 Mark Hinnebusch summarized the issues. A discussion of how ASN.1 and BER worked ensued. There was also a discussion of what was the best way of encoding MARC records in ASN.1. Wayne Davison volunteered to provide a example syntax of an ASN.1 external data type as a means of providing a framework for a discussion of ASN.1 and BER encoding. Mark Kibbey proposed that a group work on defining ASN.1 descriptions for non-MARC data structures. He is willing to coordinate such a group. Jim Michael brought up the issues of other types of emerging standards such as circulation and patron records. The consensus of the group was that at least initially data will be passed as an ASN.1 external data type of MARC provided ASN.1 compilers can handle data types of external that are not defined in ASN.1. Otherwise we will define a second object id which is external and is a MARC record ASN.1 encoded as an octet string. 16) Priority of Services Mark Hinnebusch brought up the idea of defining priority of services for implementations that are both client and server. He will put a document outlining his thoughts on this on the mailing list. 17) Other related standards Clifford discussed standard work going on with various new structured full text query languages and other things. He also discussed a new server specification for CD-ROM that is being proposed. He will send out copies of these documents if requested. 18) Browse and Index Searching Joe Zeeman asked for a status report on what is happening with Browse in standards committees and described what he meant by index browsing. He distributed a document on this which became ZIG 90-0015. Sally McCallum reported on the status of a subcommittee looking at adding this service. The intention is this would be an addendum to the existing protocol. Ralph described the existing implementation of Browse at OCLC. Ralph and Mark Kibbey offered to draft up a document as to what this service would look like. The proposal was for people to submit descriptions of a potential browse to the list. Wayne Davison will serve as the keeper of an issues list. Jim Michael will provide some examples. Clifford will also send Wayne some documentation of various types of browses. 19) Multiple profiles for different retrieval styles Joe Zeeman described the National Library of Canada's planned use of Z39.50 for known item retrieval and his need for an ability to describe what facilities the system provides so that clients would know what attribute and element set elements are supported. His document on this is included in ZIG 90-0015. There was a discussion of whether these kinds of things could be handled in the profile or a conformance statement or even by a different application context name that could be negotiated. Joe will work up some more specifics on what he needs and outline his concerns. 20) Aborting Searches Clifford discussed the need for a client to be able to cancel an individual search without aborting the entire session. Discussion ensued. The issue was raised as to whether this could be done as part of resource control since that would reduce incompatibility with SR since resource control is not in SR. There was also a discussion of adding cancel to other services such as browse and present. Clifford will draft up a document on a cancel feature. 21) Recursion and ASN.1 Ray Denenberg described a problem with the IBM OSI/CS ASN.1 compiler: its inability to handle recursion. He discussed his experience in working with IBM to get this problem fixed. He asked for information from the group as to who needed recursion in the IBM product as a way of providing input to IBM that this was an important and required feature. Ralph LeVan and Mark Hinnebusch will work up documents for Ray on the need for recursion in the IBM product. Ray will act as coordinator for all input received. 22) Proximity Operators Sara Randall brought up the issue of the lack of proximity operators and queried the group as to whether they thought this was a problem. Sally McCallum mentioned that this had been discussed at the ISO level and there had been a lack of interest. Consensus of the group was this issue needed to be dealt with and discussion was then held on how this could be handled. Ralph LeVan will put together a document describing OCLC's current implementation as a way of providing a framework for the discussion. 23) Future issues to be discussed and resolved The group put together a list of issues that needed to be discussed and resolved in the future. Among these were: a) Informatory messages on any command - how to define structure b) User information field c) Full text retrieval d) Handling non-bibliographic data e) Sorting and ranking of result sets f) Element set subsets and views - defining ad hoc views of records g) Batching Search requests and asynchronous applications h) Syntax for holdings and circulation information - Mark Hinnebusch will develop a description and put it on the list 23) Interop Birds of a Feather session Mark Needleman mentioned that there will be a BOF on Z39.50 at the Interop conference. Zig Document List ZIG 90-001 Application Profile for Z39.50, March 1990 (1st draft) ZIG 90-002 Profile PICS Proforma for Z39.50, March 1990 (1st draft) [Revised by ZIG 90-0011] ZIG 90-003 Z39.50 Maintenance Agency Terms of Reference and Procedures. November 1989 ZIG 90-004 Documentation - Search and Retrieval Service Definition. ISO/DIS 10162 ZIG 90-005 Documentation - Search and Retrieval Protocol Specification. ISO/DIS 10163 ZIG 90-006 5. Addressing requirements (extract from GOSIP documentation) ZIG 90-007 Image and full-text transfer under Z39.50 -- issues for discussion. (Clifford Lynch June 1990) ZIG 90-008 Contributions on: Z39.50 PICS Proforma, Agenda items, Stop words, Z39.50 Profile, Attributes (D. MacKinnon and J. Zeeman, Software Kinetics Ltd. June 1990) [5 items] ZIG 90-009 Extensions to ISO DP 10162/10163 to support and Explain Service. (Clifford Lynch, December 9, 1989) ZIG 90-0010 ANSI Z39.50 Version 2 Draft. September 1990 ZIG 90-0011 Profile PICS Proforma for Z39.50 June 1990 (2nd draft) [Revises ZIG 90-002] ZIG 90-0012 NISO Z39.50 Maintenance Agency Change Proposal submission form. September 1990 ZIG 90-0013 NISO Z39.50 Maintenance Agency Defect Report submission form. September 1990 ZIG 90-0014 Z35.50 Implementers Group meeting notes, June 1990 ZIG 90-0015 Software Kinetics Ltd. agenda item on Browse and Known item retrieval. September 1990 ZIG 90-0015 Software Kinetics Ltd. agenda item on known item retrieval. September 1990 Next Meeting The next meeting will be held at the University of California on Thursday 1/31/90 and Friday 2/1/90. The meeting will start at 9 AM on Thursday and run until noon on Friday. The exact meeting location is still to be determined. Attendees The following people attended the meeting: Kaushi Belani Library of Congress Jean-Eudes Beriault National Library of Canada Eric Bivona Dartmouth College Andrew Cohill VPI Computing Center Wayne Davison RLG Ray Denenberg Library of Congress Larry Dixson Library of Congress Sean Donelan Data Research Associates David Ely Corporation for National Research Initiatives Mark Hinnebusch FCLA Mark Kibbey Carnegie Mellon University John Kunze University of California-Berkeley Ralph LeVan OCLC George Loughlin IBM Clifford Lynch University of California-DLA Sally McCallum Library of Congress James Michael Data Research Associates Mark Needleman University of California-DLA Ralph Orlik Library of Congress Cecilia Preston Strategic and Competitive Research Sara Randall NOTIS Edwin Sequeira National Library of Medicine Lennie Stovel RLG Wengyik Yeong PSI Inc. Joe Zeeman Software Kinetics Ltd. Minutes were recorded by Mark Needleman and edited by Lennie Stovel.