SC AV: Meeting 2 Summary, January 10-11, 2001

	National Information Standards Organization (NISO) Standards Committee AV
	Developing a U.S. National Z39.50 Profile for Library Applications

MEETING 3 SUMMARY
Recorded and Prepared by
Mark Needleman
Margaret Quinn
&
William E. Moen, Chair

This document provides a summary of the third meeting of the National Information Standards Committee (NISO) AV that is charged with the development of a U.S. National Z39.50 Profile for Library Applications. Members of SC AV met for 2 days to continue its work to prepare a draft standard that NISO will submit for voting by its voting members. This summary does not detail all discussions at the meeting but rather is meant to communicate the important decisions and agreements based on committee discussion. This summary is ordered based on the agenda of the meeting.

Moen introduced Margaret Quinn, the NISO Research Assistant and her work on the NISO glossary project. Introductions are made.

Moen requested agenda revision. He asked to move the discussion of holdings to be the primary work item for Day 1. Moen also mentioned the UNT SLIS funding of lunch for the Tuesday meeting. Dean Philip Turner of SLIS will present news about the Texas Center for Digital Knowledge.

Carrol Lunau pointed out an error in the previous meeting minutes in the section summarizing her comments on the Bath Profile Status.

Lunau presented changes to the Bath Profile, including web location of documents and status of upcoming meetings. Minor editorial changes were incorporated, but this has not produced a new version at this time. Randall asked about approval for changes to Bath document and the amendments that have been put forward. Lunau responded that they should be considered approved. Moen mentioned the IRP approval process which the Bath Profile has been through. Williams talked about the necessity of date-time stamping of the Bath Profile to identify version. Lunau asked that the Bath online discussion list be used for opinions. Any references in the NISO Profile to the Bath Profile must be clear as to which version and date of Bath is being referenced. This will be done. The date for the next Bath Profile Group meeting has not been set.

There was general discussion of the requirements for Level 0 for Functional Area B. Moen suggested that Level 0 should be an entry level with no search. Goldner disagreed, but Needleman responded that listings are retrieved without search specifications but in a specific format.

Questions were raised about standardizing with the Bath Profile. Level 0 compliance between the two was discussed. The committee decided to move beyond the Bath Profile for Level 0 for National Profile Level 0. For this -

1) XML must be supported;
2) static element set names for returning holdings data must be supported; and
3) any returned holdings data must be consistent with the ZIG Holdings Schema.

Required reporting levels at Level 0 will be discussed at a later date. The requirement for Z39.50 Protocol Version 3 is not necessary for the lowest level.

The relationships between record syntax, record schemas and the Z39.50 CompSpec function were discussed. For simple holdings interchange, the use of primitive element set names (ESN) does not require the use of CompSpec or Espec.

Moen will confirm that the ESN names used in the National Profile for holdings match those used in the ONE-2 Profile. This relates to a proposal accepted at the last ZIG meeting put forward by the ONE-2 folks.

The Committee began by looking at the various reporting levels as defined in the Holdings Schema and put off discussion on Levels until the time being.

Wilson pointed out apparent inconsistencies in what is listed as Mandatory (M) in the table. Pattie recommended tightening M specifications. Part of the issue here us what Mandatory means. The Committee suggested a clarification along the lines of: "it is recognized that not all servers will be able to provide all the information marked as Mandatory but must return that information if available." This is part of a larger principle that needs to be written into the Profile. Mandatory refers to those elements that must always be provided, if available, to claim conformance to a particular Element Set Name.

Goldner questioned the structure of bibiteminfo since actualbibteim does not seem to be required always. Moen described the notion of tag set and tag path that is used to represent the abstract record structure of the holdings. The Committee decided that actualbibitem must be required as does holding site location. Bibitem will be a MARC record formatted in ISO2709 encapsulated in the XML. HoldingSite Location is defined to be data from the 852$a.

A question was raised about how to require a service policy in B1 when no service is asked for. The Committee agreed that in this case, a value of 0 would be sent back to mean "unknown".

A server can send back elements defined as optional for a reporting level, and the client needs to be prepared to receive them but it can ignore the optional elements. This will be the case for all reporting levels.

Discussion concerning inheritance at various levels in Functional Area B followed. Committee agreed to have each level specify inheritance and to not make inheritance default. The term "level" was not considered the most appropriate for Functional Area B. Alternatives such as "task" and "reporting level" were suggested. The primary objective was to identify the ESN and their composition required and then address how to specify a level or task.

There was an initial discussion of what is required in other, more detailed reporting levels, and included:

Status and call number were later removed from list as elements more appropriate in a lower level.

The group discussed some definitions that appear in the holdings schema or are related to holdings information including:

Moen requested assistance with summary holdings into definitions. Discussion included the following elements:

The Committee decided that for the summary level, chronology and enumeration will be unstructured and repeatable, guided by formatting described in Z39.71. UnitName/TypeofUnitDesignator is then removed. Discussion about the relationship between number of copies and enumeration followed.

For B2, actual BibItem should be included, even though it is not listed as mandatory in B2. Also, it may be necessary to include targetItemID (agreed to in Day 2 discussion).

There was also discussion about what information is needed for machine processing (e.g., ILL and NCIP) and what information is needed and should be displayed in human readable form for users. There was some sentiment that status information should be made available/ displayable for human consumption.

For circulation information, Goldner recommended bringing codes in the holdings schema and similar codes in NCIP into alignment. Moen requested that Wilson, Needleman and Goldner further investigate. Moen recommended a mapping of codes which are used in NCIP for status. Lunau asked about the compatibility of ILL codes compared to NCIP. Moen asked Quinn to look for ILL code mapping.

Riding recommended adding full name of branch or local location. Wilson countered using a foreign language example. Riding further explained that the detailed location would not be specific enough. There was no consensus on the need for a institution specific information. Moen suggested a human readable code which would help specify the location. Johnson recommended having both machine and human readable code, Randall said that was not possible according to the schema. Wilson recommended a client side translator but this appeared not feasible to the others. Goldner suggested a possibility for sending both using the element subLocation. Randall disputed the repeatability of the element. Debate on the feasibility of the subLocation element for the purpose followed. Randall requested inquiry of the schema developers for clarity on the subject. The Committee decided to recommend the use of full names and that the subLocation field should be used for this purpose for now.

There seems to be a need to ask the Holdings Schema developers why they do not have separate element for full name of location and possibly suggest one be added or the use of existing fields be better explained.

It was agreed that status is explicit for each piece, or at least mandatory when available.

Pattie asked if specific format of resource needs to be explicitly stated. It was agreed that physical form designator needs to be explicit. The capability for this occurs in B2.

In trying to decide what data should be in an ESN, the Committee agreed to the following general principle: That data are necessary in the context of Z39.50 retrieval and use, and not that it might or might not be available via another protocol (e.g., ILL NCIP). However there was a recognition that Z39.50 should retrieve sufficient information that could be passed on to an ILL or NCIP application (e.g., targeted).

Moen identified the agenda for the day's meeting, and he introduced two guests, Haley Holmes and Teresa Lepchenske of the Z39.50 Interoperability Testbed project. The document of the preceding day's definitions and element set names was distributed.

Moen described the documents he distributed. These try to reflect discussion from Day 1. Instead of using Reporting Level (B1, B2, etc.) or the ESNs adopted by the ONE-2 profile folks, these are intended to represent possible US National Profile ESNs and therefore have the following meaning:

He explained the numeration for the tagsets and nesting elements. Williams asked about holdings levels in the Bath Profile. Lunau responded with the current information. Meyer asked about the inclusion of actualbibitem in this level. Others said that it has been added to all levels.

Discussion of the summary level holdings followed. Meyer asked if this was copy view rather than bib view. Changes were made to Moen's ESN based on the discussion. There was talk about the term "copy" as it pertains to the element numberofCopies. Moen asked for a group to look specifically at the relationship between bibview and copyview and to present recommendations. For USESN2, element bibview was changed to copy view and elements typeofUnitdesignator and unitname were excised. For USESN3, element temporaryLocation will be added.

Meyer asked for the addition of the templocation field in USESN3. Goldner asked for enumerationChronology to be added in lieu of summaryEnumeration. Group agreed with the changes. Complete enumeration and chronology will be in USESN3. Group said that a standard ID number, text and info were needed. USESN3 will have more detailed Chron and Num elements. Bailey asked if Summary Level is useful to human users and would need a call number. Group decided that for now, it is not needed. Williams asked for clarification on the term "piece", does it equal item? Goldner said it might differ from system to system. Goldner asked about starting a set name committee. Moen asked for volunteers for the committee. Riding, Goldner, Meyer and Pattie agreed to participate. Lunau asked the committee to look at the implications on a national union catalog on the summary level.

Prefacing her statement as a "curveball", Pattie recommended that Levels 1 and 2 should be combined. Moen asked that the group wait until level conformance can be looked at. Ferrin and Needleman said that level one is necessary in some instances. Williams asked if NISO can accommodate the long-terms needs of this standard. Moen responded that he believed this can be done over time rather than all at once.

The Committee agreed that searching of holdings will remain a future work item, but work will begin as soon as the first version of the profile is submitted to NISO. Moen will contact Pat Harris as to the best procedure to add that function.

Moen then asked the Committee to begin discussion on the other areas of the Draft Specifications. The objective was to get agreements as possible on both General Requirements and also the specifics of Functional Area A for Search and Retrieval.

Wilson questioned the general requirement for support of result set-names (with 2 results sets required). He suggested it might be a performance resource barrier for some small systems. He also expressed concerns about the profile having too many loopholes and not enough "standard" requirements. Dixson pointed out that this is directly out of the Bath Profile which worried Wilson even more.

Debate about the relevance of the Bath Profile on the current profile occurred. Williams said that most vendors are international and they would require Bath compliance, so the US National Profile should be as well. Debate about the Precision Author Search and compliance with Bath followed. Goldner expressed concerns about the conformance between Bath and the National Profile. Pattie said that the Committee should state the purpose in the maintenance section of the Profile.

Wilson asked to lose the reference to non-conformant searches. Meyer asked to add a "no substitutions" phrase to the second to last paragraph under the Conformance section. This was agreed to.

Tables that are duplicate listing of bib-1 attributes will be removed from the general requirements.

Dixson noted that Xs need to be out in the Z-client and Z-server columns next to MARC21. A mistake was pointed out in a paragraph taken directly from the Bath Profile which states that records must be delivered in SUTRS. Debate on the "any" search followed which concluded that some word crafting will be done to indicate that the server should be able to go across author, title and subject and might include more. However, there is a difference between what the vendor implementation could do and what a particular customer site might choose.

The Committee decided to kill BL0.1 - Precision Match for Established Name Heading. It will still be in Bath and will be required for Bath compliance. Debate continued about the relationship between the US National Profile and the Bath Profile since the US Profile may not be seen as a compatible superset of Bath anymore. Moen suggested that during this meeting, we state exactly what we want in the NISO Profile and leave for the time being the requirement to maintain absolute compatibility with Bath. Moen will compile a list of preliminary differences between Bath and the NISO Profile. There is a desire to use Bath as core but not necessarily maintain complete compatibility.

The Committee agreed that at Level 0 Z-clients and Z-servers only need to support MARC21. No non-MARC record syntaxes are required. Language about character sets will detail only the sets defined in the syntax. For MARC21, only the MARC8 character set is required, not MARC with Unicode.

The sentence about character sets, the language on character sets for searching will be clarified to indicate that support for Latin-1 is only for encoding the term - this does not mean the client and server cannot support more.

The Committee decided to strike BL1.1 - Precision Match for Author. The Character Set and Language Negotiation line in Search and Retrieval Requirements is to be removed with a statement that it may be added in the future where there is more agreement on how this is to be done in Z39.50.

Some language will be put into the profile indicating that the Profile would like to serve both Canada, and possibly Mexico, and that language negotiation will be important requirements to address.

Clients will be required to support four pattern searches for all controlled vocabularies. This means adding a First Characters in Field Search to the pattern searches. The Committee decides that user systems must support at least one controlled vocabulary, assuming that their data has used a controlled vocabulary, and controlled vocabulary searches will move to Level 2.

Moen pointed out search US1.5. Discussion about 3-letter codes for material-type use attributes followed. Formats and corresponding codes will go into the body of the standard while the MARC21 record data sources will go into the appendix. Material type searches will go into Level 2. For the table of the MARC21 tags, there is a missing note about implied Boolean OR between columns, etc. This will be put in.

There was discussion of the need to add ability to search 245 subfield h (gmd) in the Level 2 type of material search. It was discussed that this was mostly useful for school libraries and did not make much sense for academic libraries. Discussion followed of whether table listing what fields in the record should be searched might better belong in a non-normative appendix since they were talking about indexing decisions. The Committee agreed to keep codes in the main portion of the profile and the table of what fields to include should move to a non-normative index.

The Committee decides to move all keyword scans to Level 2. Goldner raised questions about preferredPositionResponse. Goldner and Dietz volunteered to investigate.

The Committee confirmed the combination of attributes and values for expressing the various SCANs.

Client Side	MARC21, SUTRS, and XML with MARC DTD
Server Side	MARC21 and (XML with MARC DTD or SUTRS)

Based on the review of various MARC DTDs by Needleman and Dixson, the Committee decided to use the LC MARC DTD. There are issues, however, of how to specify a DTD in a Z39.50 present request, since only Record Syntax can be identified.

There was discussion of Randall's e-mail regarding proximity. Proximity would be considered for Level 2. Moen will work on specifications of what might be included for proximity searches.

The Committee discussed whether to push forward on work on Level 2 for this initial version of the Profile. Meyer wished to continue with discussion on Level 2. Goldner expressed concern about presenting an overly large document for consideration by the voting body. Moen stated that few searches have no use attributes, certain searches are problematic (date range, OCLC), and proximity. Goldner stated that the time factor may prohibit completion of Level 2 and anything added may have to become Level 3. Moen suggested continuation to complete Level 2 for the next draft. Further comments on Level 2 should be posted to the mail list. On the mail list, Dixson will post searches currently in Level 2 that he believes are critical and could possibly be moved to Level 1 if no Level 2 is specified.

There was discussion concerning the most appropriate MARC21 fields for inclusion in the profile. Pattie noted that usage and target audience must be considered for indexing choices. Haley Holmes of the Interop Testbed project provides explanation about the handout she created of recommendations for indexing MARC21.

Holmes will continue to work up guidelines and recommendations to the interoperability testbed. Peterson and Pattie will review the indexing guidelines. Results will be posted to the Zprofile listserv and elsewhere.

It is decided that guidelines will be mentioned in the profile but not included, although there is some interest in seeing them as a non-normative appendix.

Discussion on Moving Forward on this Version of the Profile and a Subsequent Version

Moen asked for help from an editorial review committee to review the narrative sections as he gets them written. Williams, Wilson, Dixson, Needleman and Dietz volunteered to help.

Moen will draft a letter to send to Pat Harris at NISO to tell her of the current status, plans for completion and the potential for additional work items. He suggested a September 15 date for delivery of the ballot-ready profile to NISO.

Moen will also suggest to Harris a possible starting date for the next section of work. If the Profile looks like it will be approved by NISO in fall 2001, then work could commence on other requirements in late fall 2001 or early winter 2002.

Moen will be working on Functional Area B specifications and plans to have something ready for review by the ESN committee by early May. He will also be revising Functional Area A, Levels 0 & 1 for consideration as soon as possible. A complete draft for the committee will be done by early June.

Possible meetings could coincide with IFLA in Boston or ALA in San Francisco. Moen will post potential dates for meetings.

1. Compile a list of preliminary differences between Bath and NISO profiles based on discussions at this meeting.

2. (Possibly with Randall) Work up some specifications for a number of proximity searches. How would these be incorporated into current search specifications? Are these at a higher level or what?

3. Develop a Functional Area A, Level 2 to see if it is something that needs to be out in the first version of the Profile.

Wilson, Goldner, Needleman & Quinn: Examine codes for circulation status (and others) that appear in NCIP, ILL and the Z39.50 Holdings Schema and try to bring them into alignment. Wilson, Goldner and Needleman will look at codes in NCIP and Quinn will examine ILL codes (maybe contact Lunau for assistance on this).

Riding, Goldner, Meyer & Pattie: Comprises the Element Set Name group to work with Moen on completing the specifications for three or four ESNs. It will be important to see if the ESNs work well with physical union catalogs as well as virtual catalogs.

Dixson: Post searches currently in Level 2 that he believes are critical and could possibly be moved to Level 1if no Level 2 is specified.

Peterson, Pattie & Holmes: Review indexing guidelines and recommendations coming out of the Texas work (TZIG and Z39.50 Interoperability Project).

Williams, Wilson, Dixson, Needleman & Dietz: Will serve on an editorial committee to review draft text of the profile.

Affiliation	Name	Present
UNT/SLIS	William E. Moen, Chair	X
Data Research Associates, Inc.	Mark Needleman, NISO SDC Liaison to SC AV	X

AMIGOS Library Services	Chris Peterson	X
Blue Angel Technologies	Margaret St. Pierre
California State Library	Ira Bray
CIC, Penn State University	Eric Ferrin	X
Colorado State Library	Brenda Bailey	X
Endeavor Information Systems	Sara Randall	X
Epixtech	Ed Riding	X
Follett	Michael Johnson	X
Fretwell-Downing USA	Matthew Goldner	X
GALILEO, Georgia	Phil Williams	X
Innovative Interfaces	Laurie Davidson	X
Kentucky Commonwealth Virtual Library	Miko Pattie	X
Library of Congress	Larry Dixson	X
MINITEX	Christina Perkins Meyer	X
Bath Profile Maintenance Agency	Carrol Lunau	X
OCLC	Dana Dietz	X
PALCI	Dan Iddings	X
SIRSI	Slavko Manojlovich	X
The Library Corporation (TLC)	Mark Wilson	X