A Working Draft

An International Z39.50 Profile
for Library Applications

A Working Draft

Prepared for discussion at

The Stockholm Z39.50 Implementors Group Meeting
&
The Bath Interoperability Profile Meeting

by

Carrol Lunau, National Library of Canada
<carrol.lunau@nlc-bnc.ca>

Paul Miller, Interoperability Focus, U. K. Office for Library and Information Networking (UKOLN)
<p.miller@ukoln.ac.uk>

William E. Moen, School of Library and Information Sciences, University of North Texas
<wemoen@jove.acs.unt.edu>

August 4, 1999

Table of Contents

1. Introduction to the Profile
2. Purpose and Scope
3. Functional Requirements
    3.1. Requirements for this Draft
        3.1.1. Requirements for Search
        3.1.2. Requirements for Retrieval
        3.1.3. Requirements for Search and Retrieval of Holdings Information
        3.1.4. Requirements for Browsing Indexes
        3.1.5. Requirements for Cross Domain Searching
4. Z39.50 Specifications
    4.1. Protocol Version
    4.2. Z39.50 Objects
    4.3. Z39.50 Services
        4.3.1. Init
        4.3.2. Search
            4.3.2.1. Attribute Set
        4.3.3. Retrieval: Record Syntaxes
    4.4. Diagnostic Messages
    4.5. Conformance
       4.5.1. Functional Area A: Library Catalogue Search and Retrieval
           4.5.1.1. Functional Area A: Category 1 Searching
      4.5.2. Functional Area B: Search and Retrieval of Bibliographic Holdings Information
      4.5.3. Functional Area C: Cross Domain Searching
5. References

Appendix A: Definition of Title and Subject Searches
Appendix B: Z39.50 Commentary--Z39.50 Keyword Searching of Bibliographic Systems
Appendix C: Draft Framework for Holdings

An International Z39.50 Profile
for Library Applications

1. Introduction to the Profile

This document identifies a subset of specifications from the Z39.50 Information Retrieval Protocol (ANSI/NISO Z39.50/ISO 23950) for use in Z39.50 client and server software to support search and retrieval among library catalogues and union catalogues worldwide. The Profile will evolve as the environment and the standard change and is intended to facilitate resource sharing among libraries globally.

This profile builds upon the experience of other profiling efforts addressing interoperability for library catalogue searching and the development of virtual union catalogues. In particular, the following efforts informed the profile:

• ATS-1 Profile
• CENL Profile
• DanZIG Profile
• MODELS Profile
• ONE Profile
• The Z Texas Profile
• Virtual Union Catalogue Profile
• Discussions surrounding the development of an international profile for searching library catalogues, and specifically, a document by Carrol Lunau and Joe Zeeman, Z39.50 Basic Searching of Bibliographic Systems: A Discussion Document.

The structure of the profile is modular and allows the future specification for separate but compatible requirements involving a range of applications useful to librarians and library patrons.

This draft defines several Functional Areas:

• Functional Area A for search and retrieval among library catalogues
• Functional Area B for search and retrieval of bibliographic holdings information
• Functional Area C for cross domain searching

Other functional areas may be defined in future drafts of this profile. For each functional area, Z39.50 specifications and conformance requirements will be provided.

2. Purpose and Scope

The purpose of this profile is to identify those features of the Z39.50 standard that are required to allow the effective use of Z39.50 software to search and retrieve bibliographic data from library catalogues, to support the transfer of holdings information, to support updating union catalogues, and to support cross-domain searches such as between libraries, museums and archives. Implementation of this profile by systems developers will improve system and semantic interoperability among diverse systems and improve search results within these functional areas.

Searching bibliographic databases is not an end unto itself but rather one step within a larger activity such as interlibrary loan, cataloguing, reference, acquisitions or catalogue updating. For this reason, the profile will be developed incrementally to incorporate richer functionality.

The primary goal of the profile is to increase the semantic and technical interoperability between disparate systems so that end-users can use Z39.50 clients to search catalogues and be confident that they have retrieved valid result sets. A secondary goal is to release the first version of the profile as quickly as possible rather than waiting for the complete definition of all functional areas before the initial release.

Initially, the profile will encompass:

1. A definition of attributes and attribute combinations required for basic search and retrieval of library catalogues.
2. A definition of the use of the OPAC/Holdings format.

A key component of this profile is to characterize the types of searching required by library users. To this end, the profile defines specific searches and how the semantics of those searches will be expressed in the vocabulary of Z39.50.

It is anticipated that subsequent versions of the profile will incorporate, but not necessarily be limited to, the addition of the following functions:

1. A definition of attributes and attribute combinations required for more advanced searching and retrieval of library catalogues, such as might be used by library technical service staff or reference staff.
2. A definition of attributes, attribute combinations and other requirements to support cross-domain search and retrieval.
3. A definition of requirements and specifications to support the updating of union catalogues to report new bibliographic and holdings information.

3. Functional Requirements

This section identifies the functional requirements informing the Z39.50 specifications in this document.

3.1. Requirements for this Draft

The basic requirements for this draft focus on search and retrieval between library catalogues, and the search and retrieval of bibliographic holdings information. The following sections detail these requirements.

3.1.1. Requirements for Search

Outlined below are two categories of searching requirements. Category 1 can be considered a set of core searches.  For this draft, only Category 1 searches are defined and specified.   Category 1 searches provide basic functionality that all local systems should support.

Category 1 searches include common author, title, subject searches that might be used by patrons and librarians. The following summarizes Category 1 searching requirements:

• Author searches for expressing queries that include an established name heading as well as queries that include names not under any authority control
• Title searches include phrase and keyword searching of titles. Phrase searching includes both exact match on title and first words of title.  A search on title keywords will retrieve all titles containing the specified keywords.  A search on first words of title will retrieve titles that begin with the specified string.  A search on title exact match will retrieve titles that contain the complete string as specified. (See Appendix A for a discussion paper, Definition of Title and Subject Searches)
• Subject searches include phrase and keyword searching of subject headings. A subject heading is defined as the complete subject
  heading, including all subfields. Phrase searching includes both exact match on subject heading and first words of subject heading. A
  search on subject keywords will retrieve all subject headings containing the specified keywords. A search on first words of subject
  heading will retrieve subject headings that begin with the specified string. A search on subject exact match will retrieve subject
  headings that contain the complete string as specified.
• Keyword searches against author, title, subject, and other common access points (see Appendix B for Commentary on Keyword Searching of Bibliographic Systems)
• Support for Boolean searching using AND, OR, NOT
• Support for right truncation of search terms.

Many OPACs currently allow users to browse indexes.  This is a useful feature, often used in conjunction with known-item searching. Browse-based searching can assist users in improving their selection of terms for the query.  Such browse-based searching can be seen as an alternative searching strategy for Category 1 (see 3.1.4, Requirements for Browsing Indexes).  

3.1.2. Requirements for Retrieval

This draft addresses retrieval of bibliographic and holdings records from library catalogues. Retrieval of bibliographic records given the existence of national MARC formats presents interoperability challenges.  Z-servers must return brief and/or full forms of bibliographic records in an internationally accepted MARC format.  This profile suggests that the default record syntax for retrieval is MARC 21 (the result of harmonization of the USMARC and CANMARC formats). It is assumed that there will be national or regional companion profiles to the international profile that may specify additional requirements such as a preferred MARC format when retrieving records from MARC databases in a specific country or region.   

NOTE:  The issue of the MARC record syntax profiled in this document is an issue that will be discussed at the Bath Interoperability Meeting.

Librarians and other users, however, are interested in accessing more than bibliographic data. Such access will require other structures for retrieval likely to be accommodated through the Z39.50 record syntax, Generic Record Syntax (GRS-1).  Retrieval of bibliographic holdings information requires the use of GRS-1, as specified in this document (see 3.1.3).

3.1.3. Requirements for Search and Retrieval Holdings Information

Identifying which collections contain certain information resources is a key factor in creating an infrastructure for resource sharing among libraries.  Librarians require holdings information for resource sharing, and patrons need similar holdings information for knowing where to go to use or borrow a resource.  The holdings information should provide sufficient identification and description of an item to give the user adequate information to make a decisions about requesting or retrieving the item. 

NOTE:  This requirement and the profiling of search and retrieval of holdings information will be discussed at the Bath Interoperability Meeting.

3.1.4.  Requirements for Browsing Indexes

Librarians and patrons often use a browse function on a local system to identify appropriate search terms to put in a query.  Searching term lists (or indexes) can enhance known-item searching.  Browsing indexes is a requirement for Z39.50 implementations.

3.1.5.  Requirements for Cross Domain Searching

Librarians and patrons desire integrated access to distributed resources in addition to search and retrieval of bibliographic information.  The profile specifies how Z39.50 can be used for effective cross domain searching of a range of resources including library catalogues, government information, museum systems, and archives.

NOTE:  The requirements and the profiling of cross domain searching will be discussed at the Bath Interoperability Meeting.

4. Z39.50 Specifications

This section identifies the Z39.50 specifications to support the functional requirements identified in Section 3.

4.1. Protocol Version

The profile addresses requirements related to Version 2 and Version 3 of Z39.50-1995.  All implementations are encouraged to use Version 3, but in the short-term, conformance to certain specifications prescribed by this profile is available to Version 2 implementations.  Section 4.5 details specific requirements for conformance for each of the Versions.

4.2. Z39.50 Objects

The profile addresses the following Z39.50 objects by reference to registered Object Identifiers (OIDs). For information on Z39.50 OIDs, see the Z39.50 Maintenance Agency's Definitions and Z39.50 Registry of Object Identifiers (see also Z39.50 Maintenance Agency).

Object			OID
Bib-1 attribute set			1.2.840.10003.3.1
Bib-1 diagnostic set			1.2.840.10003.4.1
USMARC record syntax			1.2.840.10003.5.10
GRS-1 record syntax			1.2.840.10003.5.105
OPACS/Holdings Schema			[to be assigned]

4.3. Z39.50 Services

The profile specifies the use of the following  Z39.50 services:

• Init
• Search
• Present
• Scan

See 4.5 Conformance.  No additional services are required for conformance to this profile. Z-clients and Z-servers optionally may use other Z39.50 services. 

Standard Z39.50 Init Service negotiation procedures control the use of all services.

4.3.1. Init

Z-clients supporting this profile may use the IDAuthentication parameters to transmit authentication information (e.g., userid and password). Z-servers supporting this profile may or may not require authentication. The profile specifies no other security requirements.

4.3.2. Search

The profile requires Z-clients and Z-servers to support Z39.50 Type 1 queries (i.e., general purpose Boolean query structures).

4.3.2.1. Attribute Set

The profile requires Z-clients and Z-servers to support the Bib-1 Attribute Set (i.e., process requests and responses that contain the OID for Bib-1). Semantics for Bib-1 attributes can be found in Attribute Set Bib-1 (Z39.50-1995): Semantics (September 1995).

Conformant Z-servers will support selected Use Attribute values according to Section 4.5. Conformance. "Support" in this context means:

• Z-clients must be able to transmit all of the specified attribute combinations in a Type 1 Query to request the specified services with relevant parameters and to accept the corresponding relevant responses from Z-servers including specified record syntaxes
• Z-servers must be able to receive all of the specified request types including attribute combinations and to return
  the corresponding relevant responses to the Z-client including specified record syntaxes. All of the specified attribute
  combinations must be supported by a Z-server, if they are relevant in the context of the corresponding databases.
• Recognize the Bib-1 OID
• Recognize the Bib-1 Use Attribute value(s) listed for a conformance level,
• Process searches using Use Attribute value(s) listed for a conformance level to produce a valid result set (which could contain 0 hits).

This implies that all implementations supporting this Profile must have search capabilities for Use Attribute values listed in each Functional Area for which conformance is claimed (see Section 4.5: Conformance). Z-servers may support additional attributes Bib-1 or other Z39.50 Attribute Sets.

NOTE:  The current assumption is that Bib-1 will be used in the profile.  But given the emergence of ZIG agreement on cross domain searching using the new attribute architecture and progress on the Bib-2 attribute set, the topic of attribute sets and the new attribute architecture is on the agenda for the Bath Interoperability Meeting.

4.3.3. Retrieval: Record Syntaxes

For interoperability, Z-clients and Z-servers must support a common record syntax.  The profile suggests that the default record syntax should be the MARC 21 Record Syntax to enable the basic interchange of bibliographic records.  Regional or national companion profiles will likely specify other Z39.50 registered MARC record syntaxes.

For retrieval of bibliographic holdings information, Z-clients and Z-servers must also support the Generic Record Syntax (GRS-1).

Z-clients and Z-servers may support other Z39.50 record syntaxes.  See Section 4.5 Conformance for specific requirements regarding record syntaxes.

4.4. Diagnostic Messages

The profile requires support for Diagnostic Set Bib-1 and Z-servers will return appropriate diagnostics.

4.5. Conformance

This profile will evolve to incorporate a range of functionality for distributed search and retrieval. To provide a modular approach to profile development, the notion of functional area is used to group similar requirements and Z39.50 specifications. Z-clients and Z-servers may claim conformance to one or more functional areas and be in compliance with this profile. Requirements and specifications for each functional area are compatible. While functional areas will differ as to their focal concern, it is likely that conformance to certain functional areas will specify conformance to other functional areas (i.e., to support a particular functional area may require conformance to one or more other functional areas).

This section identifies Z39.50 specifications Z-clients and Z-servers must support to claim conformance for each of the functional areas.  This draft specifies conformance for:

• Functional Area A for search and retrieval among library catalogues
  • Category 1 Searching
• Functional Area B for search and retrieval of bibliographic holdings information.
• Functional Area C for cross domain searching

Because of the goals, objectives, and detailed specification of this profile, Z-clients and Z-servers are precluded from "default" behavior (e.g., Z-clients should form queries using all attribute types and values identified, and Z-servers should not attempt to "reinterpret" a query because of limits on its search functionality).

4.5.1. Functional Area A: Library catalogue Search and Retrieval

Functional Area A addresses the requirements of basic search and retrieval among and between bibliographic and related databases of library catalogues. This draft specifies one category of searching within this functional area.

Functional Area A requires support of the following Z39.50 Objects:

Object	OID	Z-client	Z-server
Bib-1 attribute set	1.2.840.10003.3.1	X	X
Bib-1 diagnostic set	1.2.840.10003.4.1	X	X
USMARC record syntax	1.2.840.10003.5.10	X	X

4.5.1.1. Functional Area A: Category 1 Searching

Conformance with Category 1 Searching enables basic Author, Title, and Subject searching from the perspective of patrons and librarians.   Version 2 is required, Version 3 is recommended.  

Category 1 Searching requires support of the following Bib-1 Use Attributes:

Bib-1 Use Attributes:

Use Attribute Value	Name
4	title
21	subject heading
1003	author
1016	any

These Use Attributes are combined with other Attribute Types to express the semantics of seven Category 1 Searches.  

No default behavior (by Z-clients or Z-servers) is addressed by this profile. The profile requires that Z-clients formulate queries using all the specified attribute types and values (below), and the profile requires Z-servers to process all of the attribute types (i.e., do not ignore any attribute types or values). Z-servers that do no support specific attribute types and values must return a diagnostic.

All keyword searches are formulated according to the ZIG Commentary, a copy of which is in Appendix B. For all keyword searches using more than a single-word search term, Z-clients must formulate a multi-term keyword query by using a Boolean AND.

Z-clients and Z-servers may enhance their support of Category 1 Searching by employing the Z39.50 Scan service, and the following specifications include the use of Scan as an alternative to the exact match searches (Searches 1, 3, and 5).  Z-clients and Z-servers will negotiate the use of Scan in Init.  Scan results are relevant to a single database and thus has limited utility in broadcast searches against more than one database.

NOTE:  There is general agreement that the following searches are core searches that will be specified in the initial release of this profile, but the exact combinations of Attribute Types and Values are still under discussion.  The Bath Interoperability Meeting will discuss and finalize the attribute combinations and implications for searching.

1. Author Search -- Established Name Heading

Definition: Searches for an exact match beginning with first character for a name or parts of a name of a person or entity responsible for a particular work with right truncation. Assumes name is in some authorized form (e.g., established according to AACR2). Z-client should formulate authorized personal name query using the format:
<last name first><comma><rest of name>.

Z-servers should assume when receiving a query with Structure "normalized" that the Z-client is sending a search term that has been pre-processed and expects the Z-server to treat the search term as such.

Attribute Type	Attribute Values	Attribute Names	USMARC Fields Indexed (Suggested)
Use (1)	1003	author	100, 110, 111, 400, 410, 411, 700, 710, 711, 800, 810, 811
Relation (2)	3	equal
Position (3)	1	first in field
Structure (4)	101	normalized
Truncation (5)	100	do not truncate
Completeness (6)	1	incomplete subfield

1Alt. Scan of Author Index

Z-client will send a Scan request with that identifies the database, termlist, term, and step-size.  Termlist for this Scan is an author index.

2. Author Search -- Keyword

Definition: Searches for a word, words, character or group of characters in the name or parts of a name of a person or entity responsible for a particular work.

Attribute Type	Attribute Values	Attribute Names	USMARC Fields Indexed (Suggested)
Use (1)	1003	author	100, 110, 111, 400, 410, 411, 700, 710, 711, 800, 810, 811
Relation (2)	3 (or absent)	equal
Position (3)	3 (or absent)	any position in field
Structure (4)	2	word
Truncation (5)	*	*
Completeness (6)	1 (or absent)	incomplete subfield

* Absent or as supported by the Z-server for keyword searching

3a. Title Search -- Exact Match

Definition: Searches for an exact match beginning with the first character of first word appearing in the title of an item (i.e., names the item or the work contained in it) and contains the complete string as specified

Attribute Type	Attribute Values	Attribute Names	USMARC Fields Indexes (Suggested)
Use (1)	4*	title	130, 21X-24X, 440, 490, 730, 740, 830 840, subfield $t in the following: 400, 410, 505, 600, 610, 611, 700, 710, 711, 800, 810, 811
Relation (2)	3	equal
Position (3)	1	first in field
Structure (4)	1	phrase
Truncation (5)	100	do not truncate
Completeness (6)	3	complete field

* Use attribute value 4 should be used when Z-client wants the Z-server to search any and all title indexes.

3b. Title Search-- First Words in Title

Definition: Searches for an exact match beginning with the first character of first word appearing in the title of an item (i.e., names the item or the work contained in it).

Attribute Type	Attribute Values	Attribute Names	USMARC Fields Indexes (Suggested)
Use (1)	4*	title	130, 21X-24X, 440, 490, 730, 740, 830 840, subfield $t in the following: 400, 410, 505, 600, 610, 611, 700, 710, 711, 800, 810, 811
Relation (2)	3	equal
Position (3)	1	first in field
Structure (4)	1	phrase
Truncation (5)	1	right truncation
Completeness (6)	3	complete field

* Use attribute value 4 should be used when Z-client wants the Z-server to search any and all title indexes.

3Alt. Title Search using Scan

Z-client will send a Scan request with that identifies the database, termlist, term, and step-size.  Termlist for this Scan is a title index.

4. Title Search -- Keyword (searches using multiple search terms, queries are not limited to matching both terms in the same field).

Definition: Searches for a word, words, character or group of characters in the title of an item.

Attribute Type	Attribute Values	Attribute Names	USMARC Fields Indexed (Suggested)
Use (1)	4	title	130, 21X-24X, 440, 490, 730, 740, 830 840, subfield $t in the following: 400, 410, 505, 600, 610, 611, 700, 710, 711, 800, 810, 811
Relation (2)	3 (or absent)	equal
Position (3)	3 (or absent)	any position in field
Structure (4)	2	word
Truncation (5)	*	*
Completeness (6)	1 (or absent)	incomplete subfield

* Absent or as supported by the Z-server for keyword searching

Note: Because the profile does not allow default behavior, the ZIG Commentary on Keyword Searching may be more strictly specified for use with this profile.  This will be discussed at the Bath Interoperability Meeting.

5a. Subject Search -- Exact Match (where search term is complete subject heading but does not specify from which authoritative subject heading list). 

Definition: Searches for an exact match beginning with the first character for a word or group of words in the subject fields and contains the complete string as specified

Attribute Type	Attribute Values	Attribute Names	USMARC Fields Indexed (Suggested)
Use (1)	21	subject heading	6XX
Relation (2)	3	equal
Position (3)	1	first in field
Structure (4)	1	phrase
Truncation (5)	100	do not truncate
Completeness (6)	3	complete field

5b. Subject Search -- First Words in Subject Heading

Definition: Searches for an exact match beginning with the first character for a word or group of words in the subject fields.

Attribute Type	Attribute Values	Attribute Names	USMARC Fields Indexed (Suggested)
Use (1)	21	subject heading	6XX
Relation (2)	3	equal
Position (3)	1	first in field
Structure (4)	1	phrase
Truncation (5)	1	right truncation
Completeness (6)	3	complete field

5Alt. Subject Search using Scan

Z-client will send a Scan request with that identifies the database, termlist, term, and step-size.  Termlist for this Scan is a subject index.

6. Subject Search -- Keyword

Definition: Searches for a word, words, character or group of characters in the subject fields for an item

Attribute Type	Attribute Values	Attribute Name	USMARC Fields Indexed (Suggested)
Use (1)	21	subject heading	6XX
Relation (2)	3 (or absent)	equal
Position (3)	3 (or absent)	any position in field
Structure (4)	2	word
Truncation (5)	*	*
Completeness (6)	1 (or absent)	incomplete subfield

* Absent or as supported by the Z-server for keyword searching

Note: Because the profile does not allow default behavior, the ZIG Commentary on Keyword Searching may be more strictly specified for use with this profile.  This will be discussed at the Bath Interoperability Meeting.

7. Single Keyword Search Against Common (Z-server-defined) Access Points

Definition: Searches for a word, words, character or group of characters in fields that are commonly used as access points, as defined by the Z-server.

Attribute Type	Attribute Values	Attribute Name	USMARC Fields Indexed (Suggested)
Use (1)	1016	any	1XX, 2XX, 4XX, 5XX, 6XX, 7XX, 8XX
Relation (2)	3 (or absent)	equal
Position (3)	3 (or absent)	any position in field
Structure (4)	2	word
Truncation (5)	*	*
Completeness (6)	1 (or absent)	incomplete field

* Absent or as supported by the Z-server for keyword searching

Note: Because the profile does not allow default behavior, the ZIG Commentary on Keyword Searching may be more strictly specified for use with this profile.  This will be discussed at the Bath Interoperability Meeting.

4.5.2. Functional Area B: Search and Retrieval of Bibliographic Holdings Information

Functional Area B addresses the requirements for requesting and delivering bibliographic holdings information.  To accommodate this need, Z-clients and Z-servers area required to support the OPACs/Holdings Schema and Generic Record Syntax (GRS).  Version 2 is required and Version 3 is recommended.  

Functional Area B requires support of the following:

Z39.50 Objects:

Object	OID	Z-client	Z-server
OPAC/Holdings Schema	[to be assigned]	X	X
Bib-1 diagnostic set	1.2.840.10003.4.1	X	X
Generic Record Syntax (GRS-1)	1.2.840.10003.5.105	X	X

This Functional Area is currently under development.   See Appendix C for a discussion paper, Draft Framework for Holdings.

4.5.3. Functional Area C: Cross Domain Searching

This Functional Areas will be developed. 

5. References

Conference of European National Librarians. (1997, October 15). Z39.50 Bib-1 Attribute Set Profile for CENL, Version 1.1.
Available: <http://linnea.helsinki.fi/z3950/cenl_profile.html>.

Danish Z39.50 Implementers Group.   (1999, March 4).  DanZIG Z39.50 Profile.  Available: <http://www.bs.dk/danzig/profil.htm>

Lunau, Carrol D. and Zeeman, Joe. (1998, October 15). Z39.50 Basic Searching of Bibliographic Systems: A Discussion Document.

MODELS Library Interoperability Profile Family.    Available: <http://www.ukoln.ac.uk/dlis/models/clumps/technical/zprofile/zprofile.htm>

Moen, William E. (1999).  Handling Bibliographic Holdings Information in Z39.50: Implications for the Z Texas Profile <http://www.tsl.texas.gov/LD/z3950/Holdings27Mar99.htm>.

National Information Standards Organization. (1995). ANSI/NISO Z3950-1995. Information Retrieval (Z39.50): Application Service Definition and Protocol Specification. Bethesda, MD: NISO Press. Electronic version of Z39.50 available at the Z39.50 Maintenance Agency.   Available:  <http://lcweb.loc.gov/z3950/agency>.

National Library of Canada. (1998, January 26). Virtual Union Catalogue Z39.50 Profile, Draft Version 1.5.  Available: <http://www.nlc-bnc.ca/resource/vcuc/profil4.htm>

Husby, Ole. (1997, January 9). BIB-1 profile for ONE   Available: <http://www.bibsys.no/one-wg/bib-1.profile.html>

St-Gelais, Rolande. (1999, March 10). OPAC/Holdings Schema: Definitions and Examples.  Available: <http://www.dra.com/resources/z3950/opac-schema-defs-3.pdf>.

Texas Z39.50 Implementors Group. (1999, April).   Z Texas Profile: A Z39.50 Profile for Library Systems Applications in Texas, Release 1.0.   <http://www.tsl.texas.gov/LD/z3950/TZIGProfile99Apr20.htm>.

Z39.50 Maintenance Agency. Available: <http://lcweb.loc.gov/z3950/agency>.

Z39.50 Maintenance Agency. ATS-1 Profile. Available: <http://lcweb.loc.gov/z3950/agency/profiles/ats.html>.

Z39.50 Maintenance Agency (1995, September). Attribute Set Bib-1 (Z39.50-1995): Semantics. Available: <ftp://ftp.loc.gov/pub/z3950/defs/bib1.txt>.

Z39.50 Maintenance Agency. (1998, June). Bib-1 Attribute Set Available: <http://lcweb.loc.gov/z3950/agency/defns/bib1.html >

Appendix A
Definition of Title and Subject Searches

Action Item from ZILS Working Group Teleconference #2
Judith Pearce, National Library of Australia, 14 July 1999

Requirements for Searching

• Title searches include phrase and keyword searching of titles. Phrase searching includes both exact match on title and first words of title. A search on title keywords will retrieve all titles containing the specified keywords. A search on first words of title will retrieve titles that begin with the specified string. A search on title exact match will retrieve titles that contain the complete string as specified.
• Subject searches include phrase and keyword searching of subject headings. A subject heading is defined as the complete subject heading, including all subfields. Phrase searching includes both exact match on subject heading and first words of subject heading. A search on subject keywords will retrieve all subject headings containing the specified keywords. A search on first words of subject heading will retrieve subject headings that begin with the specified string. A search on subject exact match will retrieve subject headings that contain the complete string as specified.

First words of title is included as a category 1 searching requirement because MARC tags indexed to support this access point will include both the title proper and the remainder of title. A user searching for the journal "Rock Mechanics" will use this access point to retrieve the title "Rock Mechanics: Journal of the International Society for Rock Mechanics" as recorded in the catalogue. Exact match on title is a specialist search included to support the requirement for short title searching. A user will use this access point to retrieve the small number of works with the title " Times" as against the hundreds of works beginning with the word "Times".

First words of subject is included as a category 1 searching requirement because MARC tags indexed to support this access point will include both the main subject heading and all subdivisions. A user searching on the subject "mathematical models" will also retrieve "Mathematical Models -- Dictionaries". While a subject keyword search on mathematical models will retrieve this work it will also retrieve many works where the term "mathematical models" has been applied as a subdivision. Exact match on subject heading is a useful search that enables retrieval to be restricted to a precise subject heading. This feature may be used to retrieve general works on a subject or to find other works indexed with exactly the same subject heading. This access point may be used (in preference to first words of subject) when users exploit links from subject headings in displayed records, where this feature is supported by a client implementation.

Many OPACs use the capability to browse indexes as an alternative to the first words of title and first words of subject search. However, this feature is unlikely to be available for searching across multiple catalogues using Z39.50. It is hard to imagine how the SCAN service could be implemented in this scenario.

Conformance with subject keywords and first words of subject may require some libraries to build new indexes that are based on the bibliographic record rather than the authority file.

Draft attributes for exact match and first words searches for both title and subject

We still need Joe to clarify exactly what you get with the following combinations and whether this adds any additional functionality to title and subject searches:

• Truncation (right truncation) / Completeness (complete field)
• Truncation (right truncation) / Completeness (incomplete field)
• Truncation (do not truncate) / Completeness (incomplete field)

(Do we need incomplete subfield at all?)

Appendix B
ZIG Commentary
Z39.50 Keyword Searching of Bibliographic Systems

Issue raised by: Fay Turner

Synopsis:

Keyword searching of bibliographic systems is useful when neither the exact author, title or subject of an item is known, or when the objective of the search is to identify a broad range of documents to which the keywords apply. Characteristics of keyword searching include:

• Searching one or more words that do not have to be in any order.
• Searching for words anywhere in some or all fields indexed for the record, usually in author, title, subject and sometimes notes fields.
• Limiting the search, for example, by author, subject, title, format, language, publisher or date.
• Combining several keywords using Boolean and positional operators and expressing the relationship between the keywords: AND, OR, AND-NOT, exclusive-or, adjacent, near, with, same, etc.
• Nesting searches to construct complex searches, truncation of terms, etc.
  

Many bibliographic systems support some or all of these features.

What combination of bib-1 attributes should be used to specify a keyword search?

Response:

For bibliographic system searching, queries where all operands are constructed as follows are to be interpreted as keyword searches:

• Position: any-position-in-field (3) or absent
• Relation: equal (3) or absent
• Completeness: incomplete-field (1) or absent
• Structure: word (2)
• Truncation: absent or as supported by the Z-server for keyword searching
• Use: as supported by the Z-server for keyword searching (including ANY, if supported).
• Term: a single word

Appendix C
Draft Framework for Holdings

Action Item from ZILS Working Group Teleconference #2
Judith Pearce, National Library of Australia, 15 July 1999

Retrieval of Bibliographic Holdings Information

This draft document addresses the requirements for requesting and delivering bibliographic holdings information.

Scenarios supported in this draft include:

1. Determine which organisations have one or more copies of an item.
2. Determine the extent of run of a serial item held by an organisation.
3. Locate precisely a single copy of an item held by an organisation.
4. Retrieve detailed serial holdings information to the issues level.
5. Obtain availability and circulation information.
6. Obtain the Library's lending policy for an item.
7. Distinguish holdings items per bibliographic unit type.

Scenarios for which conformance requirements still need to be developed include:

1. Obtain number of copies held per location.
2. Locate item in a group of organisations based on a filter such as geographic location, type or region.
3. Sort holdings.
4. Update holdings.

Working note: Update could be seen as separate from the other resource discovery/ document delivery type scenarios and therefore better handled through a new functional area. However, it will be important to review the update requirements before making any decision to remove the MARC for Holdings option from the OPAC/Holdings schema.

All implementations claiming conformance in this functional area must support the Z39.50 Objects defined for Online Catalogue Search and Retrieval.These are:

Additional objects relevant to this functional area are:

Requirements for support for these objects will vary depending on whether the implementation is:

• A Single Library Catalogue.
• A Virtual Union Catalogue.
• A Physical Union Catalogue with a Holdings Database.
• A Physical Union Catalogue without a Holdings Database.

The following table is a first cut at summarising these requirements.

Version 2 Z-clients and servers would be expected to recognise the Element Set Names (ESNs) identified as mandatory in this table once these have been finalised. A future version of this profile may include a specification enabling Version 3 Z-clients to use CompSpec and eSpec-1 to request the appropriate combination of elements from a Version 3 Z-server.

It would be a requirement for conformance for both Z-clients and Z-servers that an appropriate diagnostic be returned for optional functions defined below that the implementation did not support.

Working notes

• In the table, I have assumed for shorthand purposes that the specified ESNs are independent of the OPAC/Holdings schema and that an alternative response to the OPAC/Holdings schema may be delivered in some cases.
• For the scenarios that ESN Level B-1 and ESN Level B-2 have been developed to address, I argue that a valid response for all four scenarios may be holdings fields in MARC bibliographic record syntax, with the proviso that the use of locally-defined fields to carry holdings information is deprecated.
• For all ESNs a valid response may also be a MARC for Holdings record, either in the GRS-1 record syntax using the OPAC/Holdings Schema or (argued with less confidence) as a discrete record in a MARC for Holdings record syntax.
• If separate ESNs are needed for the scenarios that involve MARC record syntaxes, there should be a direct mapping between the levels required in the response.
• ESNQBib is included as a shorthand to cover the scenario where the Z-client has asked for the holdings information in GRS-1 using the Opac/Holdings schema and wishes to specify inclusion of the bibliographic record with the holdings information. In practice, all ESNs may need to be duplicated with a Bib qualifier. [In this scenario -- using Jo Zeeman's words -- "if the server supplies the bibliographic portion in a format other than desired (or omits it), the client may follow with another present in order to get the bibliographic part of the desired format. This subsequent retrieval would be done via the classical method, that is, specifying the desired format as the record syntax, not GRS-1".]
• ESNQMARC is included as a shorthand to cover the scenario where the Z-client has asked for the holdings information in GRS-1 using the Opac/Holdings schema and wishes to specify that the local holdings data are supplied as one or many MARC for holdings records. This may or may not be a requirement, depending on whether the option to include MARC for holdings records in the OPAC/Holdings schema is retained.

References

Handling Bibliographic Holdings Information in Z39.50: Implications for the Z Texas Profile (William E. Moen)

OPAC/Holdings Profile 1: Profile for retrieving detailed library holdings in a bibliographic environment (Draft 1, January 15, 1998)

OPAC/Holdings Schema: Definitions and Examples (Fourth Draft -- July 1999)

Scenarios for the Application of the OPAC/Holdings Proposal (Draft 1, prepared by J.C. Zeeman, October 28, 1997)

Z Texas Profile: A Z39.50 Profile for Library Systems Applications in Texas (Release 1.0, Stable Draft, April 20, 1999)

Date of last edit: 08/24/99