The CIMI Profile:

Z39.50 Application Profile Specifications for

Use in Project CHIO

Draft Version 3







Prepared for the

CIMI Profile Public Review & Comment Period
June 1- July 15, 1996


by

The CIMI Profile Development Working Group

William E. Moen
CIMI Profile Development Project Manager
School of Library and Information Sciences
University of North Texas
Denton, TX 76203
email: <[email protected]>
voice: 817-565-3563
fax: 817-565-3101



CONTENTS

1. Introduction
2. Background
3. Scope and Field of Application
4. Normative References
5. Definitions
6. The Digital Collections Profile and the CIMI Profile
6.1. Two Models Underlying the CIMI Profile
6.2. Navigation of the CHIO Collection
7. Z39.50 Specification for the CIMI Profile
7.1. Protocol Version
7.2. CIMI Objects
7.3. Communication Services
7.4. Z39.50 Services
7.4.1. Init
7.4.2. Search
7.4.2.1. Attribute Sets
7.4.2.2. Proximity Operation
7.4.3. Retrieval
7.4.3.1. CIMI Descriptive Record Schema
7.4.3.2. Abstract Record Structure
7.4.3.3. Tag Types
7.4.3.4. Element Set Names
7.4.3.4.1. briefCimi
7.4.3.4.2. briefCimiHtml
7.4.3.4.3. {target token}
7.4.3.4.4. fullCimiObject
7.4.3.4.5. fullCimiObjectHtml
7.4.3.4.6. Element Set Names to Request Highlighting
7.4.3.5. Record Syntaxes
7.4.3.5.1. GRS-1
7.4.4. Access Control
7.5. Diagnostic Messages
7.6. Conformance

Appendix A: CIMI-1 Attribute Set
Appendix B: Values for PrivateProximityUnit
Appendix C: Information to Include in Brief Record
Appendix D: Non-Protocol Agreements on Client and Server Behavior/Functionality
Appendix E: Examples of CHIO Collection Resources

1. INTRODUCTION

This document describes an application profile for the use of ANSI/NISO Z39.50-1995, Information Retrieval (Z39.50): Application Service Definition and Protocol Specification (National Information Standards Organization, 1995) for search and retrieval of museum information in digital collections. This profile is named the CIMI Profile, where CIMI refers to the Consortium for the Computer Interchange of Museum Information. CIMI is the sponsor of a demonstration project, Project CHIO (Cultural Heritage Information Online). The CIMI Profile is intended to be tested and revised based on the Project CHIO demonstration implementations. Developments associated with the CIMI Profile, such as the CIMI Attribute set for searching museum information, may have utility outside of Z39.50 implementations of the CIMI Profile.

The CIMI Profile includes not only the specifications for Z39.50 in this application but also other aspects of CIMI conformant clients and servers that are outside the scope of Z39.50.

2. BACKGROUND

CIMI initiated Project CHIO as a demonstration project to enable users to search for and retrieve cultural heritage museum information from disparate and distributed information systems. Project CHIO consists of two interrelated demonstration projects -- CHIO Structure and CHIO Access -- to show respectively the utility of the Standard Generalized Markup Language (SGML) and Z39.50.

CHIO Access demonstrates the utility of Z39.50 to search and retrieve museum information captured in digital form (e.g., collections records, exhibition catalogues, images, wall text, etc.). CHIO Access demonstrates how Z39.50 offers solutions to the difficulties in achieving meaningful online searching and retrieval of museum information of different types and structure (e.g., structured records, full-text documents, images) regardless of the hardware and software used to store the data or search for it.

To implement Z39.50 in this application, a working group is developing the CIMI Profile. The CIMI Profile Working Group consists of Z39.50 experts, information systems experts, and experts in museum systems and museum information resources. The specifications included in the CIMI Profile reflect the consensus of this group and input from a range of stakeholders.

Because the CIMI Profile is being tested in the Project CHIO demonstration project rather than working implementations, the CIMI Profile Working Group agreed that a pragmatic approach should be used in developing the Z39.50 specifications. This agreement has important consequences:

The development of the CIMI Profile is documented in The CIMI Profile: Development and Specifications of a Z39.50 Application Profile for Use in Project CHIO (Moen, 1996).

3. SCOPE AND FIELD OF APPLICATION

The CIMI Profile specifies a subset of Z39.50 features, options, and parameters needed to support the functional and user requirements for the search and retrieval of museum information in digital collections. The initial CIMI Profile will be a draft specification to guide prototype Z39.50 implementations that support the demonstration needs of Project CHIO. CIMI clients will be able to interconnect with any CIMI server, and these clients will behave in a manner that allows interoperability with a CIMI server. Clients that support Z39.50 but do not implement the CIMI Profile will be able to access CIMI servers but with less than full CIMI functionality.

The initial Z39.50 implementations of the CIMI Profile will consist of Z39.50 clients and servers operating in the Internet environment and supporting access to a demonstration Project CHIO Collection. The CHIO Collection can be modeled as a hierarchical, distributed collection of digital information. The CHIO Collection consists of a number of physical and/or logical datastores of museum information, and the datastores may consist of one or more databases. A user may search the CHIO Collection to retrieve digital representations of museum information such as exhibition catalogues, wall texts, object records, images with associated text. These representations may be compound documents comprising multimedia formats of resources. Users may also access ancillary resources such as thesauri, union lists of artist names, and authority records. (Examples of CHIO Collection resources will be included in a subsequent version of the CIMI Profile.) Through Project CHIO, these digital museum resources in various data types will be accessible through Z39.50 clients and servers supporting the CIMI Profile (see 7.4.3.).

The CIMI Profile is a companion profile to the Z39.50 Profile for Access to Digital Collections (Library of Congress, 1996), which means that the CIMI Profile specifies compatible extensions to the Z39.50 Profile for Access to Digital Collections (hereafter referred to as the Collections Profile). The CIMI Profile utilizes features of the Collections Profile for navigation of the CHIO Collection and for retrieval of descriptive information about CHIO resources.

The CIMI Profile addresses Z39.50 search and retrieval in the CHIO Collection (e.g., intersystem interactions and information interchange) but does not specify user interface requirements, the internal structure of databases that contain the digital information objects, or search engine functionality.

4. NORMATIVE REFERENCES

The following list contains documents that contain provisions which, through reference in this text, constitute provisions of the CIMI Profile. At the time of this publication, the editions indicated were valid. All documents are subject to revision, and parties to agreements based on the CIMI Profile are warned against automatically applying any more recent editions of the documents listed below, since the nature of references made by the Profile to such documents, is that they may be specific to a particular edition. In addition, this list contains other documents that can be consulted for further information, background, etc.

[1] American National Standards Institute. (1985). American National Standard Z39.2­1985 Bibliographic Information Interchange. New York: American National Standards Institute.

[2] Consortium for the Computer Interchange of Museum Information. (1996). Full Text Document Type Definition. Will be available July 1996 as:
URL: ftp://ftp.cimi.org/pub/cimi/CIMI_SGML/

[3] Art Information Task Force. (1995). Categories for the Description of Works of Art. Santa Monica, CA: Art History Information Program Publications

[4] International Council of Museums. International Committee for Documentation (1995). CIDOC Relational Data Model. CIDOC Data Model Working Group: Washington, D.C.

[5] Library of Congress. (1996). Z39.50 Profile for Access to Digital Collections. Available as HTML or Postscript files:
URL: http://www.loc.gov/z3950/agency/profiles/collections.html
or URL: ftp://ftp.loc.gov/pub/z3950/profiles/collections.ps

[6] Lynch, Clifford A. (1994). RFC 1729, Using the Z39.50 Information Retrieval Protocol in the Internet Environment. Available as: URL: http://ds.internic.net/rfc/rfc1729.txt

[7] Moen, William E. (1996). The CIMI Profile: Development and Specifications of a Z39.50 Application Profile for Use in Project CHIO. Available in various formats at:
URL: ftp://ftp.cimi.org/pub/cimi/CIMI_Profile/Background

[8] National Information Standards Organization. (1995). ANSI/NISO Z3950-1995. Information Retrieval (Z39.50): Application Service Definition and Protocol Specification. Bethesda, MD: NISO Press.

[9] USMARC Format for Bibliographic Data. Washington, DC: Library of Congress, Cataloging Distribution Service.

[10] Z39.50 Implementors Agreement. (1996). Returning Diagnostics in an Init Response. Z39.50 Implementors Agreements are available from the Z39.50 Maintenance Agency,
URL: http://www.loc.gov/z3950/agency/agree.html

5. DEFINITIONS

For purposes of this Profile, the following definitions apply. For definitions of Z39.50 terms and concepts not listed here, see ANSI/NISO Z39.50-1995, Information Retrieval (Z39.50): Application Service Definition and Protocol Specification (National Information Standards Organization, 1995). For definitions of terms and concepts related to the Collections Profile, see Z39.50 Profile for Access to Digital Collections (Library of Congress, 1996).

Artist's Record. Biographical information about an artist, including name variations, nationality or citizenship attributions, birth and death dates and places, various activity dates, information about the artist's family, friends and associates, and major life events.

Associated Description. A unit of descriptive information associated with a collection or object, for example, an encoded archival description or a finding aid (for an archival collection), or a cataloging record (for library materials). (Collections Profile).

Authority record. A record holding bibliographic or authoritative reference information from a reputable source on an object, person, event or place.

Cataloging record. Refers to records of bibliographic information describing or representing books or other bibliographic entities.

Collection. A group of related objects and/or collections, possibly distributed across locations. A collection is a tree, where leaf nodes are objects and non-leaf nodes are subcollections. (Collections Profile).

Collection Descriptive Record. See Collections Profile.

Companion Profile. The CIMI Profile is a companion profile of the Collections Profile. See Collections Profile.

Context Collection. Collection A is a context collection for collection or object B if it is a superior, related collection, and the organization with responsibility for the management of B considers that although collection A may be relevant to a user who is interested in B, any collection superior to collection A is likely not relevant. (Collections Profile).

Descriptive Record. A unit of descriptive information at a higher level of abstraction than an Associated Description. A Descriptive Record may include one or more Associated Descriptions in addition to other information that describes either a collection (and possibly its contents) or an object within a collection. A Descriptive Record is either a Collection Descriptive Record or an Object Descriptive Record. (Collections Profile).


Exhibition Catalogue. A publication accompanying a museum exhibition that provides detailed information about the theme of the exhibition and the objects on display. The publication may include exhibition dates at other venues if the exhibition is scheduled to travel. The publication may provide extensive information about the socio-historical context of the objects on display. If the exhibition has works by different artists, information about the artists may be given. If the exhibition is about the work of a single artist, the exhibition catalogue may contain an extensive biography. The catalogue may provide information about the history of ownership or provenance of the objects being displayed in the exhibition.

Object Record. A record that provides descriptive information about a museum object and its component parts, measurements, weight, creation and creator, ownership, history of use, materials and techniques used in its manufacture, inscriptions, identifying numbers, historical context, rights and restrictions, credit line for display or publication, in short all information concerning an object. An object record enables a museum to be accountable for and to uniquely identify an object.

Object Descriptive Record. (See Collections Profile.).

Profile. A set of implementor agreements specifying the use of a particular standard (or group of standards) to support a particular application, function, community, environment, or class of information. A profile selects options, subsets, values of parameters, etc., where these choices are left open in the standard, and where these selections are necessary to accomplish identified functions. A profile may also specify aspects of client and server behavior that are beyond the scope of the base standards. Purpose of a profile include: (1) to provide a specification for vendors to build to, resulting in products that will interoperate; and (2) to provide a specification that customers may reference for procurement purposes.

Scope of Collection Record. Museum collections can be characterized in many ways. The entire holdings of a museum might be considered its "collection" or a collection may be any lesser number of related items -- related or similar objects, paintings by a single artist, the gifts of single donor, objects associated with a particular event or timespan, etc. A Scope of Collection record provides the ability to record the existence of significant collections by a variety of criteria.

Wall Text. Supporting words, images, and sound provided in the context of an exhibition to provide the viewer with information about the objects being exhibited. Wall text provides information about the context, meaning or significance of a group of objects or thematic section of an exhibition.

6. THE COLLECTIONS PROFILE AND THE CIMI PROFILE

The CIMI Profile is a companion profile to the Z39.50 Profile for Access to Digital Collections (Library of Congress, 1996). The CIMI Profile requires conformance with the Collections Profile at the level of Basic Conformance.

The Collections Profile provides the framework for discovering digital collections and objects (where a collection is defined as "a group of related objects and/or collections, possibly distributed across multiple locations"), for navigating within a collection and to other collections, and for obtaining descriptive information about collections or objects. The Collections Profile provides semantics for navigating digital collections. The Collections Profile defines organizational structures that describe collections and objects to represent otherwise undifferentiated digital data as hierarchical, distributed collections. The Collections Profile enables the search of structured descriptive information about collections and objects, the enumeration of objects and subcollections within in a collection, and the retrieval of a Descriptive Record for a collection or object.

The CIMI Profile specifies the use of Z39.50 for search and retrieval of museum information. The CIMI Profile focuses on Z39.50 specifications for searching datastores of museum information, where that information can take the form of digitized images, structured records, plain text, and encoded text.

It is important to note that the Collections Profile is primarily concerned with Descriptive Records, and it treats Digital Objects as "opaque," in that the Collections Profile neither addresses the structure of Digital Objects nor addresses the search and retrieval of them (except where a Digital Object is included as an element in an Object Descriptive Record or when a Digital Object is represented as a Z39.50 database record searchable by a known unique identifier). Given that the CIMI Profile is a companion profile to the Collections Profile, the following statements are warranted:

  1. The Collections Profile provides the Z39.50 specification for discovery of collections and objects in, and navigation within, the CHIO Collection.
  2. The Collections Profile defines the basic structure of Descriptive Records for Collections and Digital Objects. The CIMI Profile includes implementation agreements of those Descriptive Records for search, retrieval, and other data manipulation of the Digital Objects that comprise the CHIO Collection.
  3. The CIMI Profile addresses the search and retrieval of Digital Objects in the CHIO Collection (i.e., they are not considered opaque in the CIMI Profile even though they are considered opaque in the Collections Profile).

The scope of the CIMI Profile extends the functionality provided by the Collections Profile to search and retrieve parts or all of the Digital Objects in the CHIO Collection. The Descriptive Record structure will be used by the CIMI Profile to return both machine-processible elements and elements intended for display to the user.

6.1. Two Models Underlying the CIMI Profile

The Collections Profile introduces a model, terminology, principles, and assumptions that the CIMI Profile inherits as a companion profile. The Collections Profile's model provides a way of viewing digital collections for search and navigation. It introduces the concepts of Descriptive Records for objects (Object Descriptive Records) and for collections (Collection Descriptive Record). It also defines a Associated Description, which refer to a class of descriptive aids describing Collections and Objects. An assumption of the Collections Profile is that Collections and Objects are of primary interest to users. Associated Descriptions are assumed to be ancillary (as Associated Description, they are not members of a Collection as defined by the Collections Profile) to collections and objects and may be used in a number of ways (e.g., to derive information for the Descriptive Records). The Associated Description may be useful as an aid by the user to determine whether the object or collection it describes is or is not of interest, and as a navigational and locational aid.

The structures that the Collections Profile identifies as Associated Descriptions are not used by the CIMI Profile. Instead, some of these types of structures are viewed by the CIMI Profile as Digital Objects. The specifications of the CIMI Profile support search and retrieval of these Digital Objects.

6.2. Navigation of the CHIO Collection

Z39.50 servers supporting the CIMI Profile will provide a well-known database named CHIO that will contain at least one Collection Descriptive Record and may contain other Collection and Object Descriptive Records that are appropriate for describing collections and objects available in the CHIO Information Resource. The Collection Descriptive Record may describe a context collection to provide a starting point for navigating and searching the CHIO Collection. Specifications outlined in the Collections Profile will be used for searching and retrieving Collection and Object Descriptive Records (e.g., using the Collection-1 attribute set).

7. Z39.50 SPECIFICATIONS FOR THE CIMI PROFILE

This section details the required services available from Z39.50, describes Attribute Sets for searching digital objects, the processes by which the client may request records or elements of records to be returned, and the required the record syntax that CIMI servers will support for the transfer of digital objects from the CHIO Collection.

7.1. Protocol Version

The CIMI Profile requires that clients and servers support Z39.50 Version 2 as specified in Z39.50­1995.

7.2. CIMI Objects

The CIMI Profile requires that a client, a server, or both of them support the Z39.50 objects as indicated below

ObjectOID ClientServer
Bib-1 attribute set{ANSI-standard-Z39.50 3 1}X
Collection-1 attribute set{ANSI standard Z39.50 3 7} XX
CIMI-1 attribute set:{ANSI standard Z39.50 3 8}XX
Bib-1 diagnostic set:{ANSI standard Z39.50 4 1}XX
GRS 1 record syntax:{ANSI standard Z39.50 5 105}XX
*CIMI access control format{ANSI-standard-Z39.50 #####X
Collections Schema{ANSI-standard-Z39.50 13 XX
Collections TagSet{ANSI-standard-Z39.50 14 [####]}XX

For explicit conformance requirements regarding the support of these objects, see Section 7.6, Conformance.

*[Note: See Section 7.4.4.]

7.3. Communication Services

When the Transmission Control Protocol (TCP) is used as the transport service, the specification for Z39.50 implementation's use of TCP is found in RFC 1729, Using the Z39.50 Information Retrieval Protocol in the Internet Environment (Lynch, 1994). The use of other communication services is not defined in the CIMI Profile.

7.4 Z39.50 Services

Four Z39.50 (Version 2) services are required for conformance:

Both clients and servers are required to support Init, Search, and Present. Access Control must be supported by clients; servers are not required to support Access Control.

No additional services are required for conformance to the CIMI Profile. Other Z39.50 services, however, may be provided optionally by servers and used by clients.

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

[Note: One other Z39.50 service, Scan, will likely be profiled. Scan allows a client to scan an ordered term-list (e.g., subject terms, names, titles, etc., contained in indexes). Scan provides a useful service and profiling Scan will be based on implementation experience.]

7.4.1. Init

IDAuthentication parameter in Init is required to allow userid and password information to be passed from the client to the server. No other security requirements are specified in the CIMI Profile.

Clients supporting the CIMI Profile are required to implement IDAuthentication as recommended in the comment in Z39.50-1995 (p. 50) and should be able to select either "idPass" or "anonymous" as the CHOICE. The server may return a diagnostic if Init fails. The return of the diagnostic in the userInformationField parameter is governed by the Z39.50 Implementors Agreement, Returning Diagnostics in an Init Response available at the Z39.50 Maintenance Agency list of Implementors Agreements.

In the interests of interoperability, servers should accept the absence of IDAuthentication as an implicit "anonymous."

7.4.2. Search

The CIMI Profile requires support for Z39.50 Type 1 queries which are general purpose Boolean query structures. Additionally, support of Z39.50 Type 101 queries is required to accomplish searches involving proximity operators defined in this Profile.

The CIMI Profile requires support for named result sets (resultSetName parameter). Servers will support a minimum of two (2) concurrent result sets (this implies support for named results sets). Support for two (2) or more named result sets will facilitate the display of old result sets. Servers will also support a query which includes a single result set name ANDed with other operands. Such support will enable iterative searches.

7.4.2.1. Attribute Sets

The CIMI Profile requires that clients and servers support two registered attribute sets, CIMI-1 which is defined in this Profile and Collection-1 which is defined the Collections Profile. The CIMI attribute set has been developed based on a examination of the searching requirements for and data structures of museum information. The CIMI attribute set provides a sufficient basis for searching the subset of museum information represented in the CHIO Collection.

The CIMI attribute set imports several Bib-1 attributes. In addition to support for the CIMI attribute set (including the imported Bib-1 attributes), a server is required to support the Bib-1 attribute set at least to the following extent: A server should accept a search that specifies the Bib-1 attribute set OID and that includes only Bib-1 attributes that are imported to the CIMI attribute set. The server may, if appropriate to its database, support additional attributes from Bib-1.

The CIMI Profile defines the CIMI-1 attribute set (see Appendix A for the CIMI-1 attribute set). CIMI-1 imports all attribute types and selected attribute values from Bib-1 attribute set. CIMI-1 defines two new attribute types, Context and Authority. Additional Use attribute values defined for the CIMI Attribute Set are derived from the CIMI Full-Text DTD access points.

Not all servers will support all attribute values and types from the CIMI attribute set. Servers are required to return an appropriate diagnostic when it receives a query containing attribute values and types not supported. Appropriate diagnostics are listed in Bib-1 Diagnostic set and include, for example:

The CIMI Profile addresses only Z39.50 Version 2 implementations. Version 2 does not allow a query to combine attributes from different attribute sets, therefore multiple, iterative searches may be required when CIMI-1 and Collection-1 attributes are needed to express a search. In Version 3, attributes may be combined from different attribute sets within a single query. The support of named result sets addresses CIMI searching requirements while using Version 2, since a result set may be used as an operand in a subsequent query.

7.4.2.2. Proximity Operation

The CIMI Profile defines a private set of codes for proximityUnitCode to express complex searches where the query cannot be expressed in terms of nesting. The source for the PrivateProximityUnit is the values of the Context Attribute as listed in CIMI-1. Appendix B enumerates the values for PrivateProximityUnit. These are to be used as the value of 'private' (CHOICE [2]) of 'proximityUnitCode' of datatype ProximityOperator defined in Z39-50-APDU-95.

7.4.3. Retrieval

This section describes the components and procedures using Z39.50 to return records in response to a query.

The variety of information resources and their possible associated information types entails special requirements for the retrieval of the information objects. The following table presents the finite set of resources, information objects, and their associated data types that will be found in the CHIO Collection.

CHIO Resources
Information Objects
Data Type(s)
Object RecordsInventory records MARC, SGML, other structured records
Exhibition catalogues, Wall TextEncoded text SGML
Art Objects (e.g., pictures, paintings, etc.) Digitized objects (with associated text) Image, Audio, Video
Cataloging RecordsBibliographic records MARC
Union Lists, Thesauri, Artist's Records Authority recordsMARC, SGML, other structured records
Scope of Collection RecordsCollection records SGML, other structured records

Different information resources may satisfy the criteria of a single query, and one requirement is to return for display to the user a list (in brief form) of these satisfying information resources. Since the resources may be of different data types, the CIMI Profile specifies a mechanism to return these resources in a neutral format (i.e., not requiring separate applications to be launched on the client side to display, for example, separate records for a SGML object, a digitized image, and a cataloging record in MARC. To accommodate this requirement, the CIMI Profile specifies a the use of the Descriptive Record defined by the Collections Profile.

7.4.3.1. CIMI Descriptive Record Schema

The CIMI Profile specifies that the Collections Descriptive Record Schema will be used as the basis for one aspect of retrieval, namely identifying a set of elements from various CHIO information objects without regard to local data structures or types of those objects. The CIMI Profile extends the use of the Descriptive Record Schema for use in this application. In the CIMI Profile, a Descriptive Record may describe either a collection or an object.

This section defines a CIMI Descriptive Record Schema for a record to support the requirement that records representing different information objects and associated data types are displayed in brief format on a single screen.

7.4.3.2. Abstract Record Structure

The abstract record structure for the CIMI Descriptive Record is defined as follows. Semantics for elements not listed here are defined in the Collections Profile.

Element*Occurrence RepeatableDatatype
typeOfDescriptiveRecordmandatory noINTEGER
briefDescriptionoptional noBriefTextDescription
collectionInfooptional; occurs if and only if typeOfDescriptiveRecord is 1 noCollectionInfo
objectInfooptional; occurs if and only if typeOfDescriptiveRecord is 2 noObjectInfo
associatedDescriptionoptional yesAssociatedDescription
relatedCollectionoptional yesRelatedCollection

*See Collections Profile for semantics of the elements.

Element typeOfDescriptiveRecord is one of the following:

1 = Describes a collection
2 = Describes an object

Element briefDescription is a description of the object addressed by the Descriptive Record. The CIMI Profile provides guidance on the contents of this element. See Appendix C, Information to Include in Brief Record.

The use of associatedDescription is not specified in the CIMI Profile and may be ignored if received by a client. It is included because of interoperability considerations with other companion profiles.

The following tables provide definitions for the datatypes of the Descriptive Record Schema elements:



Datatype CollectionInfo is structured as follows:

Element*Occurrence RepeatableDatatype
collectionNamemandatory noInternationalString
databaseoptionalyes ServerAndDb
enumeratedMemberoptional yesEnumeratedMember
fullyEnumeratedoptional noBOOLEAN
childrenKnowThisParentoptional noBOOLEAN

*See Collections Profile for semantics of the elements.

Datatype ObjectInfo is structured as follows:

Element*Occurrence RepeatableDatatype
typeOfObjectmandatory noINTEGER
categoryOfObjectmandatory noInternationalString
digitalObjectoptional; may only occur if typeOfObject is less than 4 noDigitalObject

*See Collections Profile for Semantics of the elements.

Element typeOfObject is one of the following:

0 = Unspecified
1 = Object is a Digital Object
2 = Object is a Digital Object for which there is a Physical Object
3 = Object is a Digital Object for which there is no Physical Object (for example when an object is originally created electronically)
4 = Object is a Physical Object
5 = There is no (separate) object.

Element categoryOfObject indicates that a Digital Object is one of the following categories of CHIO Resources:

cimi:unspecified
cimi:artist record
cimi:cataloging record
cimi:exhibition catalogue
cimi:image
cimi:object record
cimi:scope of collection record
cimi:wall text

Some of these resources may include several forms of material, for example, an exhibition catalogue may be comprised of encoded text and embedded or linked images.



Datatype RelatedCollection is structured as follows:

Element*Occurrence RepeatableDatatype
collectionmandatory noCollectionAndDb
relationshipoptional yesINTEGER
superiorLeveloptional noINTEGER
descriptionOfRelationshipoptional noInternationalString

*See Collections Profile for semantics of the elements.

Datatype CollectionAndDb is structured as follows:

Element*Occurrence RepeatableDatatype
collectionNamemandatory noInternationalString
databasemandatoryno ServerAndDB

*See Collections Profile for semantics of the elements.

Datatype ServerAndDb is structured as follows:

Element*Occurrence RepeatableDatatype
serveroptional; if omitted, the indicated database is assumed to be on this server; must be supplied if indicated database is not on this server noInternationalString
dbmandatoryno InternationalString

*See Collections Profile for semantics of the elements.

Datatype EnumeratedMember is structured as follows:

Element*Occurrence RepeatableDatatype
briefDescriptionOfMemberoptional noBriefTextDescription
pointeroptionalno RecordPointer
whatPointerPointsTooccurs if and only if pointer occurs noINTEGER

*See Collections Profile for semantics of the elements.

Datatype Digital Object is structured as follows:

Element*Occurrence RepeatableDatatype
actualDOmandatoryno variant dependent; default is OCTET STRING
authoritativeoptional noBOOLEAN

*See Collections Profile for semantics of the elements.

Datatype RecordPointer is structured as follows:

Element*Occurrence RepeatableDatatype
databaseoccurs (in conjunction with recordID) if and only if alternative Identifier does not occur noServerAndDb
recordIDoccurs (in conjunction with database) if an only if alternative Identifier does not occur noOCTET STRING
alternativeIdentiferoccurs if and only if Database does not occur noOCTET STRING
typeOfIdentifieroptional; occurs only if alternativeIdentifier occurs noOBJECT IDENTIFIER or InternationalString

*See Collections Profile for semantics of the elements.

Datatype BriefTextDescription has type InternationalString.

The following provides a schematic representation of the Descriptive Record

typeOfDescriptiveRecord
briefDescription
collectionInfo
-collectionName
-database (repeatable) (server and db)
-enumeratedMember (repeatable)
--briefDescriptionOfMember
--pointer (database and recordID) XOR (alternativeIdentifier and typeOfIdentifier)
--whatPointerPointsTo
-fullyEnumerated (BOOLEAN)
-childrenKnowThisParent (BOOLEAN)
objectInfo
-typeOfObject
-categoryOfObject
-digitalObject
--actualDO (with variant and metadata information)
--authoritative
associatedDescription
relatedCollection (repeatable)
-collection (collectionName, server, db)
-relationship (repeatable)
-superiorLevel
-descriptionOfRelationship

7.4.3.3. Tag Types

For this schema a GRS-1 record will use the following tagTypes:

TagTypeUsage
1Elements from tagSet-M defined in Z39.50-1995, Appendix TAG, TAG.2.1. A target may include and an element from from tagSet-M at its discretion, and an origin may ignore it
2Elements from tagSet-G defined in Z39.50-1995, Appendix TAG, TAG.2.2. A target may include an element from tagSet-M at its discretion, and an origin may ignore it.
3Reserved for tags locally defined by a target. This profile does not define the usage of tagType 3.
4Elements from tagSet-digital collections defined in Z39.50 Profile for Access to Digital Collections

For the CIMI Profile, a subset of elements from tagSet-G and tagSet-M may be used to request specific elements from the database record. (See Appendix C.)

7.4.3.4. Element Set Names

The server creates a result set in response to a query. The result set is a set of pointers to records in a database. Element set names identify a group of elements in a database record for the server to return to the client.

The following element set names are defined for the CIMI Descriptive Record:

In addition, the CIMI Profile requires support for two element set names that the Collections Profile defines as mandatory

All element set names except 'target token' are literal strings. For 'target token' the element set name to be supplied is a target token that had been supplied previously by the server in a hit vector, where the hit vector corresponds to a portion of an object that satisfies the query (i.e., a chunk) that the client wishes to retrieve.

In the case when a server identifies an SGML object as containing satisfying chunks, the entire SGML object is the database record identified in the result set, and it is on that entire record that retrieval actions are requested (e.g., retrieving a single chunk, highlighting, etc.).

"Highlighting" is a function provided by Z39.50 for the server to identify satisfying portions of an database record. For the purposes of the CIMI Profile, the element set names that include "highlighting" are intended to request that the server insert chunkmarks or "hotspots" in SGML objects to indicate "chunks" of the object that satisfy a query. This approach is experimental within this demonstration project and can provide a basis for testing and refining an appropriate approach within Z39.50.

These element set names are subject to implementation experience to gain meaningful understanding of their feasibility and utility. Because of Version 2 limitations, these element set names are being used as a temporary expediency. This approach will be deprecated as the CIMI Profile evolves and Version 3 of the standard becomes the basis for the CIMI Profile.

7.4.3.4.1. briefCimi

Element set name briefCimi is a brief record of a CHIO Resource. It is defined as a Descriptive Record along with selected elements from the database record as appropriate for each of the different categories of digital objects in the CHIO resource. Several of the elements may be displayed to the user (e.g., briefDescription, categoryOfObject) and others may be used by the client for subsequent processing and navigation of the digital collection (e.g., target token). briefCimi assumes that the native format of the object is retrieved (e.g., SGML).

The element set name briefCimi is defined as follows:

ElementOccurrence in Retrieval Record
typeOfDescriptiveRecordAlways included
briefDescriptionIncluded if available [see note below]
collectionInfoWhen typeOfDescriptiveRecord is 1 (i.e., a Collection Descriptive Record) include the element collectionName only
objectInfo When typeOfDescriptiveRecord is 2 (i.e., an Object Descriptive Record) include typeOfObject, categoryOfObject, actualDO.

NOTE: The Project CHIO User Working Group determined a list of information that should be provided for each kind of CHIO resource in a brief record. Two approaches for providing that information are:

The element actualDO in briefCimi does not contain the digital object but provides a mechanism to return to the client metadata and variant information about the digital object (see below on GRS-1 in Section 7.4.3.5.1.).

In the case of SGML objects, briefCimi is intended to communicate a request to the server for GRS hits. GRS hit vectors are the mechanism by which the server conveys the number of chunks (i.e., the satisfying portions of the object) by the number of hits included. GRS hit vectors also provides the means (i.e., through a target token) for the client to subsequently retrieve a specific chunk.

7.4.3.4.2. briefCimiHtml

The element set name briefCimiHtml requests the same set of elements as briefCimi. However, for purposes of hit vector computation, this element set name requires that the server treat the Digital Object as an HTML document (as opposed to the element set name briefCimi, which requests the server treat the Digital Object as SGML for purposes of hit vector computation). These two distinct element set names, briefCimi and briefCimiHtml, are necessary because the server may assign different sets of target tokens when the document is treated as SGML as opposed to HTML; in particular, even if the logical chunks are identical, when the client subsequently retrieves a chunk via the target token, the token may indicate to the server whether to format the chunk as SGML or HTML.

7.4.3.4.3. {target token}

Unlike the other element set names defined by the CIMI Profile, {target token}provides for the retrieval of a single chunk of an SGML object. The server supplies a target token in a GRS hit vector, where the hit vector corresponds to the chunk that the client wishes to retrieve.

A target token is used by the client to request one chunk at a time.

7.4.3.4.4. fullCimiObject

Element set name fullCimiObject provides for the retrieval of the complete digital object. fullCimiObject is defined as:

ElementOccurrence in Retrieval Record
typeOfDescriptiveRecordAlways included
typeOfObjectAlways included
categoryOfObjectAlways included
actualDO Always included

The element actualDO in fullCimiObject includes the entire digital object. The server should include in the retrieved record the data available in the database record and which can be encoded in the requested record syntax.

7.4.3.4.5. fullCimiObjectHtml

The element set name fullCimiObjectHtml requests the same set of elements as fullCimiObject but the Digital Object is formatted in HTML.

7.4.3.4.6. Element Set Names to Request Highlighting

Two element set names, fullCimiObjectHighlighting and fullCimiObjectHtmlHighlighting, are used to explicitly request the server perform "highlighting" of the satisfying chunks of an SGML object.

The element set name fullCimiObjectHighlighting requests the same set of element as fullCimiObject but further specifies the SGML object be highlighted. The format of the returned actualDO is SGML.

The element set name fullCimiObjectHtmlHighlighting requests the same set of elements as fullCimiObjectHtml but further specifies the SGML object be highlighted. The format of the returned actualDO is HTML.

7.4.3.5. Record Syntaxes

The CIMI Profile requires the support of a single record syntax defined by Z39.50:

For interchange, GRS-1 records are to be treated as the complete and canonical representation of CHIO resources.

For purposes of interoperability, servers may support other record syntaxes including:


Retrieval records in other than GRS-1, however, may return only the digital object with descriptive information as prescribed in the Descriptive Record. Servers should return, if possible, the object in the syntax requested when that syntax is not GRS-1. When a server is unable to return an object in the requested record syntax, the server should return a diagnostic (e.g., Bib-1 Diagnostic #238 - Record not available in requested syntax; Bib-1 Diagnostic #239 - Record syntax not supported).

7.4.3.5.1. GRS-1

Usage of record syntax GRS-1 is defined as follows. In the GRS-1 main structure, the following parameters must be supported:

tagType
tagValue
tagOccurrence
content
metaData
appliedVariant

For ElementData:
octets Used when either {target token} or fullCimiObject, fullCimiObjectHtml, fullCimiObjectHighlighting, fullCimiObjectHtmlHighlighting, element set names are used
noDataRequestedUsed when briefCimi or briefCimiHtml element set names are used

For ElementMetadata:
hitsSupport mandatory for SGML objects when applicable

For HitVector:There is one HitVector for each satisfying chunk and each HitVector contains a sequence of the following:
satisfierOptional
offsetIntoElementOptional (IntUnit is character)
lengthOptional (IntUnit is character)
hitRankOptional
targetTokenOptional

Note: When a target token is provided in a hit vector corresponding to a specific chunk of a specific record, it may be used subsequently as an element set name, within a Present request that specifies the retrieval of that specific chunk of the record.

Note: If the server does not include a targetToken in a hitVector for a given chunk, the client cannot request the retrieval of that specific chunk and must instead retrieve the entire record to obtain the chunk.

For Variant:Variant is a sequence of the following:
variantSetIDAlways Variant-1
classAlways supplied
typeAlways supplied
valueAlways supplied

The GRS-1 element appliedVariant is used in retrieval records with all element set names.

The following defines how appliedVariant is used in briefCimi and briefCimiHtml to provide information about the size of the digital object and to provide information about the datatype of the digital objects.

Information about size of digital object:
Variant
triples
variantSetID={Z39-50-variantSet 1}(variant-1)
class=7(metadata returned)
type=2(size)
value=valueAndUnit(unit is byte)

Information about datatype of digital object:
Variant
triples
variantSetID={Z39-50-variantSet 1}(variant-1)
class=2(BodyPartType)
type=1 or 2(ianaType or Z39.50Type)
value=InternationalString

7.4.4. Access-control

Access-control is required for clients supporting the CIMI Profile. Clients must be prepared to receive Access Control messages. Servers are not required to support access-control. Access-control is intended to handle terms and conditions of use and other copyright information. The server may use access-control to send such information; alternatively, the information may be passed to the client as part of a retrieval record.

Use of access-control requires an access control format. Early implementation experience will provide a basis for developing an preliminary access control format. When stable the access control format will be registered and receive an OID.

The CIMI Profile unequivocally charges the client with responsibility to display to the user terms and conditions of use and other copyright information that is included in the AccessControlRequest.

7.5. Diagnostic Messages

The CIMI Profile requires support for Diagnostic Set Bib­1.

Based on implementation experience and the identification of actual or potential error conditions, new diagnostics may be proposed to be included in Diagnostic Set Bib-1. If the diagnostics are very specific to the CIMI Profile and its application environment, a new Diagnostic Set may be proposed.

7.6 Conformance

A statement of conformance will be developed based upon implementation experience through the Project CHIO testing of the CIMI Profile.


Appendix A:

CIMI-1 Attribute Set


This document specifies a Z39.50 attribute set for use in searching databases of museum information. It is primarily intended for searching databases consisting of or based on SGML documents, whether exhibition catalogues or object records, marked up with the CIMI Full Text Document Type Definition (DTD), Version 4 (Consortium for the Computer Interchange of Museum Information, 1996). It also allows searching of databases containing MARC-encoded bibliographic and authority records, image databases, and other resources.

Since this attribute set is intended for searching databases of museum information, the definitions of Use and Context attributes values provide a mapping to the CIMI Full Text DTD and to the Categories for the Description of Works of Art (CDWA) and International Committee for Documentation (CIDOC) Relational Data Model. Project CHIO also defined a set of CHIO access points, and a mapping from the attribute values to those access points are also included.

For guidance in mapping the attribute values, semantics for a selection of the attributes is provided.


Object Identifier

The attribute set defined in this specification is identified with the object identifier {Z39-50 attributeSet 8}. The object descriptor is "Attribute Set CIMI-1".


Attribute Types

This attribute set imports the following types from the Bib-1 attribute set defined in ANSI Z39.50: 1995:

Type Name Restrictions, Constraints
1 Use all values defined in Z39.50 Bib-1 are valid in this attribute set (but see below).
2 Position all values defined in Z39.50 Bib-1 (all values of this type may be ignored when searching SGML databases)
3 Relation all values defined in Z39.50 Bib-1
4 Structure all values defined in Z39.50 Bib-1
5 Truncation all values defined in Z39.50 Bib-1
6 Completeness all values defined in Z39.50 Bib-1 (all values of this type may be ignored when searching SGML databases)


The following additional types are defined by this attribute set:

Type Name Definition
100 Context a value identifying the context in which a use attribute occurs. Values of this attribute type map onto values of the CHIO attribute of the CONTEXT as specified in the CIMI DTD. The use of the context type in a MARC record environment is not defined.
101 Authority a value identifying the authoritative source from which a term is taken. Values of this attribute type map onto values of the AUTHORITY attributes of elements as specified in the CIMI DTD.

Values for Attributes of Type 1 (Use)

It is recommended that queries restrict their use of imported Bib-1 use attributes to the following:

Value Name Definition; Mapping to CIMI SGML DTD
1 personal name "a word/phrase giving a person's name"
CIMI = Topic element, Access-Point attribute value = "person"
CIMI = PersName element
CIDOC = People-Person Appellation Nam (NB CIDOC also has separate elements for first name, middle name, last name and generation name)
2 corporate name "the name of an organization"
CIMI = Topic element, Access-Point attribute value = "organization"
CIMI = OrgName element
CIMI = OrgTitle element
CIDOC = People-Group Appellation Nam
4 title "title of a work (book, journal, series) include. subtitles"
"a unique title for an object"
CIMI = Topic element, Access-Point value = "title-name"
CIMI = TitleName element
CIMI = Title element
CIDOC = Opus Appellation Nam
CIDOC = Object Appellation Nam
CIDOC = Event Appellation Nam
20 local number "an object's unique identity number"
CIMI = Topic element, Access-Point attribute value = "identity-number"
CIDOC = Object Identifier IDN
CDWA = Current Location - Repository Numbers
30 date "a single date in any format"
"a date range or an uncertain date"
CIMI = Topic element, Access-Point attribute value = "date-range"
CIMI = Date element
CIMI = DateRange element
CIDOC = Object Event Date Begin Tme
CIDOC = Object Event Date End Tme
CIDOC = Event Place Date Begin Tme
CIDOC = Event Place Date Begin Tme
CIDOC = Event & People & Role Dt Beg Tme
CIDOC = Event & People & Role Dt End Tme
CIDOC = Event Date Begin Tme
CIDOC = Event Date End Tme
etc.
CDWA = date subcategory (of various categories)
31 date of publication "the date of the text as given on the title page"
CIMI = DocDate element
52 number db a number assigned to a record in a database
32 date of acquisition CDWA = ownership/collecting history - dates
58 name geographic "a geopolitical place name"
CIMI = Topic element, Access-point attribute value = "place"
CIMI = PlaceName element
CIDOC = Place Geopolitical Nam
CDWA = Geographic Location subcategory
CDWA = Place subcategory
1003 author "author responsible for a bibliographic item"
CIMI = Author element
CIMI = DocAuthor element
CDWA = n/a (no category for author of a citation)
1016 any
1018 publisher "organization that publishes/distributes a book, etc."
CIMI = Publisher element
CIMI = Distributor element

This attribute set defines the following additional values for attributes of type 1 (Use). These attribute values are numbered in an unused number set of Bib-1:


Value Name Definition; mapping to CIMI SGML DTD
2000 award "prizes and other recognition"
CIMI = AwardName element
CIDOC= Award Nam
CIDOC = Award Txt
CDWA = context - historical/cultural - event name [where event type is award]
2001 bibliography words/phrases occurring within a bibliographic citation
CIMI = Bibl element
CIMI = BiblStruct element
CDWA = Citations subcategory of any CDWA category
2002 collection "a named collection [of objects]"
CIMI = Topic element, Access-Point attribute value = "collection"
CIMI = CollectionName element
CDWA = [no equivalent]
2003 concept "symbolic subjects, including iconography"
CIMI = ConceptName element
CIDOC = Concept Appellation Nam
CDWA = Subject matter - interpretation - indexing terms
2004 copyright restrictions "copyright or other intellectual property rights statement"
CIMI = Copyright element
CDWA = Copyright/restrictions - statement
2005 credit line "a credit line, e.g. source of acquisition"
CIMI = CreditLine element
CIDOC = Event Credit Line Txt
CDWA = Ownership/collecting history - credit line
2006 event "an historical event, e.g. exhibitions, wars"
CIMI = EventName element
CIDOC = Event Appellation Nam
CIDOC = Event Note Txt
CIDOC = Event Terms Txt
CDWA = Context - historical/cultural - event name
2007 inscription/mark "an inscription or mark, normally on an object"
CIMI = Mark element
CIDOC = Object Mark Transcription Txt
CIDOC = Object Mark Description Txt
CIDOC = Object Mark Interpret Nam
CIDOC = Object Mark Transcription Transln Txt
CDWA = Inscriptions/marks - transcription or description
2008 material "a material, e.g. of which an object is made"
CIMI = Topic element, Access-Point value = "material"
CIMI = MaterialName element
CIDOC = Material Appellation Nam
CDWA = Materials and techniques - materials - name
2009 nationality/culture/race "a person's nationality"
CIMI = Nationality element
CDWA = [creation - creator -] identity - nationality/culture/race
2010 object "the 'everyday' name of an object"
CIMI = Topic element, Access-Point value = "object.work"
CIMI = ObjectName element
CIDOC= Object Appellation Nam
CDWA = titles or names - text
2011 occupation "the work or business of a person"
CIMI = OccupationName element
CIDOC = Occupation Name
CDWA = [creation - creator -] identity - life roles
2012 process/technique "a means by which a task is accomplished"
CIMI = Topic element, Access-Point attribute value = "process.technique"
CIDOC = Method Nam
CIDOC = Method Description Txt
CDWA = materials and techniques - processes or techniques - name
2013 quote "a phrase or passage attributed to an external source"
CIMI = Quote element
CDWA = [no equivalent]
2014 role The name of a role played by a person or group of persons [in relation to an object] (CIDOC)
NB must be used with a context, or as part of a proximity query
CIMI = Role attribute of various elements
CIDOC = Role Nam
CDWA = role subcategory
2015 subject description CDWA = subject matter - description - indexing terms
2016 subject identification CDWA = subject matter - identification - indexing terms
2017 styles/movements "a distinctive form of artistic expression"
CIMI = StyleName
CIDOC = Style Appellation Nam
CDWA = Styles/periods/groups/movements - indexing terms
2018 technique CIMI = Topic element, Access-Point attribute value = "technique"
CIDOC = Method Nam
CIDOC = Method Description Txt
CDWA = Materials and techniques - processes or techniques - name
2019 type/classification "a category within a user-defined taxonomy"
CIMI = Topic element, Access-Point attribute value = "type.classification"
CIMI = Category element
CIDOC = Classification Term Nam
CIDOC = Classification Term Desc Txt
CDWA = object/work - type
CDWA = ojbect/work - components - type
CDWA = classification - term


Only one value of a use attribute may occur in an attribute list.

Values for Attributes of Type 100 (Context)

Value Name Definition; mapping to CIMI SGML DTD
1 award CIMI = Context element, CHIO attribute value = "award"
CIDOC = Object & People & Role & Award LDG
CIDOC = People & Role &Award LDG
CIDOC = People & Role & Event & Award LDG
CDWA = Context - historical/cultural (with event type equivalent to "award"
2 bibliography used to restrict titles, authors, etc. to those occurring within a bibliographic citation
CIMI = BiblStruct element
CIDOC = Object & Opus LDG
CDWA = Citations in any category
3 collection Context element, CHIO attribute value = "collection"
CIDOC = Object Related LDG
4 concept
5 context CDWA = Context
6 context archaeological CDWA = Context - archaeological
7 context architectural CDWA = Context - architectural
8 context historical CIMI = Context element, CHIO attribute value = "context-historical"
CDWA = Context - historical/cultural
9 copyright restrictions CDWA = Copyright/restrictions
10 creation CIMI = Context element, CHIO attribute value = "creation"
CDWA = Creation
11 creator CIMI = Context element, CHIO attribute value = "creator"
CDWA = Creation - Creator
12 credit line CDWA = Ownership/Collecting History - credit line
13 current location CIMI = Context element, CHIO attribute value = "current-location"
CDWA = Current location
14 date range ???
15 event CIMI = Context element, CHIO attribute value = "event"
CIMI = EventDesc element
CDWA = Context - historical/cultural - event name
16 identity CIMI = Context element, CHIO attribute value = "identity"
CDWA = Identity subcategory [of various categories]
17 mark CDWA = Inscriptions/Marks
18 material CDWA = Materials and Techniques - Materials
19 materials and techniques CIMI = Context element, CHIO attribute value = "materials-and-techniques"
CDWA = Materials and Techniques
20 nationality/race/culture CDWA = Nationality/Culture/Race subcategory [of various categories]
21 object/work CIMI = Context element, CHIO attribute value = "object"
CIMI = ObjectStruct element
CIMI = ObjectDesc element
CDWA = n/a (everything in CDWA is in the context of object/work)
22 occupation CDWA = Identity - Life Roles subcategory [of various categories]
23 organization
24 ownership CDWA = ownership/collecting history
25 person CDWA = various
26 place CDWA = various
27 process/technique CDWA = materials and techniques - processes or techniques
28 quote CIMI = Context element, CHIO attribute value = "Quote"
CIMI = quote element
CDWA = n/a? (quoted matter in remarks and citations subcategories?)
29 styles/movements CDWA = Styles/periods/groups/movements
30 subject CDWA = Subject matter
31 subject description CDWA = Subject matter - description
32 subject identification CIMI = Context element, CHIO attribute value = "object.work"
CDWA = Subject matter - identification
33 subject organization

Multiple values of a context attribute may occur in a single attribute list. The occurrence of multiple values is to be interpreted as follows: contexts after the first are nested within the contexts specified by preceding occurrences. For example, context 19 followed by context 10 followed by context 32 specifies that the query term is to be searched within a "creator" context within a "creation" context within an "object/work" context.

Values for Attributes of Type 101 (authority)

Value Name Definition
1 Non-Authoritative The client explicitly states that the term is not taken from any authoritative list
2 Local-to-server The term is known to the client to come from an authoritative source defined by the server
3 USMARC The term is a code taken from the US MARC manuals or associated documents, such as the set of coded values for countries languages, etc.
4 LCSH The term is from the Library of Congress Subject Heading
5 AAT The term is from the Art and Architecture Thesaurus
6 AAT_Date The term is from the Date and Geographic Name Guidelines
7 ACRL/RBMS_Binding The term is from the Binding Terms, Rare Books and Manuscripts
8 ACRL/RBMS_Genre The term is from the Genre Terms, Rare Books and Manuscripts
9 ACRL/RBMS_Paper The term is from the Paper Terms, Rare Books and Manuscripts
10 ACRL/RBMS_Printing The term is from the Printing and Publishing Evidence, Rare Books and Manuscripts
11 ACRL/RBMS_Type The term is from the Type Evidence, Rare Books and Manuscripts
12 Base_Merimee The term is from the Base Merimee Lexique
13 BGN The term is from the Board on Geographic Names
14 British_Archaeological The term is from the British Archaeological Thesaurus
15 Canadiana The term is from the Canadiana Authorities
16 Dictionarium_Museologicum The term is from the Dictionarium Museologicum
17 Garnier The term is from the Garnier's Thesaurus Iconographique
18 Geosaurus The term is from the Geosaurus, Geosystems' Thesaurus
19 Glass The term is from the Glass' Subject Index for the Visual Arts
20 ICOM_Costume The term is from the Vocabulary of Basic Terms for Cataloging Costume
21 ICONCLASS The term is from the ICONCLASS
22 Jewish_Art The term is from the Index of Jewish Art
23 ISO_Language The term is from ISO 639: Language Codes
24 ISO_Documentation The term is from ISO 5127-1: Documentation and Information
25 ISO_Iconic The term is from ISO 5127-3: Iconic Documents
26 ISO_AV The term is from ISO 5127-11: Audio-visual Documents
27 ISO_Date/Time The term is from ISO 8601: Dates & Times
28 LC_Descriptive_Graphic The term is from the LC Descriptive Terms for Graphic Materials
29 LC_Name The term is from the LC Name Authorities
30 LC_Thesaurus_Graphic The term is from the LC Thesaurus for Graphic Materials
31 Moving_Image_Materials The term is from the Moving Image Materials: Genre Terms
32 Nomenclature The term is from the Revised Nomenclature for Museum Cataloging
33 Reynies The term is from the Reynies' Le Mobilier Domestique
34 TGN The term is from the Thesaurus of Geographic Names
35 Tozzer The term is from the Tozzer Index to Anthropological Subject Headings
36 ULAN The term is from the Union List of Artist Names
37 Villard The term is from the Epoque et style des antiquites classiques
38 Yale_British_Artists The term is from the British Artists Authority


Only 1 value of an authority attribute may occur in an attribute list.

The server is to interpret the lack of an authority value in a query as the client "not saying" anything about the term. When such a case occurs, it is the server's choice in processing the term.

The list of values for the Authority attribute will be expanded as necessary based on implementation experience.

Semantics for Selected Attribute Values

The following provides semantics for a selection of Use and Context attribute values in CIMI-1. Semantics for Use attributes incorporated in CIMI-1 from Bib-1 can be found in Attribute Set BIB-1 (Z39.50-1995): Semantics.


Use or Context Attribute Value Use Attribute Name Semantics
USE 2000 Award An award or prize (e.g., Prix de Rome, first prize). [KJ]
USE 2001 Bibliography Citations for written works, such as books, journals, and exhibition catalogs. [KJ]
USE 2002 Collection
USE 2003 Concept An abstraction presented in the interpretation of the subject matter of a work of art (for example, death, innocence, salvation.) [KJ]
USE 2004 Copyright Restriction Identification of the individual or group that holds the rights to use, exhibit, or reproduce a work of art, along with an indication of any existing restrictions on its reproduction, exhibition, or use. [KJ]
USE 2005 Credit Line A formal public statement about the ownership, transfer of ownership, acquisition, source, or sponsorship of the acquisition of a work, suitable for use in a display label or publication (e.g., Samuel H. Kress Collection, Benjamin Bequest). [KJ]
USE 2006 Event An historical event or situation (e.g., French Revolution, Coronation of Richard II of England, World's Columbian Exposition). Any past occurrence, whether or not it has a specific name. [RL]
USE 2007 Inscription/Mark A description of distinguishing or identifying physical markings, lettering, annotations, texts, or labels that are a part of a work of art or are affixed, applied, stamped, written, inscribed, or attached to the work, excluding any mark or text inherent in the materials. [KJ]
USE 2008 Material An identification of the materials used to create the work of art, along with an indication of where they were employed. [RL]
USE 2009 Nationality/Culture/Race The national, cultural, or ethnic origins of a person (e.g., English, Sienese, Berber, African American). [KJ]
USE 2010 Object An art work, or any inanimate object. [RL]
USE 2011 Occupation A major professional role played by an individual in his/her lifetime, or the major roles that define the activities or purpose of an organization (e.g., architect, painter, sculptor, goldsmith, architectural firm). [KJ]
USE 2012 Process/Technique The means, method, process, or technique by which a material was used in the creation of a work. [RL]
USE 2013 Quote A piece of text taken from an external written or spoken source. Quotes will often have an associated person or bibliography. [KJ]
USE 2014 Role
USE 2015 Styles/Movements A style, historical period, school, or art movement (e.g., Baroque, Fauve, Brut, Nayarit). [KJ]
USE 2016 Subject Description A description of a work of art in terms of the generic elements of the image or images depicted in, on, or by it (e.g., woman sitting in an enclosed garden holding a baby, with a landscape in the distance). [KJ]
USE 2018 Techniques
USE 2019 Type/Classification A categorization, either formal or informal, of an entity. [RL]
CONTEXT 5 Context Political, social, economic, or religious events or movements associated with a work of art at its creation and over time. [KJ]
CONTEXT 6 Context Archaeological The circumstances in which a work of art was excavated or discovered (e.g., "The burned house and it's elaborate ironwork were covered by dense bramble, partially protecting its contents from the elements.") [KJ]
CONTEXT 7 Context Architectural The relationship between a work of art and a particular environment, structure, or open space (e.g., The statue once stood at the west side of the front door .) [KJ]
CONTEXT 8 Context Historical Political, social, economic or religious events or circumstances associated with the work of art over time (e.g., A popular activity at the bar involved creating tall tales relating the fishing expeditions where these wooden fish were supposedly "caught.) [KJ]
CONTEXT 10 Creation The creation, design, execution, or production of a work of art and its components, including all those responsible for the creation of the work, the dates of that activity, and where the creation took place. [KJ]
CONTEXT 11 Creator
CONTEXT 13 Current Location The place where a work of art is currently housed, including its geographical location. [KJ]
CONTEXT 14 Date Range A time span, normally represented by a start date and an end date (e.g., from 1931 to 1932). Date ranges can also be represented (less exactly) by descriptive terms (e.g., Dark Ages). Either, or both, ends of the range may be estimated or unknown. [KJ]
CONTEXT 16 Identity
CONTEXT 19 Materials and Techniques The substances or materials used in the creation of a work of art, as well as any production or manufacturing techniques, processes, or methods incorporated in its fabrication. This includes a description of both the materials used to create the work and the way in which they were put together. (E.g. Balsa wood wrapped with aluminum foil.) [KJ]
CONTEXT 23 Organization A named group associated together for a particular purpose (e.g., Corporation of London, National Gallery of Art, Burgess Hill Sinfonietta). [KJ]
CONTEXT 24 Ownership The history of the ownership of an object from its creation to the present. This includes the names of the owners, the means by which it passed from one owner to the next, an identification of any public sales involving the work, the names of any agents who aided the transfer of ownership, and the names of any dealers who handled the work or included it in their inventories. Ownership also includes information on the loss, theft, destruction or other ways in which a work may have vanished from public view. Also known as provenance. [KJ]
CONTEXT 25 Person An individual. This includes anonymous individuals: useful information can still be recorded even if a person's name is unknown. (Use styles-movements for informal groups of people who worked together, and organization for formal entities.) [KJ]
CONTEXT 26 Place A geographical or geopolitical location. This includes imprecisely defined places (e.g., the Hudson River Valley, the Great Lakes region). [KJ]
CONTEXT 33 Subject Organization


[Note: The semantics for these attributes are taken from the glossary developed for Project CHIO by Kody Janney and the Full Text DTD by Richard Light.]


Appendix B

Values for CIMI Profile's PrivateProximityUnit

This appendix enumerates the values for a private set of proximityUnitCode. The values of the PrivateProximityUnit are the values of the Context Attribute.

ValuePrivateProximityUnit
1award
2bibliography
3collection
4concept
5context
6context archaeological
7context architectural
8context historical
9copyright restrictions
10creation
11creator
12credit line
13current location
14date range
15event
16identity
17mark
18material
19materials and techniques
20nationality/race/culture
21object/work
22occupation
23organization
24ownership
25person
26place
27process/technique
28quote
29styles/movements
30subject
31subject description
32subject identification
33subject organization


Appendix C

Information to Include in Brief Record

The Project CHIO User Requirements Working Group has identified elements from the CHIO resources to comprise a brief record for display to the user. These elements may be included in the retrieval record element briefDescription for each of the different CHIO resources or can be identified using tags from tagSet-G.


Category of CHIO Resource
Elements in Brief Display
tagSet-G Mapping
element title [tag]
artist recordPerson name name [7]
Life dates* dateRange
authority recordName of authority source * authoritativeScheme
Preferred term* preferredTerm
a-v file (text to accompany an a-v file) Title of worktitle [1]
Subject* subject
Institutionorganization [10]
Date of workdate [8]
** [file size]
cataloging recordTitle title [1]
Authorauthor [2]
Datedate [8]
exhibition catalogueTitle of catalog title [1]
Institution responsible for catalog organization [10]
Date of catalog date [8]
** [file size]
** [number of hits]
***image
object recordTitle of object title [1]
Creator of objectname [7]
Date of creation of object date [8]
Classification type of object description [17]
Institutionorganization [10]
scope of collection record
thesaurus (hierarchical)Term plus immediate parent [?]
Preferred term* preferredTerm
Source (thesaurus name?) * authoritativeScheme
wall textTitle of wall text title [1]
Institutionorganization 10

* For the information units not able to be mapped to tagSet-G, a suggested new tagSet-G element is listed.

** The two units of information, "file size" and "number of hits" are accounted for in Z39.50 variant information and GRS metadata and are not separate elements that need to be distinguished here.

*** Images and audio-video files will likely be associated with text as found, for example, in an object record, an exhibition catalogue, wall text, etc.


Appendix D

Non-Protocol Agreements & Understandings on

Client and Server Behavior/Functionality

This section is non-normative and is intended to describe agreements among the developers of the CIMI Profile. The Profile is intended to support a demonstration project, and the details included in this appendix provide potential implementors with some background on decisions and guidance on interpreting the specifications.

Because the Profile supports a demonstration project, some of the specifications may become better understood through implementation experience. In fact, the implementation experience will provide input for revising the specifications.

I. Version 2 Implementations

The CIMI Profile is addressing Z39.50-1995 Version 2 in the initial draft of the Profile. Implementors are beginning to deploy Version 3 clients and servers, but for the short-term there will likely be more Version 2 implementations that may be able to support the Profile. Interoperability is also an issue with the installed base of Version 2 implementations.

The choice of Version 2 has implications for the specifications in the Profile. Such implications can be seen in at least two areas: attribute sets and element set names. Since Version 2 does not allow the use of more than one attribute set in a query, clients may need to conduct multiple searches to get a desired result set. In Version 3, the use of multiple attribute sets may provide for more efficient server behavior. Similarly in the area of retrieval and the use of element set names -- Version 2 implementations do not support ESpec. Therefore the profile has specified three element set names to accomplish retrieval that may be done more efficiently in Version 3.

Choosing Version 2 for the initial implementations does not preclude moving to a profile of Version 3 in the future. In fact, the demonstration project will allow the evaluation of the Version 2 specifications for supporting the functional requirements of Project CHIO.

II. Attribute Sets

The newly defined CIMI-1 attribute set is intended for searching databases of SGML objects that are marked up in the CIMI Full Text DTD Version 4. The DTD provides semantics for the access points used in the CIMI-1 attribute set. If appropriate, these semantics can be added to the CIMI-1 attribute set definition.

Implementation experience will help determine the degree to which the attribute set is adequate for expressing queries. Some validation of the attribute set will be needed, such as mapping queries to more than SGML objects, and mapping queries to Bib-1 attributes only since we anticipate that Bib-1 clients will be searching the CHIO Collection.

An implication of profiling Version 2 is that in the initial implementations it will not be possible to handle a "user-defined" context within the current CIMI-1 attribute set. Supporting such searching will require Version 3 which allows string searching of attributes.

III. Proximity Operation

While one can nest Context Attributes when searching, such nesting does not always express the semantics of a search. To accommodate a search that is looking, for example, a Creator and a Creation in the context of an Award when those elements are not contained within the same nested context in the SGML object, the Profile suggests the use of a proximity operation. The Profile defines values for a private ProximityUnitCode.

IV. Retrieving Large Objects

Z39.50 segmentation of records is not required in the CIMI Profile. Instead, clients and servers should agree to specify a reasonable record size when negotiating minimum values for maximum record size.

V. Retrieving SGML Objects

Knowledge about SGML (e.g., the DTD, associated stylesheets, and navigator files) is assumed to be localized to the extent possible. Clients should look locally for the appropriate DTD, and the server is not responsible for getting the DTD. The client is assumed to have a priori knowledge about the associated stylesheets and navigator files. For acquiring the appropriate DTD, clients should acquire it in the manner they would when operating in the SGML world. Z39.50 retrieval and management of stylesheets and navigator files is for further study.

A current limitation relates to the extent of communication between a Z39.50 client and a SGML viewer to support navigation. To address this limitation and yet accommodate the functional requirements of Project CHIO, the Profile specifies the use of "hits" and "hotspots."

A server is assumed to provide some chunking of the SGML object, where a chunk is the smallest element of the SGML object that provides the context of whatever satisfies a query. Since a single SGML object may have more than one satisfying chunk, each satisfying chunk is represented to the client as a "HitVector," which is GRS-1 metadata about the actualObject element in a Descriptive Record. Thus, the number of "hits" can be reported to the client for display to the user in a brief CIMI record (i.e., briefCimi or briefCimiHtml). This allows the user to choose one or more SGML objects that may be most relevant to their search, as well as retrieve a single chunk

The HitVector contains an element targetToken that can be used by the client to retrieve a single chunk. The server returns the single chunk, along with enough of the SGML object to make it legitimate and processible SGML for the SGML viewer. The client retrieves the chunk and can pass it to the SGML viewer.

Another feature in the CIMI Profile that supports navigation within an SGML object is the insertion of "hotspots." For example, if a single SGML object has several satisfying chunks, the user could retrieve the entire object to view in an SGML viewer. Through the use of two element set names, fullCimiObjectHighlighting and fullCimiObjectHtmlHighlighting, the client requests the records be returned with the chunks highlighted. The server inserts identifying information (i.e. "hotspots") into the SGML objects to designate the beginning and end of a chunk (i.e., it inserts hotspots). These hotspots are recognized in the DTD as legitimate. When the entire object is returned, the client passes the object to the SGML viewer and the user can then navigate to each "hotspot" as well as navigating to a section before or a section after the satisfying chunk, or elsewhere in the object.

VI. Element Set Names

The CIMI Profile defines a number of element set names. One functional requirement stated the need to display brief information about whatever is in the result set -- no matter whether it was a SGML object, MARC record, image, etc. The Descriptive Record defined in the Profile provides for this. The briefCimi and briefCimiHtml element set name includes the element actualObject and is a mechanism to ask for metadata about the element. This element set also provides the client with enough information to make a decision whether or not it can actually process and display the CHIO resource to the user (e.g., a bibliographic client that supports MARC may be able to process a MARC cataloging record but not an SGML object).

{target token} is intended to be used as a mechanism to retrieve a single chunk of the SGML object.

fullCimiObjectis intended to return the entire CHIO resource in a requested record syntax. The actualObject element in the contains the full object and will be returned to the client through this element set name. Not all CHIO resources may be available in a requested record syntax

VII. Record Syntaxes

GRS-1 is the required record syntax specified in the Profile. However, servers are afforded the latitude to attempt to provide CHIO resource in whatever record syntax was requested. Returning any non-GRS records may require server functionality not specified in the Profile. In principle, a server should return, if possible, the object in the syntax requested when that syntax is not GRS.

VIII. Access Control Format

As an initial implementor of the CIMI Profile, OCLC will develop a prototype Access Control Format, and will register it on the OCLC tree. It is anticipated that such an Access Control Format will become stabilized through the demonstration project, be specified in the Profile, and formally registered.


Appendix E

Examples of CHIO Collection Resources

[to come]