Data models for metadata:

Document details:

Summary:

Draft report describing two entity-relationship models for metadata: the model produced by the IFLA Study Group on Functional Requirements for Bibliographic Records in 1996 and an integrated model for descriptive and rights metadata developed by Godfrey Rust. A preliminary attempt is made to relate these to work being carried out with relation to Dublin Core and the Resource Description Framework.

Contents

1. Introduction
2. Selected data models
2.1. IFLA Functional Requirements for Bibliographic Records
2.2. An Integrated Model for Descriptive and Rights Metadata
2.3. DC Relations Working Group
3. Comparison of data models
4. Resource Description Framework (RDF)
5. Conclusions
6. References
7. Appendix: IFLA Entity Attributes

1. Introduction

The Dublin Core (DC) initiative is an international effort to define a core set of metadata elements for resource discovery. In 1997 the DC community began to interact with the evolving Resource Description Framework (RDF) being developed under the auspices of the World Wide Web Consortium (W3C). Stu Weibel and Juha Hakala (1998) state that DC's work with RDF "has clarified the importance of a coherent underlying data model for Dublin Core metadata". Accordingly, a working group with responsibility for formulating a Data Model emerged from the DC-5 Workshop (Helsinki) held in November 1997.

The work of this group is ongoing but its discussions have prompted a necessary re-evaluation of the data model that currently underlies DC. Data models being currently developed for bibliographic (or bibliographic-type) objects tend to use object modelling (OM) and entity relationship (ER) modelling - which identifies (or models) the relationships between independent objects (Greenberg 1997, p. 115).

Michael Heaney, for example, has reanalysed the nature of library cataloguing with reference to object-oriented modelling and has suggested major revisions of MARC that involve a "fundamental shift away from the AACR2 philosophy of description of, plus access to, physical items" (Heaney 1995, p. 152). Heaney's own object-oriented cataloguing model identifies a three-tiered structure:

• Work (text) - strings of sentences that have titles and are in a language
• Publication - a manifestation of a particular text
• Copy - "each publication is manifested in a certain number of physical copies" (p. 141)

In addition, another class of object - "Agents" - may be associated with works, publications or copies. Examples of these would be authors, editors, translators (of works), publishers (of publications) and owning agents (of copies).

Heaney's object modelling approach reflects a large-scale rethinking of the conceptual foundations of descriptive cataloguing within the library community (e.g. Tillett 1991; 1996). This rethinking has manifested itself in a number of initiatives that propose the development of new cataloguing rules based on object (or entity relationship) modelling. The most prominent of these initiatives are an report of an IFLA Study Group on Functional Requirements for Bibliographic Records (IFLA Study Group 1996; 1998) and the International Conference on the Principles and Future Development of AACR held in Toronto in October 1997 (Joint Steering Committee 1997).

The purpose of this paper is to look at some object or entity-relationship models being currently developed to represent bibliographic relationships and to relate this to work being done on Dublin Core and RDF. The two models mostly discussed here are the report of the IFLA Study Group (1996; 1998) and a model developed descriptive and rights metadata (Rust 1998). A section on the IMS Meta-Data Structure developed by the Instructional Management Systems (IMS) Project will be added in the future.

2. Selected data models

2.1 IFLA Functional Requirements for Bibliographic Records

Following the 1990 Stockholm Seminar on Bibliographic Records, sponsored by the IFLA Universal Bibliographic Control and International MARC Programme (UBCIM) and the IFLA Division of Bibliographic Control, a project was proposed to study the functional requirements for bibliographic records. The resulting study made no a priori assumptions about the content or structure of records but instead used an entity analysis technique for identifying the entries that are the "key objects of interest to users of bibliographic records" (IFLA Study Group 1996, p. 2).

For the purposes of the study, the following generic tasks were identified:

• Finding
• Identifying
• Selecting
• Obtaining

ER methodology used because "it provides a structured approach to the analysis of data requirements", in this case the key objects of interest to users of bibliographic databases. The key objects identified were in three groups:

1. The products of intellectual or artistic endeavour:

• Works - a distinct intellectual or artistic creation
• Expressions - the intellectual or artistic realisation of a "work"
• Manifestations - the physical embodiment of an "expression" of a "work"
• Items - a single exemplar of a "manifestation"

2. The entities responsible for the intellectual or artistic content of such products:

• Persons - an individual
• Corporate bodies - an organisation or group of individuals and/or organisations acting as a unit

The entities that form the subject of intellectual or artistic endeavour

• Concepts - an abstract notion or idea
• Objects - a material thing
• Events - an action or occurrence
• Places - a location

This third group also contains all the entities belonging in the first two groups.

Each of these entities will have relevant attributes, e.g. for "work":

• Title
• Form
• Date
• Other distinguishing characteristic
• Intended audience
• Context
• Medium of performance (musical work)
• Numeric designation (musical work)
• Key (musical work)
• Co-ordinates (cartographic work)
• Equinox (cartographic work)

These attributes are mapped to the International Standard Bibliographic Description (ISBD) and the Guidelines for Authority and Reference Entries (GARE) data elements and a list of coded data fields from the UNIMARC format (IFLA Study Group 1996, pp. 96-112).

The IFLA functional requirements contains a high level entity-relationship diagram

[Add diagram from report].

IFLA Relationships:

In short, the IFLA document identifies six defined relations between entities in the high-level model:

• created by
• embodied in
• exemplified by
• has as subject
• realised by
• realised through

More information on the model can be found in the Functional Requirements for Bibliographic Records document (IFLA Study Group 1996). The final version of this report has recently been published (IFLA Study Group 1998).

2.2 An Integrated Model for Descriptive and Rights Metadata

In a paper in D-Lib Magazine, Godfrey Rust (1998) proposed an integrated model for descriptive and rights metadata. This paper is primarily concerned with metadata related activity by the rights-owning communities. Rust comments that creators and rights owners will become the principal source of core metadata in the Web environment and that "metadata will be generated simultaneously and at source to meet the requirements of discovery, access, protection and reward".

The CIS (Common Information System) plan, on which Rust bases parts of his model, coined the generic term "creation" to denote a "product of human imagination and/or endeavour by one or more parties in which rights may exist". Creations come in four main structural types:

• Package - the physical medium (e.g. book, CD, photograph)
• Object - the digital medium (e.g. text, picture, audio)
• Performance - the spatio-temporal aspect (e.g. a live performance of work, a broadcast of recording)
• Work - the abstract aspect (e.g. a musical composition, a literary work)

Relationships are defined with relation to the concept of "nesting". Rust argues that "nesting" is the norm for most works, e.g. sound recordings manifest at least one corresponding composition, and printed editions manifest at least one corresponding "underlying" literary work.

Rust notes that Creations are made by people who may then assign Rights in them to other people by Agreements. Rights metadata may be expressed in three distinct categories:

• Requests - the specified use to which a party wishes to put a creation
• Offers - the terms under which a creation may be put to specified uses
• Agreements - the terms of a completed transactions

[Add diagrams on relationships]

2.3 DC Relations Working Group

The Relations working group was constituted following the Helsinki Metadata Workshop (DC-5) of October, 1997 (Weibel and Hakala 1998) and was chaired by David Bearman of Archives and Museum Informatics (DC 1998).

The nature of the relationship between the current resource and the resource identified by Relation.Identifiers is described in "Relation.Type".

The values proposed for the DC Relation.Type sub-element are:

a) Part/Whole relations:

• IsPartOf
• HasPart

b) Version relations

• IsVersionOf
• HasVersion

c) Format transformation relations

• IsFormatOf
• HasFormat

d) Reference relations

• References
• IsReferencedBy

e) Creative relations

• IsBasedOn
• IsBasisFor

f) Dependency relations

• Requires
• IsRequiredBy

3. Comparison of data models

[Text TBA]

4. Resource Description Framework (RDF)

The Resource Description Framework is a metadata initiative which is being developed under the auspices of the World Wide Web Consortium (W3C). RDF is designed to enable the encoding, exchange and reuse of structured metadata (Miller 1998). RDF provides a model for describing resources. It then proposes syntax (in Extensible Markup Language - XML) based on this data model and a specification for schema.

W3C's interest in metadata began with work on developing an infrastructure for Web page ratings known as PICS - the Platform for Internet Content Selection (Resnick and Miller 1996). A PICS-NG (Next Generation) working group was then formed to address more general resource description issues, and this in turn metamorphosed into the RDF working group.

[TBC]

5. Conclusions

[TBA - will discuss how something like IFLA data model could be represented in RDF and how DC maps on to this (?)].

6. References

Brickley, D., Guha, R.V. and Layman, A., eds., 1998, Resource Description Framework (RDF) schema specification. W3C Working Draft. <URL:http://www.w3.org/TR/WD-rdf-schema/>

Dublin Core, 1998, Relations Working Group. <URL:http://purl.oclc.org/metadata/dublin_core/wrelationdraft.html>

Greenberg, J., 1997, Reference structures: stagnation, progress and future challenges. Information Technology and Libraries, vol. 16, no. 3, pp. 108-119.

Heaney, M., 1995, Object-oriented cataloguing. Information Technology and Libraries, vol. 14, no. 3, pp. 135-153.

IFLA Study Group on the Functional Requirements for Bibliographic Records, 1996, Functional requirements for bibliographic records: draft report for world-wide review. Frankfurt am Main: IFLA Universal Bibliographic Control and International MARC Programme.

IFLA Study Group on the Functional Requirements for Bibliographic Records, 1998, Functional requirements for bibliographic records: final report. IFLA UBCIM Publications, New Series, Vol. 19. München: K.G. Saur.

Joint Steering Committee for Revision of Anglo-American Cataloguing Rules, 1997, International Conference on the Principles and Future Development of AACR. Conference papers at: <URL:http://www.nlc-bnc.ca/jsc/confpap.htm>

Lassila, O. and Swick, R., eds., 1998, Resource Description Framework (RDF) model and syntax specification. W3C Working Draft. <URL:http://www.w3.org/TR/WD-rdf-syntax/>

Miller, E., 1998, An introduction to the Resource Description Framework. D-Lib Magazine, May. <URL:http://www.dlib.org/dlib/may98/miller/05miller.html>

Powell, A., 1998, Metadata for the Web: RDF and the Dublin Core. 20th UK Online User Group (UKOLUG) Conference, Manchester Conference Centre, 15 July. <URL:http://www.ukoln.ac.uk/metadata/presentations/ukolug98/paper/intro.htm>

Resnick, P. and Miller, J., 1996, PICS: Internet access controls without censorship. Communications of the ACM, vol. 39, no. 10, pp. 87-93. <URL:http://www.w3.org/PICS/iacwcv2.htm>

Rust, G., 1998, Metadata: the right approach. An integrated model for descriptive and rights metadata in e-commerce. D-Lib Magazine, July/August. <URL:http://www.dlib.org/dlib/july98/rust/07rust.html>

Tillett, B.B., 1991, A Taxonomy of Bibliographic Relationships, Library Resources & Technical Services, 35 (2), pp. 150-158.

Tillett, B.B., 1996, Cataloguing rules and conceptual models. OCLC Distinguished Seminar Series, January 9. <URL:http://www.oclc.org:5046/~emiller/misc/tillett.html>

Weibel, S. and Hakala, J., 1998, DC-5: The Helsinki Metadata Workshop: a report on the workshop and subsequent developments. D-Lib Magazine, February. <URL:http://www.dlib.org/dlib/february98/02weibel.html>

Weibel, S., Kunze, J., Lagoze, C. and Wolf, M., 1998, Dublin Core metadata for resource discovery. Internet-Draft, 5 August. <ftp://ftp.ietf.org/internet-drafts/draft-kunze-dc-03.txt>

7. Appendix: IFLA Entity Attributes

8. Acknowledgements

UKOLN is funded by the British Library Research and Innovation Centre (BLRIC), the Joint Information Systems Committee (JISC) of the higher education funding councils, as well as by project funding from the eLib Programme and the European Union. UKOLN also receives support from the University of Bath, where it is based.

Maintained by: Michael Day of UKOLN The UK Office for Library and Information Networking, University of Bath.
First published in this form: 23-Oct-1998
Last updated: 23-Oct-1998 (v. 0.2)