FRBR and metadata decisions for repository users

At the recent DC-2007 conference I was introduced to the potential relevance of FRBR to the future of repositories.

Till now I have always tried to avoid anything that looked too heavily into FRBR discussion because it seemed too abstract and complex to mean anything to the here and now job on my desk. An article about a Dublin Core Application Profile for Scholarly Works (based on FRBR and the Dublin Core Abstract Model) also looked like it should be interesting but again I found myself wanting to relate it to my work but not seeing how to — yet.

But after DC-2007 I think I am beginning to see the light. It’s about positioning ourselves for the future. And the future is being introduced bit by bit now with work on RDA, and especially with the working together by RDA and DCMI now. It’s also about positioning repositories for the semantic web.

Think of a web user looking for a “work”, say Homer’s Iliad. She wants to know what is out there and then decide what “manifestation” is best for her (oral reading in the form of an mp3 file or illustrated text), and also what “expression” of the work is best (a simplified abridged version or the original Greek reading) — and then if she can copy the “item” or “copy” that might be available. If the richness of web searching enables such a user navigate through these sorts of web-surfing thoughts then we need to ask ourselves if we are presenting our data in the most useful way for that user.

What this means for repositories

Not a great deal just immediately. But it’s a good idea to keep the FRBR model in mind to make it easier to understand where repositories potentially fit in with the semantic web. By understanding this we can make better informed decisions about metadata in the meantime.

It has long been useful for cataloguers to describe a book added to a library collection by a single record containing such properties as:

• Title (MARC 245)
• Subject (MARC 650)
• Abstract (MARC 520)
• Author (MARC 100)
• Language (MARC 041)
• Bibliographic level (e.g. monograph or part of a monograph, serial or part of a serial, etc.) (MARC Leader 07)
• Genre/Form (MARC 655)
• Publisher (MARC 260)
• Physical description (MARC 300)
• Terms governing use and reproduction (MARC 540)

What the FRBR model does is to assign the above list of terms into their distinct conceptual boxes or entities. The idea is that this will make it easier for computers to manipulate the data so that the info in the web can be structured to give users heaps more power to interrogate the web and find things and even more things related to things.

FRBR rewrites the above list of MARC terms into 4 entities:

The Work: “a distinct intellectual or artistic creation. . . . . A work is an abstract entity; there is no single material object one can point to as the work. . . . When we speak of Homer’s Iliad as a work, our point of reference is not a particular recitation or text of the work, but the intellectual creation that lies behind all the various expressions of the work.” (IFLA 1998)

The EPrints Application Profile lists the following as among the descriptive properties of a Scholarly Work:

• Subject
• Abstract
• Creator

Compare MARC’s

• Subject (MARC 650)
• Abstract (MARC 520)
• Author (MARC 100)

The Expression: “the intellectual or artistic realization of a work in the form of alpha-numeric, musical, or choreographic notation, sound, image, object, movement, etc., or any combination of such forms. . . . An expression is the specific intellectual or artistic form that a work takes each time it is “realized.” (IFLA 1998)

The EPrints Application Profile lists the following as among the descriptive properties of an Expression:

• Status (e.g. peer reviewed)
• Type (e.g. journal article)
• Language

Compare MARC’s

• Language (MARC 041)
• Bibliographic level (e.g. monograph or part of a monograph, serial or part of a serial, etc.) (MARC Leader 07)
• Genre/Form (MARC 655)
  – But note neither the bibliographic level in the MARC leader nor the 655 field strictly matches conceptually the FRBR concept of Expression.

(The EPrints Application Profile also recommends best practice to provide a value URI for a class from the Eprints Type Vocabulary Encoding Scheme.)

The Manifestation: “the physical embodiment of an expression of a work. . . . encompasses a wide range of materials, including manuscripts, books, periodicals, maps, posters, sound recordings, films, video recordings, CD-ROMs, multimedia kits, etc..” (IFLA 1998)

The EPrints Application Profile lists the following as among the descriptive properties of a Manifestation:

• Publisher
• Format (e.g. IANA MIME types)

Compare MARC’s

• Publisher (MARC 260)
• Physical description (MARC 300)

The Item: “a single exemplar of a manifestation. The entity defined as item is a concrete entity. It is in many instances a single physical object (e.g., a copy of a one-volume monograph, a single audio cassette, etc.).” (IFLA 1998)

The EPrints Application Profile lists the following as among the descriptive properties of an Copy or Item:

• Access rights
• Date available

Compare MARC’s

• Terms governing use and reproduction (MARC 540)

So the above MARC data would be broken into something like this:

The Work:

• Title (MARC 245 or other)
• Subject (MARC 650)
• Abstract (MARC 520)
• Author (MARC 100)

The Expression:

• Title (MARC 245 or other)
• Language (MARC 041)
• Bibliographic level (e.g. monograph or part of a monograph, serial or part of a serial, etc.) (MARC Leader 07)
• Genre/Form (MARC 655)

The Manifestation:

• Publisher (MARC 260)
• Physical description (MARC 300)

The Copy/Item:

• Terms governing use and reproduction (MARC 540)

There is some conceptual merging of FRBR Expressions and Manifestations in some of the MARC fields. But Dublin Core modeling can sort some of that overlap out.

What’s the point?

The real value of the above modeling and breakdown of concepts is going to be with the RDF applications for the semantic web. But even apart from the semantic web, with repositories collecting virtual works and so many different versions of works (pre-prints, post-prints, powerpoint version, paper version, . . . . ) keeping this FRBR model in mind can be useful just for helping us clarify the best way to use the tools we are working with.

Web users looking for a “work”, say Homer’s Iliad, are not looking for a specific book. They want to know what is out there and then decide what “manifestation” is best for them (oral reading in the form of an mp3 file or a retelling for children), and also what “expression” is best (a simplified abridged version or the original Greek reading) — and then if they can copy the item that might be available.

Keeping user needs and interests in mind through models like this can help us understand what we ought to be doing when we construct metadata models and schema.