2012-01-05. Frozen archive - links may not resolve - see directory of files at MoinMoin wiki archive

> CdapReview

About this draft review

This document is a 'trial run' for a full review of the CD Application Profile by the Usage Board. The draft is work-in-progress; it is intended for internal use of the Usage Board and should not be cited or quoted outside that context. The purpose of this document is to help the Usage Board visualize the end result of the review process -- its outline and (roughly) its length and granularity of detail.

Usage Board Review of the Collection Description Application Profile -- Draft

The DCMI Usage Board undertook an assessment of the Collection Description Application Profile (CDAP) at a meeting in Manzanillo Mexico [1] on Sunday, 1 October 2006. Members present were Tom Baker (chair), Diane Hillmann, Andy Powell, Akira Miyazawa, Stuart Sutton, Joe Tennis, Andrew Wilson (designated shepherd of the Collection Description proposal), and Makx Dekkers (ex officio). This note holds an interim response to the DCMI Collection Description Working Group, short of being a formal review of a submitted application profile. This note includes comments on the Collection Description application profile as a whole. See [1] for links to the actual documents reviewed.

== About the Collection Description Application Profile ==

The following sections in italics are pasted from [2] and should be edited down to summaries.

The purpose of the CD Application Profile is to "provide a means of creating simple descriptions of collections and collection-descriptions suitable for a broad range of collections. It is designed primarily to support the discovery and selection of collections, though it may be used to support other functions such as collection management too. It is not intended to describe every possible characteristic of every type of collection".

Collections and collection-level description

The term "collection" can be applied to any aggregation of physical or digital items. Those items may be of any type, so examples might include aggregations of natural objects, created objects, "born-digital" items, digital surrogates of physical items, and the catalogues of such collections (as aggregations of metadata records). The criteria for aggregation may vary: e.g. by location, by type or form of the items, by provenance of the items, by source or ownership, and so on. Collections may contain any number of items and may have varying levels of permanence.

A "collection-level description" provides a description of the collection as a unit: the resource described by a collection-level description is the collection, rather than the individual items within that collection. Collection-level descriptions are referred to in Michael Heaney's An Analytical Model of Collections and their Catalogues as "unitary finding-aids" [AMCC].

Collection-level description enables a collection provider to

It enables a user to

Increasingly, many of these functions - notably the discovery, location, selection and comparison of collections - are being carried out by software acting on behalf of a human user, perhaps in accordance with user preferences or with parameters describing the scope of a particular service.

Functional Requirements of the DC CD AP

The DC CD AP is intended to provide a means of creating simple descriptions of collections and collection-descriptions suitable for a broad range of collections. It is designed primarily to support the discovery and selection of collections, though it may be used to support other functions such as collection management too. It is not intended to describe every possible characteristic of every type of collection.

The aim is that the DC CD AP should support:

CDAP Application Model

The DC CD AP is based on a data model which is derived from that described in Michael Heaney's An Analytical Model of Collections and their Catalogues [AMCC] and Users and Information Resources: An Extension of the Analytical Model of Collections and their Catalogues into Usage and Transactions [EAMCC]. The model used here is both a subset and a simplification of that model. The entity type which in that model is referred to as "Resource Mediator" is referred to here as "Service".

Review of CDAP by the Usage Board

The Board is of the view that this application profile is almost ready to be submitted as a complete proposal to the Usage Board for formal review and compliments the Working Group and the chair Pete Johnston, on the dedication and effort that has been put into the work. The Board asks the working group to resubmit its proposal for re-consideration and recommends the following changes and additions:

1. The data model used for the CDAP needs a better explanation. This explanation should include a statement about how the model in the Application Profile diverges from or otherwise amends the AMCC model cited as the basis for the profile. The Board feels that more explanatory material is needed to assist in relating the data model to the AP where dependencies exist.

2. The document should include a summary statement of the purpose and scope of the CDAP.

3. The Board feels that the distinction between "use of CDAP for collection descriptions" and "use of CDAP for collections of collection descriptions" is a bit confusing, and the CDAP should more clearly separate the two. This might be achieved by splitting the current CDAP into two separate APs -- one for collection descriptions and one for collections of collection descriptions. Editorial decisions of this nature are at the discretion of the working group.

4. The Board suggests the Working Group reconsider the change of label for the term 'Collection-Description' [2]. The Board feels that the old label, 'Catalogue or collection description' [3] is clearer and easier to understand because the label itself is an example of a typical use for the term.

5. The term 'Content' is not defined in the AP, but is used in a definition [4]. The Board asks the Working Group to include a definition of content in the Glossary. The same applies to some terms used in the CDAP that are defined in the DCMI Abstract Model, such as 'Rich Description'.

In general, the Board prefers that Application Profiles be able to 'stand alone' as documents; ideally, readers of the CDAP should not need to refer to external documents in order to understand its logic and suggests that the working group consider adding some additional contextual information from external documents as an aid to the reader.

Documentation Submitted

This list to be turned into a more narrative characterization of documentation submitted.

[1] http://dublincore.org/usage/meetings/2006/09/manzanillo/profile-cdap/html/

[2] http://dublincore.org/groups/collections/collection-terms/2006-08-24/#cldcollectionDescription

[3] http://www.ukoln.ac.uk/metadata/dcmi/collection-application-profile/2005-08-25/#dcdescription

[4] http://dublincore.org/groups/collections/collection-terms/2006-08-24/#clditemType

Note - links to current documents about Application Profiles at: http://dublincore.org/usage/meetings/2006/04/profile-review/


Pete on the RDF representations - 2006-08-29

-- http://dublincore.org/groups/collections/collection-terms/2006-08-24/cldterms.rdf -- http://dublincore.org/groups/collections/colldesc-type/2006-08-24/cdtype.rdf -- http://dublincore.org/groups/collections/frequency/2006-08-24/freq.rdf -- http://dublincore.org/groups/collections/accrual-method/2006-08-24/accmeth.rdf -- http://dublincore.org/groups/collections/accrual-policy/2006-08-24/accpol.rdf

The RDF data should essentially be an alternative representation of what is in the HTML docs i.e. the only question marks are to do with that representation (e.g. the relationship etween a value in a VES and the VES, if we change the DCAM so that it is not is-instance-of (rdf:type))

I put up RDF/XML representations of the "collection description terms", the type vocabulary, and these three vocabularies, and the PURLs for the terms should de-reference to those RDF/XML docs (i.e. in the same way DCMI serves one doc per "namespace"). But some of that data is incomplete/tentative, pending some decisions about the DCAM and/or property ranges/domains, so at the moment it's probably best to treat the RDF/XML stuff more or less as a "placeholder".

All the term URIs (I hope) de-reference to something useful -- at the moment that's an RDF/XML document, but in the future we might set things up so that agents can get alternative representations (e.g. HTML for a browser displaying stuff to a human reader, RDF/XML for an app that wants to get the data about the relationships between terms in a form it can act on).

Note that the current RDF representations listed above are tentative/incomplete, pending discussions in the DC Architecture WG about changes to the DCMI Abstract Model, which would have an impact on how e.g. we describe the relationship between a vocabulary encoding scheme and a member term/value within that vocabulary encoding scheme.