Contextual Wrappers 2

packaging University museum collections information as part of open metadata provision

Monthly Archives: March 2012

Designer Briefs

The project’s aim to develop a prototype search-portal of UK University Museums took its first steps during March, through further engagement with stakeholders: with a brief being agreed with UMG: Design brief for prototype University Museum search and a call being issued to UMG members to provide images that represent their collections and services: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1203&L=UNIVERSITY-MUSEUMS-GROUP-UK&F=&S=&P=6910.

Advertisements

Collection Records Profile for Grid: Part 2

Recap:

These kinds of records are being developed by the project to support users of the Culture Grid aggregator service to:

  • Understand the scope, legal status and usage-terms of the ‘collections’ Culture Grid holds and makes openly available
  • Make content selections for different kinds of services/applications to be produced
  • Drive interface features based on the information and relationships inherent in these records

Additionally, these kinds of records will also support service-management tasks, such as reporting.

The records will use an ‘application profile’ similar to the RSLP-CLD (http://www.ukoln.ac.uk/metadata/rslp/) based one that was used for the Wrappers1 project.

Innovative approaches in the implementation of the profile will include:

  • Using controlled term lists from the Culture Grid Vocabulary Bank, including the JACS terminology to assign course/module-based Higher Education relevance (http://culturegrid.lexaurus.net/culturegrid/browse)
  • Auto-generating values from ‘look-up’ and other records contained within the Culture Grid

The latest version of the profile:

Following ‘offline’ testing of the RSLP-CLD based Wrappers1 profile for describing ‘collections’ in the Culture Grid, the profile, which will be tested in iteration 2 of the system developments, now has the following mandatory and recommended elements, with a number of other RSLP elements being auto-generated and the rest remaining as optional:

Mandatory elements, (which were kept to a minimum to allow identification, provenance, legal status and system-status to be recorded)

Identifier – single value. Auto-generated from system.

Title – dc:title. The name of the collection. Single value. 

Owner – multi value. Uses institution:identifier

Legal Status – A statement of the legal status of the collection.  Single-value, with an editable default value to indicate the provenance and rights status of the originating collection/items as a whole and the terms of use of the records from Culture Grid.

Type – Multi value.  From controlled list of: item; collection; institution; top-level; provisional

 

Recommended Elements (to describe coverage of the collection/items referenced in collections)

 Description dc:description. A description of the collection. Single value

Temporal coverage dc:temporal. The temporal coverage of the items in the collection. Multi-values from extensible list including MICHAEL Period terms from http://culturegrid.lexaurus.net/culturegrid/browse

Spatial coverage – dc:spatial. The spatial coverage of the items in the collection. Multi-value. (options being investigated to to populate/control by ‘look up’).

Subject – dc:subject. A concept (keyword) of the items in the collection.  Auto-generated from top 50 terms from subject-facet.

Subject & encoding scheme = JACS – (non RSLP) Controlled terms from HE course/module related thesauri at: http://culturegrid.lexaurus.net/culturegrid/browseMulti-values.

 

Auto-Generated Elements:

 Size (non RSLP) dcterms:extent. The size of the collection. Single value auto generated from the number of records in Grid collection once ingested.

Accrual status- cld:accrualStatus. A statement of accrual policy (closed, passive, active, partial/selective), accrual method (purchase, deposit)) and accrual periodicity (closed, irregular, periodic). Single value.  Auto-generated (oai=active; upload = passive; method = oai-harvest or upload; period = irregular for upload, harvesting default for oai, and closed for any known to ‘never come again’)

Super-collection dc:isPartOf. The identifier or name of a second [Grid] collection that contains the current collection. Single value. Auto generated.

 Sub-collection dc:hasPart. The identifier or name of a second [Grid] collection contained within the current collection.  Multi-value.

 Associated collection cld:has Association. The identifier or name of a second collection that is associated by provenance with the current collection. Multi-value. Auto generated from collections with same owner/provider.

 Optional

All other RSLP/CW1 elements will be optional

Collection-records of collections of collections of collection-records – the sometimes convoluted world of aggregation

The project has been active in several areas, following a team meeting at end of Feb, which saw the project further align with the Open Book project, (http://www.jisc.ac.uk/whatwedo/programmes/di_informationandlibraries/resourcediscovery/openbook.aspx)

Activities include dissemination of the consultation findings through different stakeholder channels and further engaging with University Museums to confirm focus groups sessions in May and June to review the prototype search services that the project will develop over March/April.

The first of the focus-groups should be in late April at the Fitzwilliam Museum, University of Cambridge, combining with the engagement activities for Open Book.  Another is confirmed for 14th May at The Egypt Centre, University of Swansea, (http://www.egypt.swansea.ac.uk/ ) in conjunction with ACCES, the Subject Specialist Network for museum curators looking after archaeological collections from Egypt and Sudan (http://ssndevelopment.org/acces_ssndevelopment/home/acces/public_html/ ).  Other focus groups are being confirmed with UCL, through join-up with the OBL4HE project (http://blogs.reading.ac.uk/oblhe/) ; The University of Aberdeen and  Queens University Belfast.

A major activity prior to next aggregator-system & interface developments is producing an application-profile for ‘Grid-collections’ i.e. for the collections of resource-discovery records that the Culture Grid aggregates.  These are required to support consumers of data from Culture Grid – to easily understand the scope and usage-terms of the ‘collections’ that Culture Grid  provides; developers to potentially build interfaces based on the relationships inherent in these records; and for service-management purposes, such as reporting etc.

Producing this application-profile is proving to be something of a brain-twister, as Culture Grid as an aggregation-service ingests and holds collections of item-records; collections of collection-records; and collections of institution-records, all of which potentially have poly-hierarchical relationships and associations with one-another, so the situation could arise where a collection record of collections of collections of collection-records would be required!

The aim is to try to reduce this potential complexity to something fairly simple to implement and the approach taken to try to achieve this is to test the RSLP-based application-profile that was developed for the first Wrappers project as a basis for the Grid-collection records, by doing some sample cataloguing of the different kinds of ‘collections’ within Culture Grid, to see which elements should be mandatory, which controlled or constrained, and any extra elements to include etc.

This exercise also includes looking at the use of the JACS vocabulary for subject-terms to potentially allow subject-browsing for HE audiences at the level of ‘Grid-collections’.  (JACS has been incorporated into the Culture Grid Terminology Bank, (http://culturegrid.lexaurus.net/culturegrid/browse) for this purpose.)  The profile-investigations have also included a revisit of the RSLP schema (http://www.ukoln.ac.uk/metadata/rslp/) and the DC collection-application-profile  (http://dublincore.org/groups/collections/collection-application-profile/).  The results of these explorations should be known shortly to inform second iteration developments.

A similar exercise testing the CW1 profile, is also to be undertaken by The Fitzwilliam with regard to other University of Cambridge collections; and the ACCES SSN are looking at applying it for their records of archaeological collections from Egypt and Sudan in the UK, particularly with regard to adding basic excavation information as part of their collection records and controlling temporal and spatial terms by appropriate authority-lists.

Phill Purdy,
Culture Grid Manager, Collections Trust