Informal meeting with web group about access and indexing requirements and wishes.

Formalities

Participants: [Kåre, Toke, Stefan], [Hans, Hans, Mads, Bolette].

Vision from the web group

The DOMS is an object repository, that stores objects, metadata and relations. It handles basic queries and provides object access.

End-user searching is handled 100% by software outside of the DOMS. The DOMS is indexed by the web groups system, which handles advanced queries with weighting of results, correction of spelling mistakes and so on. Access to specific objects is done by querying the DOMS and transforming the result to human readable format.

Entities in the DOMS are typically made up of several objects, with strong interconnections. A cd-entity could be represented by an object pointing to a cover-object, a lyrics-object and some track-objects. RDF seems like a possible choice.

Objects in the DOMS are descriptive objects. Actual files for the digital resources are stored and - typically - accessed outside of the DOMS.

Preliminary requirements

Access to the DOMS by basic calls:

• getModifiedOrNew
  In order to update existing indexes.

• getAll
  Get all objects, probably as a remote iterator.

• simpleQuery
  Example: Get all objects belonging to a specific project.

• getObject
  Get an XML representation of a specific object.

Random thoughts

Splitting entities into several objects does represent a challenge for end-user presentation. Example: A CD could consist of 50 tracks. A model where the CD-object in unaware of the type of the objects it relates to would require that each associated object should be queried for basic data.

Caching or explicit knowledge about the layout of the different entity-representations are possible solutions.