Draft Minutes of ICOM TC Meeting, March 30, 2011

[Laura Dragan <laura.dragan@deri.org>]

Sorry about the delay with the meeting minutes,
Laura

---------

Minutes of ICOM TC Meeting, March 30

Agenda 

1. Roll Call
2. Approve draft minute for March 16 TC Meeting
3. Overview of the ICOM ontologies.
4. AOB


1. The following members were present.

Eric Chan
Peter Yim
Patrick Durusau
Deirdre Lee
Gofran Shukair
Laura Dragan


2. Draft minutes from March 16 TC Meeting was approved.


3. Overview of the ICOM ontologies & discussion.

Laura, Deidre and Gofran shortly presented the status of the ICOM ontologies, main structure, and issues that arose during transformation of the model to RDF. 
ICOM ontologies are modular, following the ICOM model structure, except for exclusion of some classes and properties that are redundant for RDF, like RoleDefintion, PropertyDefinition, RelationshipDefintion, etc.

Eric proposed DERI write a document specification for ICOM RDF similar to the specification for UML. Laura & Deirdre argued that writing such a specification would most likely lead to duplication, as the two models are the same. It would probably be more useful to write a set of guidelines on how to use the RDF vocabularies. Patrick suggested to start gradually with implementation of examples in RDF to discover possible problems with the model and then write the document.

Peter suggested we ask someone with OWL expertise to get involved. 
Eric suggested that we can ask members of the OWL2 WG to get involved, thus providing better visibility for ICOM. Peter argued that having the work done within our group would result a non-normative specification for ICOM.
In the past we discussed using only RDF for the ICOM ontologies, as there was no apparent need for OWL and it would add additional overhead that wasn't needed. However, we might need to revisit the decision and incorporate some OWL constructs. We already use small part of OWL, like max/min cardinality.

Discussion about where to host the ontologies, for the public review and after. Peter had proposed OOR (Open Ontology Repository) but it was down when Laura and Deirdre tried to use it. It also doesn't currently have a stable endpoint. There is the OOR sandbox - http://sandbox.oor.net/ . More information at http://ontolog.cim3.net/cgi-bin/wiki.pl?OpenOntologyRepository . UML and support for relational database schema is only on the "wishlist" so far, so those parts of the ICOM model cannot be stored there, if we decided to do so.
Deirdre & Laura proposed using http://vocab.deri.ie to host the ontologies. It is a public server with Neologism, hosted in DERI. Neologism provides an easy and convenient way to browse the ontologies.
Alternatively we can publish in the ICOM wiki HTML descriptions of the ontologies, auto-generated from the rdf files.

Discussion on the naming of properties - with or without "has", singular or plural form to indicate cardinality.
SIOC defines has_attachment and attachment, which is confusing. We should decide on a notation and be consistent through all the modules, not to confuse users. Ideally we should use in RDF the same notation as in the normative specification, for simplicity - without "has" in property names.
Java representation requires systematic changes of normative property names.

JIRA issue: missing body/data attribute in SimpleContent


4. AOB.

Agenda for next meeting could include: 
- follow up on Patrick's email, mixing ICOM data in the cloud, facilitated by ICOM security model
- follow up on progress of RDF modelling
- follow up on Patrick's discussion thread on promotion/marketing of ICOM
 

The TC Meeting was adjourned.