Minutes of ICOM TC Meeting, March 31, taken by Eric S. Chan
Agenda
1. Roll Call
2. Approve draft minutes from March 3 and 24 TC
Meetings
3. Review RDF ontology in Drupal Neologism
4. Review the questions posted by Deirdre and
Laura
5. AOB
1. The following eligible members were present
Laura Dragan
Deirdre Lee
Patrick Durusau
Eric Chan
2. Approve draft minutes from March 3 and 24 TC
Meetings
Draft Meeting minutes approved.
3. Review RDF ontology in Drupal Neologism
Participants reviewed the RDF model of ICOM in http://resources.smile.deri.ie/icom/
(thanks to Laura and Deirdre for facilitating the Neologism ontology authoring
tool). Accounts have been created for Laura, Deirdre, Patrick, Marc, and Eric. Please
send a request to Laura if you need an account for this online tool.
Deirdre and Laura provided the following installation
resources for Neologism:
http://code.google.com/p/neologism/downloads/list
(neologism-1.0-rc8.zip “Neologism 1.0 Release Candidate 8” is
the current version)
http://drupal.org/requirements
http://drupal.org/node/628292
(in the installation guide, wherever it says 'Drupal', replace with 'Neologism')
The RDF ontology can be downloaded in RDF XML or N3 format
by clicking on the following icons in the Neologism tool.
Eric asked about hiding some nodes and properties from the
diagram layout to provide different perspectives of the ontology. Eric to
create a new RDF model for Figure 8 in http://wiki.oasis-open.org/icom.
Labels in the diagram can be dragged by simultaneous pressing Ctrl and mouse click.
Eric noted that icom:Identifier is not in the RDF ontology. Need
to review whether it should be defined in RDF or whether rdfs:Resource can map
to icom:Identifiable.
Some auxiliary concepts, such as Identifiable, Parental,
Addressable, are in adjective terms, which make sense as “mixed in”
for objects in OO. Need to review whether the auxiliary concepts should be
renamed to nouns.
Icom:Interface is introduced in RDF ontology to designate
the abstract class concepts. Need to define a similar concept to designate the
abstract properties. Need to define axioms regarding the abstract class and
property.
Deidre mentioned about a workshop at DERI around end of
April to get feedbacks (from DERI internal only) for the RDF ontology for ICOM
and the mechanisms for abstract classes and properties.
Participants exchanged ideas for ICOM connectors for Drupal.
Although Drupal may not provide API for custom clients other than the PHP web interface,
it is possible to access Drupal contents in RDF format. At SemTech 2010, there is
a session by DERI researchers to present the RDF access for Drupal contents, which
will provide access to SIOC and other ontologies in Drupal. It is possible for ICOM
to map to RDF store in Drupal. ICOM as a frontend to the RDF store opens up an
access path from Java and other language API’s to Drupal contents.
4. Review the questions posted by Deirdre and
Laura
1.
Are AccessControlList, AccessControlEntry,
and Marker all root classes? i.e. they are not subclasses of Entity?
Marker
is a subclass of Entity. AccesssControlList and AccessControlEntity are not
entities.
2.
Is the description property only
associated with Scope, Space, and Folder,
or is it associated with all Entities?
The TC to review whether
the Description should be defined in Artifact, and why it is not defined in
Entity (this is a schema issue for UML model)
3.
Should we change subCommunities to communities,
as in the Group class we have groups and not subgroups?
Participants agreed to rename
the attributes to subCommunities and subgroups.
4.
Could we replace elements property in Space
and Folder to folders and artifacts
to distinguish?
The elements property is
polymorphic at the superclass ArtifactContainer of Space and Folder. It
allows us to define elementOfArtifactContainer from which we can extend two sub-properties
elementOfSpace and elementOfFolder. Also need to consider the range for
elementOfFolderContainer.
5.
Are the userCreatedOn/userModifiedOn
properties of Artifact the same as createdOn/modifiedOn
properties inherited from Entity?
The userCreatedOn and UserModifiedOn
are different from the attribute in Entity. The userCreatedOn reflects
the time when the artifact is created (may be while the client is offline)
while createdOn is when the artifact is actually committed into the repository.
Similarly for the userModifiedOn time.
6.
Regarding the folder models, we would suggest to create
separate extension models for each type of folder, e.g. calendar, addressbook,
forum. Are each of these types modeled in more detail anywhere?
The TC to define the specialized
folders soon.
7.
We are unsure what to do with TagApplication
and CategoryApplication. Can we discuss their use-cases some
more?
They are needed only for
UML model, please refer to the explanation for their role in http://wiki.oasis-open.org/icom,
“UML
Links: Category Application, Tag Application, and Bond Entity Relation” Section 9.0.
(There are no counterparts for CategoryApplication,
TagApplication, and BondEntityRelation in the RDF model).
8.
Similarly, we would like to suggest not using Bonds
for common use-cases. For example, if we make spaces hierarchical, there is no
need for Bonds in the workspace use-case. From the wiki, it suggests that Bonds
are included to address issues in translating the model to RDF. However the
model should be language independent. Also, we believe that the inclusion of
complex model structures may be a barrier to adoption. We are open to
discussion on this point.
Eric to research the use
cases of Relationship model in common repositories, such as MOF.
5. AOB
Eric to send out the hotel rates
in Redwood Shores around June 28.
Eric to include Laura as a
co-presenter of ICOM poster in SemTech 2010.
The TC Meeting was adjourned.