Oct. 3, 2005
Agenda and Notes for Telecon, Oct. 3, 2005
This conference call needs to be sharply focused on infrastructure/technology questions
1. Commitment to Resource Description Framework (RDF) and its possible implications:
[Discussion] Stuart talked about the value of using RDF, where the structure of the data is encoded in the data itself instead of relying totally on a relational database. Ryan and Stuart's experience has been that this is very useful, and supports inferencing. There are still issues of how much of the implementation is based on RDF. Jon feels that there are intrinsic limitations on what other technologies you can use. Jon is particularly concerned with what languages can be used, particularly if we want to distribute the software. Ryan agrees that we need to pick technologies that are supported over time. Stuart points out that we didn't promise to distribute software, rather to support distributed registries that could communicate with one another. Diane points out that we do need to think about distribution, lest we be the only extant registry (an unsustainable prospect). Stuart agrees that we need to think about this as we move along.
Action item: Ryan and Jon will explore the specific choices about RDF servers/data stores and query options and make a recommendation.
2. SKOS as "default" scheme generation framework
Stuart feels that SKOS is the only available choice--really not much out there. Early attempts did not work out. Using SKOS will enable us (require us) to provide some tools to assist users to generate vocabularies. Stuart believes (and Diane agrees) that SKOS exceeds the power that most vocabularies need and is extensible. Jon pointed out that Alastair says it's built on OWL and extensible in that direction. All agree that SKOS is a good starting place and will allow us to build usable interfaces.
Stuart would like to implement an editor that uses SKOS, and believes that this will fulfill most of our needs for the immediate future and leave OWL for later. We are unlikely to do full inferencing or support OWL for anything except import/export. We all recognize that we need to be able to ingest ontologies encoded in OWL as well as SKOS but that we need not support creation/editing etc. in the foreseeable future (if ever)--so, it is ingest and display.
Jon asked about ISO11179, and Stuart quoted Rachel H.'s feeling that the ISO community was very disconnected from Semantic Web vocabulary issues. Stuart thinks that it may have some effect on describing schemas at some point. Ryan asked about application profiles and Stuart mentioned that we will talk about some of those issues at our F2F meeting.
3. Issues of Zope/Plone (or any other content management system) to handle select functionality.
Stuart: The big issue is whether we can use Zope/Plone for efficient implementation of functionality where we don't want to spend a lot of effort, in hopes of spending more time on things that are more the focus of the project. Workflow, authentication, role management, display are all handled fairly well, from Stuart and Ryan's point of view.
Diane suggested that because she and Jon are not used to the product that some effort be made to have things more transparent and know what's defaulted and what intentional. Stuart suggested that we focus on questions of what might not be possible or not easily possible, that we'd want to do. Ryan expressed a fondness for the snap-in approach. [Jon and Ryan discussed some technical issues that I couldn't track well].
We talked briefly about the Blog issues, including some questions about how new users will be added. Right now Diane will look into the site regularly to approve people, but Ryan will look into implementing a mail option (will be needed for later implementations.
Action item: Jon will contact the KNOTES people with a bug report on the 'floating scrollbar' Action item: Diane will issue invitations to the Registry WG as outlined in her message earlier today.
Call ended, items below deferred.
4. (Time permitting) Immediate next steps in preparation for meeting in Ithaca:
- Completion of first iteration of draft Use Cases
- Begin work on RDF resource modeling
- Agent class (stemming from Use Cases)
- Registry Managers
- System Administrator
- Registered User
- Organization Contact
- SKOS (working vocabulary representation)
- Concept Schemes
- Agent class (stemming from Use Cases)