Difference between revisions of "Feb. 14, 2007"

Latest revision as of 10:02, 15 February 2007

Valentine's Day Telecon and Notes, Feb. 15, 2007 (Moved from 2/14)

1. Change in telecon time to accommodate Joe.

• Change to Mondays at 8 Pacific and 11 Eastern.

1a. Static on the SKOS list...important? Jon thinks that some of this was provoked by a recasting of SKOS documentation by Alistair. There's also some functional requirements documents that are being discussed.

2. Diane's discussion with Karen Henry, notes:

• What pathways are trying to gather user annotations (even for local use), does she think they'd be interested in working with us on this. [Ask Kim about this, too]
  • Sounds like they still don't know what they're doing with this, going into it very slowly. I brought up the idea we'd had years ago about harvesting them, and she sounded like that was a new idea to her. I pointed out that it was important to be able to categorize them, and the Registry was the tool for that.
• How do they think Expert Voices is going? Are they getting much data? Are they still planning to use the simple model of relating the annotation data, or has something else been developed?
  • They're thinking of enabling some limited social tagging for their initial annotation service, so we might add something about relating tags to established vocabularies in our terminology services.
• Are they looking to offload Metadata Harvesting? Or is that seen as something that must be tightly tied to the NDR?
  • She was responsive to the notion that the OAI world has not embraced the NDR APIs and that coordinating and scheduling services might be useful. Interestingly, she also mentioned that audience project that SDSC is doing, and I think we should include that as one of our Terminology Service examples.
• Ask what the status is with the iVia contract--is that something that NSDL would welcome having us contribute to?
  • Sounds like they're putting lots of pressure on Johannes, they think iVia hasn't been providing much value for the money they've spent. They're working on another contract with them and seem to want to impose numbers on them (60% is what she threw out), which is clearly unwise and probably impossible. No pathways or projects have been able to use the service, mostly, I think because they don't understand the limitations or the product. There could be some possibilities here to include something in evaluation (is this collection one that could benefit from iVia services? Or NLP?) as well as in terminology services (maybe use the FAST stuff instead of LCSH, or pull out geographics?) Mapping from their rich text might also be an option.

3. Proposal discussions.

• One thing Karen warned about is biting off more than we can chew, so we might think about focusing the proposal more, maybe leaving some things (the GEM stuff maybe?) for next year. I think we don't want them to think we've no chance of delivering.
• Given Karen's reactions to the human evaluation portions, I think we should definitely beef that up a bit and talk about our ideas about moving to machine evaluation via APs.
  • Organization and Narrative
    • Jon will add some information about the Hub and relation to our orchestration paper. Diane will reorganize so that that what's now in Project Goals is Project Design with the hub stuff at the head. Diane will also add some concrete examples and descriptions based on some of the ideas from Karen's phone call.
  • Milestones
  • Budget and Overhead
    • We have to look into how to include Metadata Services Unit information
  • Tables and images
    • Table looks okay, will need some revision as we move on

4. Questions to NSF, answers from Alpha:

• The 100,000 limit applies to our Registry extension. They don't always allow up to $100,000
• Level of documentation needed: Need a full proposal, which means all of the pieces need to be there but not necessarily 15 pages. Additional questions identified in the documentation.
  • Diane will start that
• Those categorized as "continuing" is in NSF document 04-32--Stuart will scope that out.