Partial minutes from PI call 28 March 2003

(Sorry I missed the first two items - Mick?)

Attendees: David Karger, Eric Miller, Mick Bass, Rob Tansley, Kevin
Smathers, John Erickson, Mark Butler


David: Haven't looked at the document yet - will mail feedback over the
weekend.

Eric: I started reading it, but got pulled off. I have some comments on
paper, not yet fleshed out.

Hygiene question: web publishing with CVS history logs is a useful way of
providing history on this stuff.

David: I also prefer getting a pointer on to something on the web, rather
than an email attachment.

Eric: Does it had CVS version history?

Mick: No it does not.

Kevin: You could diff the two files as they are both on the MIT edu site.

Eric: I'm lazy, there must be a better way of doing it.

Mick: I'm happy to use CVS, Mark and I have talked about it, Eric if you are
up to do it then we can take an action item to do it. 

In the meantime we'll continue editing in parallel. 

Eric: We are doing some of this for SWAD-E, so we should be able to use it
for this project also.

Mick: Let's take this off-line, action item for Mick, Mark and Eric.

We did have feedback from most of the other folks who are on the call. There
is an issue list that summarises all of the feedback we have received
accessible  from the title page of the document. Rob would you like to
summarise your feedback?

Rob: Yes I'm not sure the motivating problems is discussing this, or
discussing appropriate techniques. I'm not sure we needed the split between
the  relationship and the interoperability sections. Also I think we are
lacking in HCI issues. We discuss it for schemas, but not more generally.
Finally of  these use cases, only part of them leap out as candidates for
using SW techniques so we need to consider the other options and demonstrate
the utility of the  SW approach.

Mick: Kevin?

Kevin: All my comments did make it on to the issue list. I'd emphasise the
importance of considering how the use cases address the need for
distribution. I  think this is a very strong reason for using RDF, so we
need to consider it up front.

Eric: Can I jump here? Both of those are good points. Why we are taking the
particular approaches compared to the techniques we are advocating. But one
thing  we need to accept up front is we don't necessarily know all the ways
upfront that this data is likely to be used. So the kind of data for
controlled  vocabularies, may be useful in other applications e.g.
discovery, browsing, facet searching. We're in it for the long haul, to
service applications we see  now, but also to provide data that can easily
be incorporated into new applications.

Kevin: That needs to be characterised into the use cases.

Eric: Yes, and if we can prove it then its more successful. 

Kevin: We (Genesis) want to cover distribution in the SIMILE use cases. 

Mick: So we want to add the unexpected use of data and distribution to the
use cases.

John: I willing to engage in conversation to understand what our intentions
are here.

ACTION: Mick and John to speak about this. 

Eric: I'd like to join that conversation as well.

John: I'll write an email, and we start a conversation there.

Mark: That goes to www-rdf-dspace.

Mick: I'd just like to review the changes we've made to the document:
- we've pulled in some material from the original project proposal
- we've included some of the comments and feedback, and filling out some of
the feedback here
- we've managed to fill up some of the sections that says "to be completed".
Some of the other PIs can also be helpful here, so in particular Rob and
Dave  I'd encourage you to take a look at the make this stuff easy for
humans to use sections.

Planning for the 9th

Mick: We anticipate having another version of the document turned around in
advance of the 9th. The 9th will be the first chance we've all had to look
at the  feedback, sit around the table, and discuss the motivating problems
and use cases. It seems it would be an excellent use of time to walk though
the document  and discuss
- are the use cases adequately captured
- how do we prioritise the motivating problems
- what is the relationship between the two
- we'd like to look at some kind of strawman for the type of prototypes and
demonstrators produced by Mark Butler and myself. So we can have a session
mapping use cases to demonstrators and we can make sure the demonstrators
map onto the use cases and motivating problems. In particular, we need to
think  what we can do in the first year. 

Finally we'd like to spend some time in a working session reviewing and
crafting job descriptions so that we can get those posted. 

Eric: So a successful meeting on the 9th would be if we finished the job
descriptions, we agreed on the proposals on the first year deliverables, and
we  provided feedback on the use case documents. 

Mick: Let's live in the full breadth of the use cases for a little while,
before drilling in on particular demonstrators, so that the demonstrators
are in  the center of the space. 

Eric: I've had enquiries about how the VIAF (?) maps onto different subject
areas. We could spend lots of time thinking about that, so we need to
balance  getting work done with creating new areas. 

Mick: We need to drive aggressively towards the first year plan, and the
other thing I think we need is to increase the amount of the bandwidth
available to  the project. We can do this by speeding up the hiring process.


Any other items before we close?

No.

Dr Mark H. Butler
Research Scientist                HP Labs Bristol
mark-h_butler@hp.com
Internet: http://www-uk.hpl.hp.com/people/marbut/

Received on Friday, 28 March 2003 13:57:52 UTC