[infod-wg] Minutes 28 April 2005

[Chris Kantarjiev]

Dieter, Cecile, Abdeslem, Susan, Vijay, Chris (minutes)

Special guest: Stephen Davey

Apologies: Steve Fisher

Stephen presented a use case for NextGrid from the National e-Science Centre.
(See the email archives for the PPT.) We discussed how well INFOD would
be suited for this - both from scalability and usability perspectives.

There was a lot of back and forth to try to tune the use case and
requirements, and discussion of how the interfaces match up.
Cecile will work with Stephen to map the interfaces to the use case.
Dieter will work with Stephen to map the requirements.

We had a long discussion about a "suitability" measure to indicate
how use cases match the scope of our interfaces - what do we do
when a use case requires one or more interfaces that don't match
something in INFOD, but we consider out of scope?

Susan will try to figure out what to do about merging suitability
requirements into the use case document.

Other news:

Not much progress on the use case doc.

Cecile reported on work done with Dieter to refactor/reframe the
artifacts/roles/interfaces: they have come up with a model that
cleans things up considerably and will work this into the next
version of the doc. The basic idea is to provide a common structure
for all the artifacts which is intended to make understanding
the registration manager entries and interfaces much simpler.

Vijay reported on work done to concretize the model, in order to
come up with enough detail to build an implementation.  Broke into
three categories: artifact type, operational constraints, instance
level requirements for particular artifacts (e.g., scheduling details
for a publication). There has also been some work to come up with a more
concrete format for INFOD messages, as well as details about
what will go into the registration manager.

Vijay will write something up for the list.

There will be a west coast working session on May 17, probably at Oracle.