INFOD-WG  8 Jun 2004  14:00 to 15:30

There were about 30 present. Their attention was drawn to the GGF IP
policy.

Susan Malaika introduced the agenda and explained that we are now
approved as a working group and that this was therefore the first
meeting of that group.

We have to understand the link between INFOD and a number of other
activities and standards. For example can we build on top of WSN?
There are also obvious links to the OGSA-WG data and information
services work. Norman Paton explained that INFOD originated in the
DAIS-WG as they realised that flexible data delivery and 3rd party
delivery were major issues which did not fit well under the DAIS
umbrella. Concerning the scope of INFOD it was stated that INFOD does
not concern itself with data transport. At GGF10 it was thought that
there may be a link with the OREP-WG however this no longer appears to
be the case.

Dieter Gawlick then presented his view on the scope of INFOD. It was
based on a previous presentation at GGF10. David Martin asked if
communication was outside the scope of INFOD. Dieter replied that we
must know that communication has occurred and know the properties of
the communication channel. Dieter explained that WS-NF was not
adequate to generate a notification when a fish is more than 3 days
old. Evaluation of use cases is necessary to understand the
limitations of the classical pub/sub. We need rules which are able to
watch the evolution of information. In response to a query, Dieter
explained that the publisher (which will probably be renamed producer)
is assumed to have an SQL query engine. A concern was expressed that
the query engine would need to be able to process temporal
queries. Somebody then stated that as 90% of the data in the world is
unstructured, INFOD was much too RDBMS oriented. A fear was expressed
that it was a big step from what people are used to. David Martin said
that the spec would have to carefully distinguish between the "musts"
and the "shoulds" and the "mays". Norman Paton however said that if
the infrastructure could only be relied upon to deliver the "musts"
the rest of the spec was not of much use. There was then a brief
discussion of how this might fit info WS-NF and whether or not it
could be layered on top of WS-NF.

It became clear that the scope of INFOD needs defining very carefully.

Susan Malaika then showed the next steps: a pair of draft documents by
GGF12. Work must start on the use cases immediately. Susan reminded us
that the next conference call will be 2 weeks on Friday and will be
announced to the mailing list.


Steve Fisher