[dais-wg] DAIS Telcon Mins from 07/06/05

[Mario Antonioletti]

DAIS Telcon - 07/06/05
======================

Chair: Norman Paton
Notes: Mario Antonioletti

Attendees:
	Norman Paton, University of Manchester
        Mario Antonioletti, EPCC
	Dave Pearson, Oracle
	Allen Luniewski, IBM
	Savas Parastatidis, University of Newcastle

Agenda:

 o Specification issues.

Actions:

[Mario] Find out what practice is used for versioning GGF documents.

+--

Norman: Trying to come up with an update facility for DAIS. This is
        still a work in progress.

        Mario do you know what the status of the specs is at the
        moment?

Mario: have been iterating the relational spec with Susan.  I think
       Simon has the token on the core spec. I'm not sure what
       progress on that is. Not sure if Amy has made done the XML
       spec ... she was busy last week.

...

Mario: For GGF14 do we aim to go in as a proposed recommendation?

Norman: submitting to the proposed recommendation track happens after
	GGF. Hope the version submitted to GGF will be close to the
        proposed recommendation.

Mario: Ah I thought we were submitting then. How much can we change
       after?

Norman: well there should be no known bugs in the specs. They ought to
	do what we think it should and describe what we intend it to
	do.

Mario: what things like ws-address/wsrf which will not be ratified in the
       same time frame?

Norman: specs will have to change - external dependencies on ws-addressing
	and wsrf need to be explicitly noted.

Mario: the fact that submission happens after GGF14 also means that we
       do not have to specify the principal issue contact on the
       footer of the specs as required? We can worry about this later.

...

Mario: Is the WSDL being produced supposed to be normative? Some of
       the language is a bit loose on this.

Norman: it will be normative. The language needs to be tightened.

Mario: Is the data description stuff an interface or a collection of
       properties? It's currently being referred to as an interface 
       in the relational spec...

Norman: depends how we use the word interface - it really is a
        collection of properties which will be associated with some
        operations.  Need to be careful how we use the word
        interface. Need to go and look at how Simon used it in the
        core. Check how Simon defined it.

Mario: How closely should the WSDL follow the spec in terms of naming?
       e.g. SQLAccess and the like - should that map to a portType.

Norman: because of the issue with portTypes in WSDL ... need to check.
        We need to know what we are saying ... probably have to be
        silent if they are going to be composed and the definitions
        are going to be extended the definitions, e.g. result sets.

Mario: having a problems with the usage of CIM in the relational spec
       as we do not have a definition as yet - so we need to do
       something for GGF. Could use an xsd:any and remove the specific
       detail about the content from the relational spec or we could
       have an xsd:any which is restricted to the DMTF/CIM namespace
       but this too has not been decided on yet. Another alternative
       is to come up with our own elements names with the proviso that
       this will be swapped for the proper ones later on ...

Norman: needs to be an extensibility point - do we provide a set of
	relative specific set of properties - need to make sure that
	they line up with the corresponding CIM model. A plausible way
	would be to have the different properties - and each of these
	would be an any and would describe the schema that it conforms
	to and then there would be a statement ... to the effect that
	this will be tightened up - to produce a separate spec. For
	now allow people to plug things in and in due course that will
	be tightened up and the CIM stuff will come up in time.

Mario: ok can discuss at some later point ... so that's most of my
       current issues addressed ... though not necessarily resolved.

Norman: Allen can you say a couple of word about what the OGSA Data WG
        are doing at GGF14?

Allen: have put a version of the Data strawman architecture up on Grid
       Forge. There's stuff on replication, data transfer, a section
       on storage ... aim to have three sessions at GGF14 - on
       introducing the group to the community, a session that will
       concentrate on data transfer .... part of that is going to be
       looking at infod - the third session is with the transaction
       management research group ... neither of the chairs will make
       it to GGF - also the group is in the process of shutting down -
       submitting their final report. Want to do transactions as it
       crosses all the pieces. Do the same with security ...

Norman: people from DAIS should try to have a look.

...

Norman: Savas did Simon copy you in the last version of the WSDL?

Savas: had a call with Simon where the conduit was discussed. I think
       I understand it now. Told me that he was going to work on the
       WSDL - told me not to look.

Norman: it would be good to have your input on the WSDL once it comes
	out.

...

Norman: GGF have not published the schedule yet - would be helpful if
	they could schedule all the DAIS sessions before Wed 2pm.

...


+-----------------------------------------------------------------------+
|Mario Antonioletti:EPCC,JCMB,The King's Buildings,Edinburgh EH9 3JZ.   |
|Tel:0131 650 5141|mario at epcc.ed.ac.uk|http://www.epcc.ed.ac.uk/~mario/ |
+-----------------------------------------------------------------------+