OGSA F2F
========

15 March 2007

Particiipants
=============
Michel Drescher, Fujitsu
Fabio Benedetti, IBM (phone)
Fred Brisard, CA (phone)
Donal Fellows, UoM
Andrew Grimshaw, UVa
Chrs Kantarijev, Oracle
Hiro Kishimoto, Fujitsu
Allen Luniewski, IBM
Steven McGough, IC
Mark Morgan, UVa
Andreas Savva, Fujitsu
David Snelling, Fujitsu
Ellen Stokes, IBM
Ravi Subramaniam (phone)
ZKert


1. JSDL
=======
- Andreass gives an update on JSDL
- same sliddeset as given at OGF19
- Andrew: There are a number of BES and hence JSDL implementations out there
  -- This "user base" must be respeccted
- Andrew iterates on the DataStaging element in JSDL 1.0 and its semantics
  -- How does this influence JSDDL 2.x?
  -- current implementation is an easy way to describe dat preconditions 
     and data postconditions
  -- The group discusses the different approaches on
     --- pre-/post-conditions
     --- requirements/cappabilities 
  -- Andrew proposes that pre- and post-conditions are separate elements in 
     a JSDL document sibling to requirements/capabilities
  -- Which information should be in requirements, and which in pre-conditions?
  -- Fabio: Concerns expressed about how one job can express a dependency 
            upon the completion of a second job
  -- Andreas: roll out proposed new JSDL features as extensions to V1.x. 
              Use this as a means to test proposed new JSDL features and then
              will be in a better position to define JSDL V2
- Jay: Propose to publish a draft of JSDL2
  -- new namespace, new schema
  -- selected new elements of JSDL 2 get back ports to JSDL 1.x
- The group agrees to nail down a very early draft of JSDL 2.0
  -- the specification for JSDL 2.0 will contain non-normative text, i.e.
     immplementation notes, rationales, etc.
- Jay: Propose to provide an XSLT that converts a JSDL 1.x <Resources> section 
  into JSDL 2.x requirements
  -- Rationale in this should include pointers to tooling that would help
     users creating the XQuery
- Andreas: Propose to allow non-XQuery requirements
- Andreas continues to give a mind map based overview about the issues 
  for JSDL 2


2. Information Model
====================
- Ellen introduces a draft set of XML tags for the Information Model
  -- Ellen provides some background information about info model work in
     BES and GLUE schema
- Dave: What's the virtue of using/starting out with the GLUE schema?
  -- Ellen: GLUE has been around for a while in Grid context, and fairly 
     wide adopted
  -- political opportunity to tag team with the GLUE-WG in OGF
- The group discusses the benefits and drawbacks why to start off on GLUE or not
- The group takes the GLUE schema 1.3 and starts to modify selected elements 
  to the Information Model needs
  -- This is not so much as to modify GLUE 1.3 but more of a profile excercise
     to further the Information Model work for a concrete tag list