Notes from the OASIS
WSRF TC teleconference
7th March 2005
Roll call
The roll call is kept on the TC web site
under the meeting record.
See http://www.oasis-open.org/apps/org/workgroup/wsrf/event.php?event_id=4815
Approval of minutes from the face-to-face meeting (21st
February)
See http://www.oasis-open.org/committees/download.php/11627
(DaveS) Correction: wrong title for
discussion of issueWSRF66 – should be about aggregation
PS: The corrected version is here:
http://www.oasis-open.org/committees/download.php/11714
(DaveS) Any objections to approving the amended
minutes?
None
Action Review
(Chairs) Investigate WIKI site for Developing AppNotes (carried forward
from 4th Feb) Carry fwd
(Bryan) Move issues 97, 98,99 and 100 to ‘Open’. Done
(TomM) Respond, hopefully with a proposal to resolve issue wsrf44. In
Progress
(Bryan) Move issues 58, 59 to resolved. Done
(Sam, TomM, DaveS) Propose resolution to issue 65. Carry fwd
(All) Review current drafts. Final reviews needed are now.
(DaveS) Has
anyone reviewed anything they didn’t write?
(IanR) We
need a ballot on this.
(DaveS) Yes
Action: Set
ballot for 7 day period. If successful the working drafts will become the new
drafts on the front page of the web site.
(JemT) We’ve checked all the external references
(IanR) Are the wsdl/xsd up to date with the recently posted SG spec?
(TomM) Not
Yet.
New issues to consider
- Bryan
WSRF 101: Use of non-normative references
(TomR) There
were some inconsistencies.
(SteveG) Are
these non-normative refs a problem?
(TomR) The namespaces
referenced are ibm namespaces. We need to justify the use of these things.
(SteveG) But
the references may be helpful. We should balance the advantage of removal.
(TomR) It
would take a week to examine all of them. MHO it is that we should remove
unnecessary references.
(WillamV) The
experience from WSDM is that we should only have the things that are really
needed.
(DaveS) It
sounds like an issue worth opening.
No objections
Action:
Move to Open
WSRF102: Invalid delete
(Bryan) There
is no fault corresponding to the update/insert.
(SteveG) Or
we could have one fault to address all of the modify
(DaveS) We
should move this to open. Any objections?
None.
Action:
Move to Open
Change to Public Web Site - Bryan
See: http://www.oasis-open.org/apps/org/workgroup/wsrf/email/archives/200503/msg00005.html
(BryanM)
There are no links from the public page of the web site to the June 2004
versions of the specs on which the WSDM specs depend.. Also, the schema for the
base faults spec is missing.
(IanR) I
think this would be confusing. These are old working drafts and they are still
at available at the URLs referenced in the WSDM Specs.
(TomM) We
could start putting a ‘Current version’ pointer in all specs, but that wouldn’t
fix past versions.
(WilliamV) We
could put pointers on the WSDM web site to the specs, but people go to the WSRF
TC site to look for those references.
(IanR) It’s
the wrong way around to put links from WSRF to all the consuming specs.
(WilliamV)
But WSDM is an important (the most important) consumer.
(DaveS) We’ve
spent a while discussing this. Is there anyone who wants a vote?
(TimB) I
propose a vote
(IanR) Seconded.
(DaveS) Let’s
hear from those in favaour of posting a pointer to the wsdm TC site.
Result: 6 in
favour, 10 abstains, 7 opposed
(DaveS) So we
should call this a vote against the motion.
PS: (DaveS)
There’s a typo in the email: b for B in ‘basefaults’ which explains the
apparently ‘missing’ spec.
Draft Status Check
(e.g. Service Group?)
(TomM) This will be there by midnight.
CD Schedule
(DaveS) We
have a set of drafts which are nearly finished in the sense that there are few
outstanding issues. (Some SG and WS-A). What’s the process for WS-AQ to go to
last call?
(TomR) We can
find out at today’s call. After ‘last call’ during which comments are taken
into consideration, then there is the approval draft, then the official
process.
(DaveS) So,
we have a stronger basis for our Committee Draft after the ‘last call’ is
issued.
(IanR) Does
the ‘last call’ have a namespace that we can reference?
(DaveS) All
the working drafts has a date-stamped namespace.
Action:
(TomR) Identify a point in time draft which is as stable (or better) then the
2004/08 on which WSRf depends, but closer to the expected final version.
(TomR) By the
way, Policy has gone as first class item, being replaced by ‘metadata’
(TomM) What is our schedule on this?
(DaveS) MHO
is that we should wait until we have a stable reference on WS-A.
(IanR) I
suggest we don’t have a reference to WS-A. If there is a stable point when we
are ready, then that’s good, but if the W3C is still in progress and that is
the only thing outstanding, we should go to CD.
(DaveS) I
don’t disagree with the CD, but we should not go for OASIS standard. If we go
to CD, we can wait as long as needed to wait for public comments.
(TomR) Public
is a separate vote to go for public comment, There can be several CD’s.
(DaveS) So,
hopefully, things will happen in W3C as we move along. We might expect two
CD’s. Any objections to keeping going?
None.
(IanR) Can we
go back to the base faults schema question? Which is the missing file?
(Bryan) I get
a 404 response
(DaveS)
There’s a typo in the email b for B in ‘basefaults’
Issue review
- Dave
WSRF65:
Need a mechanism to aggregate an operation across several resources in a
ServiceGroup
(TomM) The
usual use case (from Savas and the eScience community and Tony Hey’s article):
when an EPR is used to address a resource there is no opportunity to include a
by-value list representing multiple resources. This raises questions about the
value of the ‘implied’ mode of hiding the target identity.
(DaveS) The
choices are to build multiple operations with different signatures, or build an
entity like service group. However, ServiceGroup currently has no operation.
(TomM) Does
globus have this today?
(SamM) We do
not do the operations at the collection level. We have talked about doing
operations which combine a selection and internally drive an operation. For
example, select all jobs owned by me and kill them.
(DaveS) This
allows for optimization. The use cases are interesting, and are similar to the
WSDM use cases.
(TomM) Is
this an engineering exercise that we leave to an an AppNote, or should we try
to tackle it?
(TimB) We
could start with simple steps: an interator and GetNext.
(IanR) Do we
need this urgently, or can t be delayed until the next versdion of WSRF?
(WillamV) I
think it is needed soon.
(DaveS) Sue,
is DAIS able to invent it’s own iterator based on ServiceGroup?
(SueM) We
don’t use ServiceGroup, but we could look into them.
(DaveS) This
issue is an important use case, but an optional operation. I think it could be
layered in later in a subsequent release. We do have some precedents for
dealing with operations with multiple components,
(WilliamV) But
if we delay then we may find that communities like OGSA and WSDM diverge.
(DaveS) We
could work on the range of simple facilities from GetNext to Combining
operations with ServiceGroup.
(TomM) We
also need to gather input from the various communities.
Action:
(TomM/DaveS) Gather input and make proposals
Straggler Roll Call and Close
Summary of actions
(Chairs) Investigate WIKI site for Developing AppNotes
(Carried forward from 4th Feb)
(Sam, TomM, DaveS) Propose resolution to issue 65. Carried
fwd from 21st Feb
(Bryan) Move issues WSRF 101,102 to ‘open’.
(TomR)
Identify a point in time draft for WS-Addressing which is as stable (or better)
then the 2004/08 submission on which WSRF depends, but closer to the expected
final version.
(TomM/DaveS)
Gather input and make proposals for issue WSRF65