Notes from the OASIS
WSRF TC Teleconference
June 28th 2004
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=4797
Approval of minutes from previous meeting
See: http://www.oasis-open.org/apps/org/workgroup/wsrf/download.php/7266/
There were no comments on the minutes and no objections to
approving them.
Confirmation of next call
NB: Due to US public holiday on 5th
July, there is a change in the duration of the call on July 12th (to
share the time with WS-N). The call will be 17:00-18:00 UK-Time.
(?) Is there a way to avoid these changes – by avoiding
Mondays?
(DaveS) We already searched hard to find a mutually
satisfactory time.
(?) Perhaps the schedule could be weekly for an hour?
(SteveG) But the roll call takes 15 mins – leaves little
time from an hour.
Acceptance of New Issues
(Bryan) A new step has been added to the issue closing
process (See http://www.oasis-open.org/apps/org/workgroup/wsrf/download.php/7439/WSRF_IssuesList.doc)
Namely, that ‘The completion of the changes must be
reported by the primary editor and verified by the secondary editor (or other
TC member)’
(DaveS) Proposed: to accept this new process. There
were no objections.
(Bryan) The issue describing ‘Simple WSDL Schema changes” is
complete – needs to be checked (per new process).
Action: ?
Issue 49: Use of Pseudo-schema in specs
Action: Move to Open.
Issue 50: Common faults Namespace
Action: Move to Open.
Issue 51: WS-AtomicTransaction
(see http://www.oasis-open.org/archives/wsrf/200406/msg00059.html
and http://www.oasis-open.org/archives/wsrf/200406/msg00085.html)
(Ian) The issue is to be inclusive of multiple specs (don’t
single one out)
Action: (Bryan) – add this point to the issue.
Action: More to Open.
Issue 52: Delete will always fault
See http://www.oasis-open.org/archives/wsrf/200406/msg00102.html
Action: Move to Open.
Issue 53: WSRP namespace
(see http://www.oasis-open.org/archives/wsrf/200406/msg00098.html)
Action: Move to Open.
Status of Published Documents
(Ian) The home page points to all the new
specs and WSDLs.
Ballot result was: 86% Yes, 14% Abstain
with 70% Quorum.
Other Action Review
(ALL) Ballot for working
drafts closes 18th July. Please review and ballot. Done
(Bryan) Change status of issue 6 to “resolved” Done
(Jeff) Raise new issue to
describe the need for description of notation. Done
(SteveG) Send a sample property
to illustrate maxoccurs>1 (ref Issue 15)
to the mailing
list for consideration. Done
(Bryan/Ian) Formulate the proposal
for Issue 16 and publish to mailing list. Done
(SamM) Clarify Issue 48 and
propose resolution via mailing list. Done
(William) Write proposed
recommendation to combining the ops for Issue 23. Done
(Bryan) Remove cross reference from Issue 21 to issue 32 (which
has been closed)
Done
(GlenDaniels) Write and publish an
expanded the proposal for Issue 21 (treating this
separately from Issue 23). This proposal should spell out the
operations
and why they should be split. Action: Carry forward (see discussion below).
(SteveG) Re issue 23 - write a
justification for the status quo (of split operations
for query)
and publish to mailing list. Action: Carry forward (see discussion below)
(SteveG) Provide a summary of
options to resolve issue 9. Done.
(Ian) Investigate a non-toll-free number for US access. Action: continue to investigate
(Ian) Add F2F details to
TC calendar. Done
Issue review
Issue 15: [Example of maxoccurs>1]
(SteveG) CIM Model for storage media
provided a couple of examples detailed here:
http://www.oasis-open.org/apps/org/workgroup/wsrf/download.php/7279
(DaveS) Proposed: to move the issue to
‘Resolved’ awaiting review. No objections.
Action: move
to resolved.
Issue 16: [Property to advertise Query expression dialects]
(Bryan) proposed the resolution described here:
http://www.oasis-open.org/archives/wsrf/200406/msg00095.html
Q(DaveS) there has to be minoccurs=1 so that the operation works?
A(Bryan) yes.
(DaveS) Proposed: to adopt the
solution. There were no objections
Action: move
to resolve.
Issue 48: [nillable properties]
Discuss proposal from Sam M.
http://www.oasis-open.org/archives/wsrf/200406/msg00049.html
+ Threads
Discussion Postponed to next meeting.
Issue 21 & 23: [Factoring of set/inquire operations]
Discussion based on completed actions
from William, Bryan, Glen, and Steve.
See:
http://www.oasis-open.org/archives/wsrf/200406/msg00067.html
http://www.oasis-open.org/archives/wsrf/200406/msg00096.html
(DaveS) OGSI had getProperty and GetMultipleProperties
(Glen) We need to keep this simple so that people
don’t feel motivated to invent their own.
(?) This can be done with a dialect in a
more generic op.
(Glen) But this has to be wrapped in a
generic operation.
(SteveG) Yes, we need to keep simple things
simple.
(TimB) We lose the return codes if things
are wrapped in a generic op.
(SteveG) There are ways to describe return
codes.
(DaveS) the first implementation would be
to wrap up the simple gets.
Q(IanR) What
are the objections to having three operations?
A(Bryan) Three interfaces seems more
complicated than one.
(TimB) But only one of the operations is
mandatory.
Q(DaveS) Do
we need more debate?
A(DaveO) Yes
– I’m not sure about the proposals or the tradeoffs.
(DaveS) The issue seems to be the
simplicity of the minimal implementation vs the simplicity of the maximally
configured interface.
(DaveO) About XPath.– the results are an
xml document, so may contain bits of the original – it may not validate.
(?) That’s right.
(DaveO) So, is this generic structure valid?
– a subset may be valid: a resource, a set of properties or a single property.
(DaveS) The returned document depends on
the dialect and the query, which is complicated. This seems (IMHO) to favour
the three operations.
(DaveS) Is there still a need for
discussion?
(William) I’m happy.
(DaveS) Then we need to vote. Those who
want a single op, speak up.
(4 votes)
(DaveS) Abstentions?
(8 votes)
(DaveS) In favour of 3 ops?
(8 votes)
(DaveS) Do we have a quorum?
(MartinC) A quorum is not needed on the
vote, so this means the matter is resolved in favour of the three ops.
(DaveS) Moved that issue 23 is
resolved. We just need to check that the doc is right (unchanged).
Now for issue 21 (split up the SetResourceProperties
operations)
(Glen) The simple implementation is to have
separate insert/delete/update. We shouldn’t require complex boxcar stuff.
(?) Right. We can add in the atomicity
properties later.
Q(DaveS) Why
were there not three ops in the first place?
A(SteveG)
From OGSI? However the three ops seems good as a new option.
Q(William)
If we had the three, why would we need the boxcar method.?
A(SteveG) To
reduce the number of round trips.
(DaveS) Let’s talk about adding the three
first, (Also need to talk about semantics of multiple ops vs the boxcar)
(Glen) We are specifying something whose
semantics are up in the air.
(SteveG) So should we talk about the policy
statements?
(Glen) Better to talk abut the simple ops
first.
(MartinC) The box car is a hands-off
approach to semantics – it’s an application issue.
(DaveO) So… let’s have a ‘do it’ op and the
rest is application defined.
(SteveG) SetResourceProperties is optional.
(Igor) Need to clarify support for
insert/update/delete vs boxcar – which can a client assume is supported? – it’s
complicated. On the other hand, the simple operations seem prerequisite to the
boxcar.
(SteveG) Can we get input from WSDM?
Action: Carry
forward discussion to next call (Ian). Get input from WSDM (DaveS/William)
Issue 9 (API to List Properties)
Postponed to next meeting.
Any Other Business
None.
Meeting closed 13:30 EDT.
Summary of actions
(Chairs) Continue to investigate conference call access
numbers for US participants.
(IanR) Review resolution of ‘Simple WSDL/Schema changes’
issue (see last page of issues document)
(Bryan) Move issues 49, 50, 51, 52 and 53 to ‘Open’ status,
and incorporate Ian’s proposal into issue 51.
(Bryan) Move issues 15, 16 and 23 to ‘Resolved’.
(?any?) Verify the resolution of issue 21 (ie doc is
unchanged).
(DaveS/William) Gather input from WSDM on requirements re
issue 21 (need for boxcar ‘Set’ operation).
(Chairs) Continue discussion of Issues 21, 9, 48.