3. Approve minutes of 30th August
Proposed: Mark Little,
Scnd: Greg
No comments or
objections
Approved
4. Conformance section
for WS-CONTEXT
Greg's text (
http://www.oasis-open.org/archives/ws-caf/200409/msg00013.html)
was intended as compromise between the more and less wishes of various
people
Location would be at the end of the
document
Tony: Would like something, so Greg's text is good.
But, proposed text doesn't clearly
bring out what conformance can be
claimed for. Some specs there's only one thing - in this
case there
are several - documents, services, specs. There are likely to be
different
requirements on each, and whole of spec won't necessarily
apply to each kind of
thing that implements/uses/references
WS-Context.
Tony suggests that these be called out as
subsections. (In answer to clarification: as
sub-headings within a
conformance section). cf Tony's original proposal, and recent
email.
?: isn't all this implied.
Eric: we'ed look
to XML processing specs, rather than include our
own.
Formality:
Proposed: that Greg's text be
included in the text, at a location to be determined
by
editors
proposed: Jeff; seconded:
Mark
Jeff: what is the
"context-service-user"
Tony: It's defined in the spec. -
(location to be sought :-)
Tony:
Proposes friendly amendment,
add sub-heads that
identify the things that can conform, as in
http://www.oasis-open.org/archives/ws-caf/200409/msg00073.html
Martin: the sub-headings in that email don't quite
correspond to the Greg's
paragraph
Tony: yes, we would need to a bit of merging
work.
Peter: Phrase
as:
that each paragraph have a sub-head that
identifies the kind of artifact that
is constrained
by the conformance requirements of that paragraph.
proposed as amendment: Peter, scnded
Tony
Doug: (comment on
para 4, deferred till after amendment)
Eric: clarifying - that this amendment would be
editorial, not changing the
actual requirements, but
calling out what they applied to.
Jeff: For some paragraphs (e.g. 4) it is fairly clear what they apply to (a
context mgr), for others (e.g. para 3)
it is less clear
what the artifact would be.
Greg:
it would be the context structure in a message as exchanged, would have to
a semantically useful, derefenceable
reference
Jeff: If I'm selling
something, and want to claim it is conformant, so what
is
constrained. Tony has a point, it isn't quite clear what is
constrained.
Martin: it's saying
the reference scheme must be valid
..
Greg: is para 3 redundant and should be deleted
?
..
Peter: isn't para 3 describing "correct" use, rather than
conformant - a conformant
implementation can be (mis-)
used to send rubbish
Martin went
through the paragraphs - all but para 3 seem clear.
Eric: question: is this an editorial or substantive
question
Jeff: when we agree, it's
editorial, when not, substance.
Martin: we need to get clarity, closing the spec up.
Any objections to the amendment
Eric: object, it adds unnecessary
complexity
Voice vote: 6 for,
6 against, 2 abstain
tie - amendment fails by
robert's rules
Doug: in the light
of the discussion, proposes deleting para 3
2nded:
Eric
Jeff: what was the intent ?
the words don't quite capture it - what would we lose
by
deletion
Greg: intent was to
disallow meaningless garbage, but perhaps that's not quite
a conformance claim. Have to use identifiable addressing schemes (long
discussion
on this in section 2)
This is generally covered in section
2
Martin: any objections:
none,
amendment
passes
Doug: on (original)
para 4: what exactly is a protocol in "systems and protocols" -
should that be "systems and referencing
specification".
anyone can do
anything with the pass-by-reference must implement the ctx mgr, which is
too strong
Jeff: but a ref spec isn't an implementation
Tony: in another forum, they distinguished
implementations conforming and specifications
complying
Peter: that was a bit
silly really
Doug: questions his
own amendment !
clarifies : delete "and
protocols" from para 4
Jeff:
seconds
Doug: that paragraph now
says pass-by-reference use means must implement
the
context mgr, but really it only needs to support either offering the
service or interacting with it.
Martin: only the thing passing out the context needs to
implement the mgr service
Doug:
wishes to clarify the distinction between implementations that
must
implement and those that just
use
Proposes text to that effect, Peter
seconds
Peter: the text would
appear to say any system using pass-by-ref must offer the Ctx
mgr
service, though it clearly doesn't need
to
general disagreement that it
could be so interpreted
Voice vote: 6 for , 8 against, 1
abstain
amendment
fails
The main
motion was taken, as amended (once) -
Passed, no
objections
5. Security
considerations for WS-Context
Once the document is ready, Martin will open a 10 day ballot on Kavi
to approve
it as Committee Draft.
Action
MARTIN