Hi Ram,
We agreed that the only textual change was the one proposed in my first
e-mail on this subject:
http://www.oasis-open.org/apps/org/workgroup/ws-tx/email/archives/200606/msg00022.html
I would point out that this change had in fact already been
incorporated by Andy, in WD-06, which is now CD-02, as he rightly
viewed it as an editorial reflection of a voted decision of the TC. So
we've voted to make this change twice (or three times, depending on
your view :-) ).
Alastair
Ram Jeyaraman wrote:
Alastair, Monica,
Can one of you please summarize the editorial changes we agreed to,
relating to issue 45, in the last TC call?
It will be easier for the issues list to point to a specific email,
instead of grafts from several emails.
Thanks.
-----Original Message-----
From: Monica.Martin@Sun.COM [mailto:Monica.Martin@Sun.COM]
Sent: Monday, June 05, 2006 8:20 AM
To: Alastair Green
Cc: Joseph Fialli; ws-tx@lists.oasis-open.org
Subject: Re: [ws-tx] Unfinished editorial business re wsp:Optional
green: Hi Monica,
We definitely agreed to remove the sentence you quote, which also
includes the words "SHOULD NOT", and Andy has shown this deletion
correctly in the latest WD. I recognize that that sentence also
contained the word "MAY".
I think we all overlooked the fact that there was a sentence in the
preamble to section 5 (explaining the overall intent), which also
includes these words, so my suggestion is aimed at tidying that up in
line with the main change that you refer to.
My recollection is that we voted to support the semantics MUST and
MAY (Mark's motion to get agreement on design, before we hardened
implementation).
It was then agreed that the optional syntax amounted to MAY: the
server was declaring capability to accept a context and subordinate
processsing thereto, but was offering the client the alternative of
following the MUST or deciding what to do itself ("behavior
undefined"). If it decided what to do itself then it could decide to
send a context (same effect as following MUST) or not to send one.
This is MAY because the choice is in the client's hands.
So, I believe that saying "MUST or MAY" is right, but I could be
missing a subtlety here. An alternative approach would be to zap the
preamble altogether
fialli: Independent of whether the MAY flow concept is allowed to stay
in document or not,
I just wanted to point out that MAY occurs in two other places than
the preamble.
It also occurs in Section 5.4 "Assertion Example" on lines 314 and 322
in latest working draft[1]
-Joe
mm2: Perhaps the TC, then, should discuss any differences that exist
between syntax and semantics to achieve greater clarity. Thanks.
[1] link:
http://www.oasis-open.org/apps/org/workgroup/ws-tx/download.php/18557/ws
tx-wsat-1.1-spec-wd-06.pdf
green: New uploaded WD 0.6 ll. 245-246:
"1. whether a requester MAY, MUST or SHOULD NOT include an
AtomicTransaction CoordinationContext flowed with the message."
I believe we agreed that we only aimed to find a way of
representing the semantics MAY and MUST, and therefore that the
sentence should read:
"1. whether a requester MUST or MAY include an AtomicTransaction
CoordinationContext flowed with the message."
Alastair
I think we forgot to do something about the following sentence, when
discussing/resolving 045 at the F2F:
mm1: Alastair, I believe this is the paragraph that was to agreed to
be deleted; therefore, is not MUST the only specific requirement?
This is the sentence that was deleted which includes the deletion of
MAY and SHOULD NOT.
"Presence of both policy alternatives indicates that the behavior
indicated by the assertion is optional, such that an atomic
transaction MAY be flowed inside a requester's message. The
absence
of the assertion is interpreted to mean that a transaction SHOULD
NOT be flowed inside a requester's message.)"
|