ws-rx-editors message
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]
Subject: Need help on i093
- From: "Gilbert Pilz" <Gilbert.Pilz@bea.com>
- To: <ws-rx-editors@lists.oasis-open.org>
- Date: Mon, 8 May 2006 16:16:09 -0700
My
fellow simians,
I
need some help on i093. After talking things over with Doug Bunting I think
I have a pretty good understanding of the issue. Here's my sense of
it:
Imagine an English sentence with [subect], [verb], [object]. Doug
Bunting is saying that, if that sentence carries RFC2119 semantics, the
[subject] cannot be a thing like an XML element or a part of some document,
etc. but instead should be some kind of software component that someone
implements. For example it is incorrect to say "The
<foo:baz> element MUST contain a <foo:splat> attribute . . .". It
is correct to say "Foo Service implementations that send
<foo:baz> messages MUST include <foo:splat> as an attribute of the
<foo:baz> element . . ."
Unfortunately most of the WS-RM document (haven't looked as WS-RM Policy)
uses phrasing like "This REQUIRED element MUST . . .". I've fixed the examples
that Doug Bunting called out in the issue description and cleaned up all of
SequenceAcknowledgement. However that still leaves CreateSequence,
CreateSequenceResponse, CloseSequence, CloseSequenceResponse, TerminateSequence,
TerminateSequenceResponse, Sequence, and AckRequested left to clean
up.
I'm
a bit busy right now and I need some help in order to turn this around in a
timely fashion. If everyone could do a couple of sections it shouldn't take too
much time. You can cut and paste from my phrasing or (obviously) concoct your
own. Thanks (in advance) for any help.
-
gp
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]