I always though this could be
done on a spec by spec basis. At one point in time
the RM spec didn't even require WSA
at all. And I think right now it really only
requires it for the AcksTo EPR. So,
if the RM spec could find the right wording then
it seems like it would be possible to
say "RM uses the EPR concept from WSA -
however that is expressed or used based
on which version of WSA the RMS and
RMD choose to mutually support".
Some other specs might actually need to go
deeper in specific features of WSA (or
whatever spec) and in those cases a tight
coupling between the specs would make
more sense. In our case though,
I think referencing the EPR concept
w/o tying ourselves to a fixed WSA version
would be safe.
One of the reasons I think this
is important is become some people read the
specs very very literally and if spec
X references a certain version of spec Y
then, in their mind, under no circumstance
should X ever use a different version
of spec Y - period. Perhaps, I'm
being too permissive but this seems awfully
extreme and if possible I'd like to
leave an out for those folks who live and die
by every word in the spec :-)
-Doug
"Green, Alastair J."
<>
09/21/2005 10:28 AM
To
Doug Davis/Raleigh/IBM@IBMUS,
<>
cc
Subject
RE: [ws-rx] NEW ISSUE: Which
version of WS-Addressing spec?
I think that you have to be
very careful about this.
The sole purpose of this spec
is interoperable reliable messaging.
The markitectural half-truth
of “composable specs” cannot be alchemically converted into a whole truth
to be relied upon for the purposes of interoperability.
Any given set of versions of the composable specs may be truly composable
(to ensure this at some basic level is why WS-I exists), but this cannot
be assumed.
In the case of any one spec,
it is the job of the spec to create an island of composability by narrowing
its references to specific version(s). If this is not done then the resulting
spec is open to implementation variation which could be (almost certainly
will be) lethal to interoperation. If there are multiple versions specified
then implementations can state the matrix of interop that they will support.
If there are no specified versions then the statement “implements the
interoperable protocol” becomes close to meaningless (will only be true
by sheer luck).
Alastair
Alastair J. Green
CEO and CTO
Choreology Ltd
68 Lombard Street
London EC3V 9LJ
www.choreology.com
+44 207 868 2316 (office)
+44 870 739 0051 (fax)
+44 795 841 2107 (mobile)
-----Original Message-----
From: Doug Davis [mailto:]
Sent: 21 September 2005 14:21
To:
Subject: Re: [ws-rx] NEW ISSUE: Which version of WS-Addressing spec?
Do we really need to point to a specific version of WS-Addr? I thought
that one
of the composibility aspects of these WS specs was that we wouldn't need
to change
all of them just because one of them was modified. As long as the
basic ideas of
WSA (that RM needs) don't change - like EPRs, MI Headers... can't we just
talk about
them in generic terms and let the implementations decide which version
they want to
support? Should one that (for some reason) can't step up to the latest
WSA be tagged
as non-RM compliant? I don't think it should. It may not be
as interoperable, but that's
a different story.
I've seen other specs go thru quite a lot of pain because different specs
they're
composing with point to different levels of WSA - and they actually try
to support
more than one at the same time - what a hassle.
-Doug
Anish Karmarkar <>
09/20/2005 04:01 PM
To
wsrx <>
cc
Subject
[ws-rx] NEW ISSUE: Which
version of WS-Addressing spec?
Title: Which version of WS-Addressing spec?
Description:
Page 25, lines 664-665 at [1] says:
"WS-ReliableMessaging faults MUST include as the [action] property
the
default fault action URI defined in the version of WS-Addressing used in
the message."
This can be interpreted as any version of WS-Addressing is allowed with
WSRM. WSRM spec should specify which version of WS-Addressing is used by
the spec.
A related issue is that:
On page 25, lines 664-666 talk about the default
"http://schemas.xmlsoap.org/ws/2004/08/addressing/fault" as the
Fault
[action] property. This default is defined only for the SOAP binding and
is meant to be used with WS-Addr faults not WSRM faults.
Justification:
Without clearly indicating which version of WS-Addressing is
required/used by the spec, independent implementation will not
interoperate. WS-Addressing specification has changed substantially (in
certain sections/artifacts of the WS-Addressing spec) over the years.
Target: core
Type: design
Proposal:
Use the CR version of the spec [2] (in this paragraph as well as the
normative reference for the spec) for now and make changes as the
addressing spec transitions through the process of becoming a REC. Based
on the WS-Addr schedule and WSRM schedule, WS-Addr is slated to become
a
REC before WSRM is final.
For the related issue:
change line 664 from --
"WS-ReliableMessaging faults MUST include as the [action] property
the
default fault"
to --
"WS-ReliableMessaging faults MUST include as the [action] property
as
defined by WS-Addressing [ref]."
and
delete lines 665-667
Related issues: none
[1]
http://www.oasis-open.org/apps/org/workgroup/ws-rx/download.php/14548/wsrm-1.1-spec-wd-03.pdf
[2] http://www.w3.org/TR/2005/CR-ws-addr-core-20050817/
