This WS-I consortium entrusts itself with
"defining" interoperability. The founders are permanent board
members and board minutes are never public (even to the contributing
members). Later some of the same founding companies that sit on the
"interoperability" board author a set of standards (the WS-X series) and
offer them up to their local international standards body, OASIS. Some
of the proposals are unique or complimentary to existing work and some are
overlapping. I think it's safe to say that some of the aspects of the
proposals are causing fear, uncertainty, and doubt, in the industry (FUD)
and that the current situation would seem to be at odds with the goal of
WS-I, the interoperability consortium set up to alleviate confusion in the
first place
(remember the noble cause).
I don't know much of the working of WS-I but I do know that whenever
we've approached them about transactions the response has been "interesting,
but too much other stuff to do at the moment". For something like
transactions, especially when you haven't got any spec. yourself, I'd suspect
that waiting for WS-I to be ready isn't on the cards.
[Mark
Potts]
The
triplicate of Security, Reliable Messaging and Transactions were noted as
likely candidates after the basic profile by WS-I, however there are likely to
be other topics competing, and considering there is no real "consensus"
besides Security I think Reliability and Transactions will be the latter
topics for profiling.
Now OASIS is in the painful
position of having to take action and account for the interoperability
consortium's influence - a consortium made up of some of the biggest
players in the industry who are also involved with these standards
organizations at some level.
It's my belief that the best way to deal with this situation is for
the standards body to address proposals in light of existing working groups
and this is the job of the standards body leadership. OASIS has been
fairly silent up to now, I'm sure, because they are examining their options
carefully as we are about to set some precedence (not unlike patent/RAND
issues).
I don't think there's anything to be gained by having multiple
specifications on the same topic covered by different standards bodies. But
then that's a personal opinion.
[Mark Potts]
Agreed our goal should be to influence the placement of SW_coordination
and WS_Transaction to OASIS as a was done with WS_Security such
that overlapping standards can be incorporated.
I hope that the standards bodies find the
strength to drive the standards process as best meets the needs of their
entire constituency. In this case, that means examining each proposal
in depth and taking a clear, well communicated, course of action that
eliminates redundancy and confusion. If the standards body cannot
accept new proposals based on existing TC work, then the standards loose
their effectiveness in defining what is "standard in the first place".
Once clear set of standards is much more effective than many ambiguous
and overlapping standards. Situations like this easily cost the
industry millions of dollars in wasted effort.
WS-Transaction - This proposal is in essence a
simplified non-normative version of BTP (with exception of some
material).
Without going into too much detail, WS-T
isn't as complete or as polished as it should be. It's more a marketechture
statement than a specification like BTP.
If the redundant parts of the
proposal have capabilities that do not exist in the existing standards
proposals then the submitters should join the affected WG and modify the
existing standard (that happens to already be in public review after a long
processing cycle in TC). In essence a new version can be hatched based
on the merits of the proposal.
Yes and no. One of the things that instigated WSCF (and HPs initial
submission to OASIS) was that a single transaction protocol may not suit
everyone and neither does a single coordination protocol. Rather than create a
bloated specification that has lots of caveats and optional features and is
horribly complex to read and implement, why not have micro-protocols that a
geared just for a single (say) domain? It's like the micro-kernel approach to
operating systems. So, if there are things in WS-T that are useful to some
businesses *unless* they are useful to all, I'd prefer to see a separate
transaction protocol. Maybe BTP should be the Business Transaction
Protocols?
[Mark Potts]
Hmmmmm.... not sure about that but its a long
conversation!
WS-Coordination - This proposal is unique
and stands on its own as a way of setting up "distributed SOAP transaction
servers" - aka a VAN.
Agreed.
BPEL4WS - This proposal is
redundant but has the distinct benefit of merging XLANG and WSFL, two of the
three major business process languages (BPML is one other I can think of
easily). I'd be happy to get rid of three standards in exchange for
two any day.
I think the
implication here is that if OASIS does not adopt the proposals future
submissions will go to W3C (or yet another new body)
I don't know about that. I don't know what the current agenda is, but I
can say that originally we definitely intended going to a standards body and
taking *all* that that implied, i.e., that what goes in may not be what comes
out (just look at BTP).
and we still don't have clarity.
What this assumption does not take into account is that W3C and OASIS
*are* fundamentally interested in interoperability and that they may not be
willing to compromise their integrity by "clouding the water" even
more.
Personally I think there's a lot of confusion in general about OASIS and
W3C!
A little marketing for
OASIS/W3C/SOAP Builder interoperability efforts might be in
order.
Regards and let's hope for the best.
Cheers,
Mark.
----------------------------------------------
Dr. Mark Little,
Distinguished Engineer,
Transactions Architect, HP Arjuna Labs
Email:
mark_little@hp.comPhone: +44 191
2606216
Fax : +44 191 2606250
Bill Flood
Sybase
| Mark Little
<mark_little@hp.com>
08/10/2002 02:00 AM
|
To: Mark Potts
<mark.potts@talkingblocks.com>, "Cho, Pyounguk"
<pyounguk.cho@iona.com>, Peter Furniss
<peter.furniss@choreology.com>, BT - spec
<bt-spec@lists.oasis-open.org> cc:
Subject: Re: [bt-spec]
WS-transaction..
|
Mark, I agree that a teleconference is in order. Middle of next
week would
be better for me (and either before 5pm our time or after 7pm
would be even
better, given that I'll be juggling a baby at the same time
:-)
No revisionist history here: WSCF was always intended to support
BTP as just
another extended transaction model. So, if my colleagues in
IBM continue in
that light, it should be possible for us to co-exist. We
will definitely
have to see what happens when this hits a standards body
though.
Mark.
----------------------------------------------
Dr.
Mark Little, Distinguished Engineer,
Transactions Architect, HP Arjuna
Labs
Email: mark_little@hp.com
Phone: +44 191 2606216
Fax :
+44 191 2606250
----- Original Message -----
From: "Mark
Potts" <mark.potts@talkingblocks.com>
To: "Mark Little"
<mark_little@hp.com>; "Cho,
Pyounguk"
<pyounguk.cho@iona.com>; "Peter Furniss"
<peter.furniss@choreology.com>;
"BT - spec"
<bt-spec@lists.oasis-open.org>
Sent: Friday, August 09, 2002 7:56
PM
Subject: RE: [bt-spec] WS-transaction..
Hey before we all
get fired up ( just like the ebXML folks did with us
originally ) lets
take a step back and look at this rationally....
This is just my
perspective, and yes I did know this was coming somewhat,
but there is
little difference between this spec and BTP -
WS-Coordination is
separated from the transactions types (WS-Transaction )
such that
WS-Coordination can coordinate different transactions models, BTP
simply
declares the transaction model being used ( atom or cohesion ) within
the
context. With some changes ( minor ) on either side BTP can fit in
easily
to WS-Coordination and therefore BPEL. Of course there are
other
differences but on the whole anyone involved in BTP should
recognise a lot
in the specs.
We (as BTP TC) should be looking to
this as an opportunity for "convergence"
where we can have IBM,
Microsoft, BEA, Oracle and some of us smaller minows
all back one
standard. There is no definition of where this work will end up
as a
standard (as yet) but again we should encourage in whatever way
we
can, those involved to take this to OASIS where we can help get
WS-
Coordination and WS-Transaction "fully baked" and leverage the
work
completed in the BTP TC.
I suspect the TC needs to get
together on a call so at least we have a
common statement and position on
this we can discuss with customers and our
companies.
Regards
Mark.
> -----Original Message-----
> From: Mark Little
[mailto:mark_little@hp.com]
> Sent: Friday, August 09, 2002 11:20
AM
> To: Cho, Pyounguk; alex@ceponkus.org; Mark Potts; Peter Furniss;
BT -
> spec
> Subject: Re: [bt-spec]
WS-transaction..
>
>
> FYI this actually started out as
the Web Services
> Coordination Framework
> with HP, IBM and
Iona but due to some politics which I don't
> even want to
>
understand, the participant list changed. (A very long and
> not so
happy
> story!)
>
> Mark.
>
> ----- Original
Message -----
> From: "Cho, Pyounguk"
<pyounguk.cho@iona.com>
> To: <alex@ceponkus.org>; "Mark
Potts"
> <mark.potts@talkingblocks.com>; "Peter
> Furniss"
<peter.furniss@choreology.com>; "BT - spec"
>
<bt-spec@lists.oasis-open.org>
> Sent: Friday, August 09, 2002
6:08 PM
> Subject: [bt-spec] WS-transaction..
>
>
>
Hello BTPers,
> Please, check out the following link.
Finally, IBM and MS
> together with
> BEA have come up with
ws-transaction spec.
>
http://zdnet.com.com/2100-1106-949049.html
>
> Regards,
>
Pyounguk
>
>
----------------------------------------------------------------
> To
subscribe or unsubscribe from this elist use the subscription
>
manager:
<http://lists.oasis-open.org/ob/adm.pl>
>
>
>
----------------------------------------------------------------
To
subscribe or unsubscribe from this elist use the subscription
manager:
<http://lists.oasis-open.org/ob/adm.pl>