Re: Summary: 22-24 Sept 2003 WS Desc FTF

ARGH! So all those hours we argued about 6 vs. 4 were because of
a random pattern???!!! Hmm, gotta think again about the no-shoot
rule. ;-)

OK, so where does this leave us? Back to 4 patterns, but more
clarified? Boy, that was a slightly long route to clarifying
what those patterns meant.

Welcome to standards ;-).

Sanjiva.

----- Original Message -----
From: "Roberto Chinnici" <Roberto.Chinnici@Sun.COM>
To: "Amelia A. Lewis" <alewis@tibco.com>
Cc: "Sanjiva Weerawarana" <sanjiva@watson.ibm.com>; <jmarsh@microsoft.com>;
<www-ws-desc@w3.org>
Sent: Wednesday, October 01, 2003 12:20 AM
Subject: Re: Summary: 22-24 Sept 2003 WS Desc FTF


> Now that you mention it, I think that I proposed in-multi-out at the
> Scottsdale f2f. It was simply meant to be an example of a pattern other
> than the usual one-in/one-out ones. I can vaguely imagine some use for
> it, like a stock ticker service (don't shoot me, please). Hopefully it
> served some purpose in helping define the pattern framework, but I don't
> think it (or the other "multi") deserve a place in the WSDL 1.2 spec.
>
> Roberto
>
>
> Amelia A. Lewis wrote:
> > *shrug*
> >
> > Probably.  But outbound-first is unrelated to the "multis".  The
> > "multis" are in-multi-out, assuming only two participants, with the
> > service streaming messages until it (somehow) decides to stop, and
> > out-multi-in, assuming only two participants, with the
> > [not-acting-as-a-service-participant] replying to a single question from
> > the service with a barrage of answers, one after the other.
> >
> > Whatever those were intended, they were not proposed or advocated by
> > TIBCO, and I can't defend them because I don't understand them, and
> > don't see any use in them.
> >
> > I wasn't asked to supply justification for the outbound-first
> > operations, which TIBCO *has* advocated, strongly.  The disappearance of
> > the old multicast solicit response is a result of changes in the
> > definitions, such that the outbound-first operations are now,
> > theoretically (apart from the trifling problem that fault replaces
> > message is probably inappropriate), modeled by existing outbound-first
> > patterns.
> >
> > Amy!
> > On Tue, 30 Sep 2003 19:56:40 +0600
> > Sanjiva Weerawarana <sanjiva@watson.ibm.com> wrote:
> >
> >
> >>"Amelia A. Lewis" <alewis@tibco.com> writes:
> >>
> >>>The "multis" are not multicast-related, and I was never a proponent
> >>>of them.  I do not, in fact, understand what networking paradigm
> >>>they are thought to embody, or who has advocated them.
> >>>
> >>>The "multis" appear to be serial unicast: a trigger message starts a
> >>>flow of messages from some other participant, which eventually
> >>>stops.  I feel certain that someone has a reason for proposing such
> >>>patterns, but it wasn't me, and I don't know what the reason was or
> >>>is.
> >>
> >>I'm confused Amy .. I recall that Tibco and MSFT had different
> >>interpretations of the old outbound operations and I had always
> >>thought that that difference was recognized by these two patterns.
> >>
> >>Is that not the case? Is there another pattern we should be
> >>including that has a single outbound / single inbound combination
> >>yet does something different that we should be including?
> >>
> >>Sanjiva.

Received on Tuesday, 30 September 2003 14:29:31 UTC