Re: Proposed MEP TF recommendations

One immediate response.

On Tue, 22 Jul 2003 15:53:00 -0400
David Booth <dbooth@w3.org> wrote:
> 5.  Fault rules.  Simplify to "message triggers fault", and further
> specify that the fault is sent to the client whose message triggered
> the fault.  For example, in a pattern such as p2d that involves two
> clients A and B, if client A sends an initial request message that
> causes fault F to be generated, then fault F is sent to client A, not
> client B.

I don't agree at all.  I do not think such a suggestion should be part
of the task force recommendations, partly because I do not agree, and
partly because what Umit has recommended, in a similar vein, is entirely
different (she has proposed "message triggers fault" but with the
message always going to the designated recipient, so the fault in the
example above would go to B, not A).

These slight differences are the kind of thing that make me really want
to keep the ability to define different fault generation rulesets.  I
think that when people suggest winnowing them all out, it's because one
set always works for their environment ... forgetting that other
environments may have different requirements (where possibly one ruleset
*also* works ... but it's a different ruleset).

Amy!
-- 
Amelia A. Lewis
Architect, TIBCO/Extensibility, Inc.
alewis@tibco.com

Received on Tuesday, 22 July 2003 16:29:40 UTC