Re: Composition model

On Mon, 26 Jul 2004 12:39:25 -0400
Glen Daniels <gdaniels@sonicsoftware.com> wrote:
> Unfortunately, I still think that allowing the "widening" of potential
> values (i.e. feature requiredness and property values/constraints) is a
> dangerous place to go.  Imagine this situation: I define a standard

As I recall, this was discussed at the face to face, where it was resolved
in favor of a single scoping rule, rather than
scope-modified-by-requisite.

An issue fell out of that resolution, that it should be *possible* to
express 'requisite, cannot be made optional', and a proposal was
submitted.  In that discussion, many people expressed antipathy to the
idea (and the proposal's author was, in fact, convinced by these
arguments).

I think that, in general, we get in trouble when we try to head down the
path of *prescribing* behavior, rather than *describing* it.  I would so
characterize the decision made that requires some wsdl:required
meaningless-feature for dispatch.  Granting that there are best current
practices, ruling out other practices is [several seconds of inventive
appalachian obscenity in a form that translates to a single adjective
meaning 'extremely'] bad practice on our part.

We MUST provide a language for service description.  We MAY mention best
practices, but we MUST NOT gut the language in order to prevent something
considered stylistically abusive (even if it's dressed up in
missile-launch terminology, for better scare value).

SO.  I object to reopening this issue, as the scenario offered was also
presented at the face to face, and after, in the discussion of the 'final'
bit.  Lacking new information, new proposals, or at least new arguments, I
believe we are chartered not to reopen the issue.

Amy!
(subtext: amy threatens to reopen several *other* issues.  whoosh!  the
chakram pinballs off seventeen known issues and ...)
-- 
Amelia A. Lewis
Senior Architect
TIBCO/Extensibility, Inc.
alewis@tibco.com

Received on Monday, 26 July 2004 14:40:49 UTC