Re: Extensions

On Fri, 20 Sep 2002, Lofton Henderson wrote:

> Major disagreement here.  As I understand, you're saying that it is
> not necessary for SpecGL to enumerate or state checkpoints on any of
> the individual dimensions

It is not necessary for SpecGL to enumerate or state checkpoints on
any of the individual dimensions iff those checkpoints are sub-cases
of other, more general, checkpoints.

> First, a general observation.  In SpecGL we are addressing (a
> half-dozen or so) specific "dimensions", such as Extensions, because
> in the collective experience of QAWG members, these have repeatedly
> been associated with serious interoperability problems.

Yes, I understand that. I believe that Checkpoint 9.5 above does not,
in general, solve the Extension problem you are trying to solve (while
requiring a lot of overhead to comply). Registering an extension may
or may not solve interoperability problem, depending on the spec and
on the registry. We need to look deeper.

[ lots of good stuff snipped for the above reason ]

> >By its very definition, extension is something
> >that can be added without being known to others and still work. That's
> >what important.
>
> Bingo!  You and I apparently have different concepts about what extensions
> are.  In the examples that I'm quoting, if the extension is not known to
> others, in fact the content does NOT work.

SpecGL should define extension before introducing checkpoints that
cover them. For example, we could agree that "extension" is something
that can be added without others knowing anything about it, except for
the fact that it is an extension. Then we could proceed with defining
a "good extension mechanism".

What SpecGL is doing now is providing an implicit instruction for
authors how to make some extension work: "register your extensions and
you are covered!". What SpecGL should be doing is requiring that
extension mechanisms maintain interoperability among parties that MAY
NOT be aware of each extension definition. SpecGL should address the
_core_ of the problem instead of demanding a specific solution to one
of its known side-effects.

I am skipping lots of smaller problems with "extension registries" to
keep the thread shorter. SpecGL already tries to define a "good
registry", but there are many things the current attempt does not
cover. My point is that SpecGL should not digress to a specific and
imperfect solution; it should address the core problem
(interoperability in the presence of extensions).

Alex.

-- 
                            | HTTP performance - Web Polygraph benchmark
www.measurement-factory.com | HTTP compliance+ - Co-Advisor test suite
                            | all of the above - PolyBox appliance

Received on Friday, 20 September 2002 16:09:57 UTC