Re: discussion on Extension text of SpecLite

At 01:22 PM 4/21/2004 +0200, Dominique Hazaël-Massieux wrote:
>Le mar 20/04/2004 à 17:15, Lynne Rosenthal a écrit :
> > 1.  Designing extensibility into your specification - can this be done for
> > all specifications

Yes, it can.  But should it?  (No, IMO.)

>or are there some technologies and areas (e.g., I18N,
> > Accessibility) that don't apply?

I think there will be standards where extensibility should not be designed 
in.  "Don't apply" might be one reason.  "Violates interoperability 
requirements" could be another.  There may be others.

>[...]
> > 2. The complexity inherent to DoV interaction is not presented.  Should
> > there be a discussion and/or requirement about the interrelationship
> > between extensions and other DoV?  Can this be accomplished with through
> > examples and scenarios?
>
>I would love seeing this explained through examples ;

+1

>the way I envision
>it would be a general statement about how extensions increase
>variability between implementations, and then one or two well-spotted
>examples illustrating how this can affect interoperability, and how this
>can be resolved ; it's unlikely I'll have time to come up with these
>examples before Monday, but I'll try to get something in the wiki then.
>
> > 3.  It may be the case that the second area, (extensions affect
> > conformance) is not exclusive to extensions, applying to any optional
> > behavior.   Are there things in this section that only apply to extensions?
> >
>I think so ; as David stated Monday:
>* not contradicting the core specification is something specific to the
>extensibility mechanism
>* the notion of strict conformance probably belongs to this area, too,
>since that's pretty specific to extensions IIRC.

IMO, Extensibility should remain a separate topic.  Certainly for 
FPWD.  But I haven't yet heard anything to convince me that we should put 
aside our original reasons for special attention to Extensibility.

Regards,
-Lofton.

Received on Saturday, 24 April 2004 16:36:09 UTC