Re: Carmelo's Post on Gathering Discretionary Implementation Issues

[David_Marston@lotus.com]

JohnRobert ("atman") was covering the topic of discretionary
implementation,
as granted in the spec, but then wrote:

>All this is proposed simply to start gathering unclear spec ideas easily
and
>systematically when they come up, to encourage the appropriate plethora of
>--or caucophony of--voices,

Note that "unclear spec ideas" (areas?) is something different! The part
that
we are now discussing is where the spec explicitly uses the word "may" and
describes two or more choices from which an implementor may choose. We also
need a way to gather our list of unclear areas, typically where:
A. The spec does not provide sufficient detail
B. Two different parts of the spec interact, but neither part says enough
to
   explain the interaction sufficiently
C. The spec blithely refers to some other spec, but the other doesn't cover
   all XSLT-related situations in sufficient detail
This is another area where we'll do better if we have a quick way to submit
issues as we think of them. However, there will probably be disagreements
about "sufficiency" and whether to bring questions/quibbles to the XSLT
Working Group.

Contrast that to discretionary implementation areas, which are explicity
specified as such and (probably) not subject to disagreement.
.................David Marston