Re: 8879 changes?
Jon Bosak wrote:
> There is general consensus on the changes that we actually need for
> XML (for example, allowing overlapping enumerated attribute values, so
> that you can use "YES" as a value for more than one enumerated
> attribute, and allowing attributes on an element type to be specified
> in more than one place). The uncontroversial nature of these changes
> was what emboldened us to assume them for purposes of the latest
> draft. WG8 may see the corrigendum as an opportunity to make
> additional changes beyond those needed for XML if general agreement
> can be found for them; it's this part that's still undecided. So
> there won't be an actual TC for you to look at until after the meeting
> in Barcelona. Charles Goldfarb has been carrying the ball on this and
> doing a mighty fine job of it.
If the parts are so uncontraversial, there should be no reason to not
have a basic draft for Barcelona. Charles seems to have done all the
legwork. If there are things that are contraversial, then they should
belong to the general 8879 "review" system until they are popularly
Jon, as far as you can see now, there are no things in XML-link or the
XML-stylesheet proposals that may alter any TC requirements?
The TC mechanism may be suited for day-to-day SGML improvements that
are needed to bring it into line with current practise or needs or
to align with other standards, as far as I see. XML is clearly, even
as we speak, an example of current practise.
In other words, I think TCs should be short and highly targetted, and
reflecting actual user needs, not because of kitchen-sink-ism. And I
think SGML should keep in complete sync with XML. So I prefer a TC
rather than later, and simpler rather than 'complete'.
(delegate for Standards Australia, WG8)