This echos a thread that died out about two weeks ago. The
gist of the argument then was similar to yours: We have people
running off half-cocked after multiple, often competing, specs
(e.g., schemas) - and talking about core portions of the 1.0
spec as if they were "legacy" items. Specs are flying right and
left, timely ideas are being ignored, and tempers are flaring.
Nobody can even agree on what an FPI is or what it's really do-
ing in the XML spec.
Although the discussions have quieted down, and people are talking
more about compatibility of formats and goals, the damage is al-
ready too obvious to be ignored: XML is starting to be seen as
amorphous and unstable.
Exacerbating the problem is the distance of the W3C from some
important thought currents. Namespaces are the obvious example.
We had any number of thoughtful, interesting postings about how
to make namespaces truly useful (and how to integrate them fully
with validation, as it is defined in the 1.0 spec).
Very few productive responses were garnered from people whose
vote counts. And many people were left with the feeling that the
1.0 validation/schema mechanism is going to be a legacy item be-
fore it's ever even used.
Once that feeling settled in, we again started to see the whole
SGML-compatibility question re-opened. And we started seeing
the sorts of gripes that reveal widespread ambivalence and con-
fusion about what XML is really supposed to do.
Worse yet, it is becoming clearer that the initial goals set out
in the XML 1.0 spec are falling by the wayside. This can't help
but give the rest of us a profound sense of uncertainty. Some
random observations along these lines:
3) XML shall be compatible with SGML
(the entire schema mechanism is being debated, and name-
spaces, despite protests to the contrary, weaken and com-
plicate use of DTDs)
4) It shall be easy to write programs which process XML documents
(with every new spec, this prospect becomes more remote)
8) The design of XML shall be formal and concise
(ditto)
10) Terseness in XML markup is of minimal importance
(this position is reversed with namespace defaulting)
I've always had confidence that there is enough brainpower here
to satisfy most of the core constituencies. And especially since
late September we're seeing more of an effort to do just this.
My guess is that if the standards bodies don't make any bonehead
moves, and if everyone will make stability and smooth transitions
into prime goals, that everything will turn out okay. It will just
take longer than predicted.
Richard Goerwitz