RE: DTD for protocol-03?

	From: Yaron Goland <yarong@microsoft.com>
	
	Let me put this another way.
	
	1) 99% of the people who read our specs do not understand DTDs, they do
	understand BNF.

Oh, you have some survey results to offer perhaps?

	2) The main XML rallying cry was "NO MORE DTDS!!" DTDs are a deprecated
	syntax which is to be replaced by the XML Schema draft.

I'm sorry to be so blunt, but this is total bullshit. I have been
involved in the XML definition effort from the beginning (it's XMl
that has kept me from staying active with this group). And I can say
that this is simply not true. As in false, as in misinformation that
should not be propagated by any responsible person, that is not
supported by the people writing the sdtandard, or the text of that
draft itself.

Let me put this another way: I don't know what evidence you have for
this assertion, but I know that it does not reflect the opinions of
the XML working group or special interest group, nor does is reflect
the draft as written, nor does it reflect any proposed change to the
draft under consideration. It may reflect the opinions of some
individuals working on XML, although this is not a position that is
under consideration at the moment. So even insofar as your erroneous
"history" is considered as a position in the current work, it is a
position without public advocacy.

Stop wasting our time with this. XML allows DTDless processing,
because shippping DTDs for well-known document types is wasteful, and
because not all applications require a markup parser to validate their
input. 

	3) DTDs are NOT required to validate XML syntax. DTDs are not required
	for ANYTHING having to do with XML.

If you had read the standard, you would see, as Terry explained, that
there are two forms of XML: well formed, and valid. As you might guess
from the names (valid XML, which possesses a DTD, and conforms to it)
is the preferred level of conformance for XML. Well-formed XML is a
less-stringent variety for applications that don't need to validate
against a formal model of the document.

Schemas have been proposed as a possible way to supplement DTDs, (or
perhaps generate DTDs) for applications that require _more_ validation
than DTDs can provide. They are currently a proposal by one vendor
that may or may not ever be integrated into XML itself.
	
	So given that we MUST provide a BNF definition and that DTDs are old
	news, I see no reason to waste valuable time trying to maintain two
	separate definitions of the same material. If and when the XML Schema
	group produces a usable finished product I will be the first to support
	its adoption, until then, time is a scarce resource and there is no
	reason to waste it trying to write DTDs which only a tiny fraction of
	our community can use.

I'd like to see support for the assertion that only a tiny fraction of
the community can read DTDs, but since there's no way to prove such
things (unless you do have access to some significant social science
research that you're not telling us), I think that requiring the
reading of less than 20 pages of the XML standard is a small price to
pay for consistency and conformance.
	
	That having been said, if someone would like to volunteer their time to
	write up the DTDs and have them inserted into an Appendix in the draft I
	am sure the authors will be happy to let this person see the draft
	before we submit it as an ID so this volunteer can write up the DTDs and
	we can submit them as part of the draft.

It seems to me that given the current state of the standards process,
not to mention the pressure to use XML, that truth is that the authors
now bear the burden of proof that the XML standard is grossly
inadequate to meet your needs. If that proof, as I suspect, is not
possible then I suggest you sit down, do your homework and stop
throwing red herrings into the discussion!

   -- David
	
------------------------------------------+----------------------------
David Durand                 dgd@cs.bu.edu| david@dynamicDiagrams.com
Boston University Computer Science        | Dynamic Diagrams
http://www.cs.bu.edu/students/grads/dgd/  | http://dynamicDiagrams.com/
                                          | MAPA: mapping for the WWW

Received on Tuesday, 7 October 1997 14:09:48 UTC