XML-Data: advantages over DTD syntax?

[Paul Prescod]

> From your mouth to the vendors' ears.  If you ever see an SGML
> editor that supports DTD editing with more than an ASCII editor and
> a help screen saying "Gee, you're on your own here, kid", then let
> me know.  The only DTD editors I know of are specialized tools which
> do not provide any document-instance interface at all.

That's because no user interface designer with a modicum of pride 
would argue that documents and instances *should* use the same graphical
interface. This is in fact a huge step backwards. Editor user interfaces
should be getting more and more customized towards particular SGML 
applications, not more and more generalized, trying to be all things to all
people.

We could encode programming languages and user interfaces in
SGML too, but I'm not going to give up my UI builder or IDE!

> Well, my reason for thinking it makes sense is not any of the the
> above.  I tell people every day that SGML is a good syntax for
> structured information of all kinds.  There are some who seem to do
> this just for laughs -- me, I believe it.  

Do you think SGML is a good syntax for Scheme programs? For C++ programs? 
For regular expressions? For BNF grammars? I don't -- not by a long
shot. Trying to shove everything into SGML may be convenient from a
processing point of view, but it would be a nightmare for those that
must work with the actual textual representation of those formats.

> So it seems natural to me
> to ask why I shouldn't use SGML syntax for structured data about
> document types.  The answers I've heard so far (because 8879 blessed
> this other syntax; because SGML isn't that good after all; because it
> is heresy) are all laughable.

SGML *is* that good after all -- for what it was designed to do. It is NOT
that good for washing windows, fixing cars or describing content models. 

Maybe SGML should be extended so that it is friendly for representing 
arbitrary non-document data, but right now, it is pretty much a crap shoot 
whether a particular non-document domain will find a convenient 
representation in SGML. If you don't intend to edit the data directly,
then it doesn't matter, but if you do, then it certainly does.

> Yes.  And it seems far more likely that we'll get some useful 
> new thinking about document grammar and validation if we do the
> preliminary thinking in a notation that doesn't steer our minds
> straight back to clause 11 of 8879.

That's true. And perhaps once the ideas are hashed out they can be re-expressed
in a format that is type-able (SGML DTD syntax or not). SGML *is* very 
convenient for prototyping formats that you have not written a parser for
yet.

 Paul Prescod


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)