Lists Home |
Date Index |
- From: Paul Prescod <firstname.lastname@example.org>
- To: email@example.com
- Date: Fri, 03 Oct 1997 22:09:34 -0400
Sean Mc Grath wrote:
> Yes it reduces the number of parsers "a good thing (tm)" but isn't the big
> win of the XML-Data approach is the extra miles per gallon that
> accrue to all the XML application software?
To all of the XML application software that cannot parse DTDs. It still
comes back to saving software developers from processing two formats.
> All of a sudden it becomes possible to typeset documentation
> of schemata by processing them with straight XML typesetting tools.
You can do this with SGML. Check out Earl Hood's software. Check out
Near&Far. Yes, this is *easier* with XML-Data, because you only need one
parser instead of two. Nobody has ever disputed that.
> It becomes possible to load schemata into XML databases as first class
> citizens. XML greppers can grep 'em. XML web harversters can harvest
You can do this with SGML documents. SGML DTDs are represented in a
database by a specialized form of grove created by the parser and grove
builer. XML-DATA merely adds a level of processing. The XML-Parser must
build an XML grove which is then transformed into the specialized grove.
> It becomes possible to contemplate a schema derivation mechanism
> based on using XLL to "cherry-pick" from a collection of existing schemata.
There are proposals for SGML schema derivations. The question is only
whether they use XLL or not. This is fundamentally the same question as
whether XML should use instance syntax.
> It becomes possible to contemplate using an XML to XML transformation
> system to transform schemata and then auto-generate the instance
I don't know what you mean here, but I confidently predict it can be
done with SGML DTDs too. If you use a decent parser like SP, the grove
that is built provides access to all of the information in the DTD, in a
grove that is optimized for DTD navigation (and thus transformation).
> For me the big win is the simplification it could bring to base XML
> application development and the sheer intellectual appeal of it.
> It is a very computer science-ish, Lisp-ish, Dame Ada Lovelace-type,
> KISS way of looking at things. A grand unifying theory of a sort.
I share this interest in unifying concepts. I've done instance syntaxes
for DTDs myself in the past. I have no problem with the concept it's an
obvious one...I just don't think it needs to be, nor should be, the
*standard* way for creating DTDs. Users should have the option of the
> The XML world would then have the freedom to create
> new and better syntaxes for schemata safe in the knowledge that
> todays tools that can process XML will process these new syntactic
> sugars via transformation to base XML - the mother of all syntaxes.
Sean, we have always had this freedom. Dozens of different projects have
taken advantage of it in the past. You are re-describing the current
situation in SGML. We can experiment with instance syntax and convert to
DTD syntax for compatibility. Making such a converter is about a day's
xml-dev: A list for W3C XML Developers. To post, mailto:firstname.lastname@example.org
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:email@example.com the following message;
To subscribe to the digests, mailto:firstname.lastname@example.org the following message;
List coordinator, Henry Rzepa (mailto:email@example.com)