Subject: Re: Alternative *ML syntaxes
From: Erik Naggum <erik@naggum.no>
Date: 19 Nov 2002 00:01:02 +0000
Newsgroups: comp.lang.lisp
Message-ID: <3246652862008190@naggum.no>

* Henrik Motakef
| May I infer that you have a DTD parser lying around somewhere?  Maybe
| even a fully SGML-conformant one?  In Common Lisp?  Care to share? ;-)

  As sharing in any way could be construed as encouragement to use SGML or
  XML, I have to decline.

| However, even if DTDs are still somehow priviledged in XML compared to
| other schema languages (simply by being part of the XML 1.* spec), in
| practice the only thing you really /need/ a DTD for are entity
| declarations (and notations, but they aren't too useful in XML anyway).

  That the DTD is sort of "implied" by the application or the instance in
  no way negates it.  Although some elements may have sub-element contents
  according to their own position in the hierarchy, which could be hard to
  express in a "standard" DTD, sender and receiver even of XML-encoded data
  need to agree on /something/.  The situation may not actually be better
  than it was absent the stupid tags, but the only thing you have on an XML
  without a pre-agreed structure is attributes and sub-elements and data.

| What do you do with XML applications that lack a DTD, most prominent
| example being XSLT? Do you simply ignore them, or do you construct a DTD
| from what you get?

  I try to ignore XSLT.  Most of the time, document instances have enough
  information to construct useful content models and keep track of which
  things are attributes.  It is not particularly hard.  The biggest problem
  is in fact that some users actually need the retarded syntax.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.