- From: John Whelan <whelan@itp.unibe.ch>
- Date: Tue, 12 Oct 1999 12:58:32 -0400 (EDT)
- To: www-html@w3.org
> - You cannot parse general SGML based document, unless you have its > DTD. > For example, the simplest case, you have no idea which elements can > be included inside others, and thus you have no way of knowing > whether "<x>foo<y>bar" means really "<x>foo<y>bar</y></x>" or > "<x>foo</x><y>bar</y>". Or, for that matter, <x>foo<y/>bar</x> or <x/>foo<y>bar</y> etc., to apply the corresponding XML notation. Having fed an HTML4.0 document to an HTML3.2 parser and gotten back something like <table> <col> <col> <col> <tbody> <tr> <td> </td> </tr> </tbody> </col> </col> </col> </table> I can say that this is sometimes an issue. (Yes, the parser *should* read the DTD, but moving to XML seems simpler to me than getting lowbrow HTML applications to do the extra work required to get SGML to work smoothly.) John T. Whelan whelan@iname.com http://www.slack.net/~whelan/
Received on Tuesday, 12 October 1999 13:01:38 UTC