- From: Jay Bazuzi <jaybaz@microsoft.com>
- Date: Sun, 21 Jul 1996 00:28:31 -0700
- To: "'www-html@w3.org'" <www-html@w3.org>
>From: Marc Salomon[SMTP:marc@ckm.ucsf.edu]
>
>What about doing it in <SPAN CLASS="ACRONYM">HTML</SPAN> this way, <SPAN
>CLASS="PAUTHOR">Dave Raggett</SPAN> of the <SPAN CLASS="IAUTHOR">W3C<SPAN>?
>
>Fight Tagitis!
>
This is an interesting idea that has show up a couple times here lately.
To generalize your suggestion, what you seem to be proposing is that
the tag name becomes an attribute of some generic, non-meaningful tag.
Consider, for example, the following DTD snippet:
<!ELEMENT TAG - - (#PCDATA | TAG)* >
<!ATTLIST TAG
ELEMNAME NAME #REQUIRED -- Element Name --
>
While we're at it, why not make the content of an element be an
attribute:
<!ELEMENT TAG - - (TAG*) >
<!ATTLIST TAG
ELEMENT NAME #REQUIRED -- Element Name --
CONTENT PCDATA #REQUIRED -- Content --
>
I didn't come up with this all by myself, of course! I the idea in the
first chapter of _The SGML Handbook_, an early paper by Goldfarb about a
structured markup languages. He calls the ELEMENT attribute 'GI' and
suggests that the 'GI=' part be implied by its placement next to STAGO
('<'), and that the 'CONTENT=' part be implied by its placement between
the open and closing tags of the element.
It seems absurd to me to embed a generic coding language within SGML.
SGML is there and waiting for us. It's well documented, well thought
out, and it's even an International Standard. We're so close here to
the Right Solution*, but it'll be a while before we get there. Grumble.
-Jay
*I mean a system to handle any DTD and a powerful stylesheet mechanism
like DSSSL.
------------------------------
jbazuzi@microsoft.com I just work here.
Received on Sunday, 21 July 1996 03:30:18 UTC