Re: Digressions (was: Footnotes)

Daniel W. Connolly (connolly@beach.w3.org)
Wed, 01 May 1996 00:41:19 -0400


Message-Id: <m0uETj5-0002TqC@beach.w3.org>
To: murray@spyglass.com (Murray Altheim)
Cc: Francois Stragier <stragier@email.enst.fr>, www-html@w3.org
Subject: Re: Digressions (was: Footnotes) 
In-Reply-To: Your message of "Mon, 29 Apr 1996 10:53:23 CDT."
             <v02110100adaa925d026e@[140.186.34.50]> 
Date: Wed, 01 May 1996 00:41:19 -0400
From: "Daniel W. Connolly" <connolly@beach.w3.org>

In message <v02110100adaa925d026e@[140.186.34.50]>, Murray Altheim writes:
 On talking
>with some folks here at Spyglass, we use "NOTATION", as the more formal
>SGML term already in use in other places. So, as an example with CODE and
>PRE:
>
>    <PRE NOTATION="text/pascal">
>      NumToStr(paramPtr, theID, IDStr);
>      NumToStr(paramPtr, SizeResource(theResource), sizeStr);
>      BlockMove(@theType, POINTER(ORD4(@typeStr)+1), 4);
>    </PRE>

We considered that, but text/pascall _isn't_ a notation in the sense
of <!notation ...>. And making it a notation would mean changing the
DTD every time a new data format is added, breaking down independence
between HTML and stylesheets (and graphic formats and...).

That, or supporting <!notation ...> declarations on a per-document
basis. But it seemed silly to have the notation attribute be a
pointer to a <!notation ...> declaration at the top of a document,
which was just a pointer to a MIME type, ala:

	<!notation css system "text/css">

when <style type="text/css"> accomplishes the same thing.

So we chose type, as a mnemonic short-hand for content-type or
media-type, in the RFC1522 (MIME) sense.

Dan