TEXT/HTML wide open?

Paul Prescod (papresco@calum.csclub.uwaterloo.ca)
Wed, 6 Mar 1996 10:44:42 -0500


Date: Wed, 6 Mar 1996 10:44:42 -0500
Message-Id: <199603061544.KAA24831@undergrad.math.uwaterloo.ca>
To: "Daniel W. Connolly" <connolly@beach.w3.org>
From: Paul Prescod <papresco@calum.csclub.uwaterloo.ca>
Subject: TEXT/HTML wide open?
Cc: Paul Prescod <papresco@calum.csclub.uwaterloo.ca>, html-wg@w3.org,

>>Fine, but if the text/html document type
>
>Ummm... terminology mismatch. are you talking about the text/html
>internet media type, or one of the HTML document types?

The proposed media type.

>> has no real teeth to enforce
>>HTMLish-ness, why not remove the requirement?
>
>The teeth to enforce HTMLish-ness comes from the installed base of
>clients, which only moves so fast.

Fine.  But why have it in the standard too?  I think a typical reader of
your proposed standard would come to the (erroneous) conclusion that
text/html conformant documents will have certain particular elements (esp.
<TITLE>!).  I would suggest that you explicitly state that 

 * text/html is wide open
 * the only tags users can depend on are those in the standardized tag sets
that HTML compliant user agents are required to support

and dump the teethless "includes" requirement on user DTDs.

"Rather than binding text/html to any particular DTD, we define it to be any
SGML document type whose DTD is externally referenced by URL; i.e. the body
begins with

<!doctype ... public "..." system "...">
"

"Where elements in a text/html document correspond with element names in the
standardized HTML DTDs, user agents should interpret them as having the same
semantics."

I feel that this is a little bit more clear, flexible and straightforward.

 Paul Prescod