Re: The Final Word On HTML (fwd)

marc@ckm.ucsf.edu
Thu, 26 Sep 1996 22:19:26 -0700


From: marc@ckm.ucsf.edu
Date: Thu, 26 Sep 1996 22:19:26 -0700
Message-Id: <199609270519.WAA05750@pele.ckm.ucsf.edu>
To: www-html@w3.org
Subject: Re: The Final Word On HTML (fwd)

MACRIDES@sci.wfbr.edu wrote:
|MegaZone <megazone@livingston.com> wrote:
|>Once upon a time Carl Morris shaped the electrons to say...
|>><OBJECT DATA="hello.png" TYPE="image/png">
|>> <OBJECT DATA="hello.gif" TYPE="image/gif">
|>> whoops no variant
|>></OBJECT>
|>></OBJECT>

Let the protocol do the work it is supposed to do.

The server sends out:

<HTML>
...
<OBJECT DATA="hello" TYPE="video/*">
<OBJECT DATA="hello" TYPE="image/*">
<OBJECT DATA="hello" TYPE="text/*">
goodbye
</OBJECT>
</OBJECT>
</OBJECT>
...
</HTML>

Which says: negotiate me a video, if you can't do that try an still image, 
OK I'll take whatever text we can settle on before I bail.

"Inserting objects into HTML" [1] doesn't address using wildcard minor IMT's 
in the TYPE attribute.  HTTP uses this convention in its Accept header, so it 
seems reasonable that HTML inherit it.

The proper variant, if any, can be negotiated instead of building every
contingency of an external object into each HTML document that references 
it.

-marc
[1] http://www.w3.org/pub/WWW/TR/WD-object