Re: A proposal for addition to HTML 3.0: Frames

From: Joe English <>
Date: Tue, 19 Sep 95 23:06:05 EDT

Marc Salomon <> wrote:

> Christian Mogens writes:
> |Still, I can't help thinking that FRAMEs are really DIV and/or
> |FIG with some different syntax and added semantics.
> Not to mention nifty processing.  HTML 3 should be able to deal with it like:
> <FIG REL=EMBED HREF=&uri; > 
> </DIV>
> With the behavior of CLASS=FRAME defined with the style information.  Perhaps
> the W3C could maintain a library of definitions of well-known base styles, su
> as FRAME, which could be implemented to a base level by the UA vendor and 
> customized using the style mechanism.

I disagree.  It isn't feasible to push *everything* off into stylesheets,
and this feature is one of the things that belongs in the document

It should be safe for browsers to completely ignore
a stylesheet, replace the author's stylesheet with
the reader's, or combine them in some fashion.  
A FRAMES compound document is intimately linked with the
frame specification; putting one part in content and
the other in a stylesheet would have terribly unpredictable

Plus, the stylesheet syntax(es) are still far from complete,
whereas SGML can provide a stable, rigorous definition
of the markup.  (I have it on good authority that a
formal DTD fragment *is* forthcoming.)

Lastly, I don't think that <DIV CLASS=FRAME> is an appropriate 
use of the <DIV> element or the CLASS attribute: the namespace
of CLASS attribute values should belong to the document author,
and I'd hate to see <DIV> overloaded with so many functions that
it becomes another <INPUT>.

--Joe English

P.S.  I don't think that this:

> Makes you wonder what they're thinking down in Mtn. View:
> 1.  Alleged adherence to standards process is only for marketing purposes.
> 2.  Send marketing guy to standards process with oblique reference to 
>     published fait accompli.

is an accurate (or appropriate) assessment of the situation.
Netscape is doing *exactly* the right thing here: they've 
published a draft specification for review before releasing
anything to the public at large, and (will have) provided a 
sample implementation for evaluation.  The sample implementation 
is/will be a *beta* release, so there will be plenty of time for 
corrections to the spec before it's finalized; and on the other
hand, Netscape's release schedule for 2.0 will ensure that the
review process goes quickly (certainly *noone* can complain 
about *that*.)