Re: Frames - does anyone like them?

Sunil Mishra (smishra@cc.gatech.edu)
Wed, 21 Aug 1996 13:36:04 -0400 (EDT)


Date: Wed, 21 Aug 1996 13:36:04 -0400 (EDT)
Message-Id: <199608211736.NAA05167@cleon.cc.gatech.edu>
From: Sunil Mishra <smishra@cc.gatech.edu>
To: www-html@w3.org
In-reply-to: <Pine.LNX.3.95.960821193428.405C-100000@trillian.hol.gr> (message
Subject: Re: Frames - does anyone like them?

\\ The way I think this would best be implemented is using multiple BODY
\\ elements in combination with stylesheets, who already seperate the canvas
\\ (referred to using the HTML element) and the actual document (referred to
\\ using the BODY element). Multiple bodies could be given float atributes and
\\ scrolling/non-scrolling attributes as per CSS1, so you could even easily
\\ have non-rectangular frames. I.e a box in the top right corner could be used
\\ for navigational purposes while the rest of the screen could be used for the
\\ actual information.

Multiple body elements would be a not-so-good idea, IMHO. A single body
element is great because it points out the primary content of a page. I
think we already have a mechanism in place (two in fact) that would allow
for such a layout.

The first is the LINK element, (almost) always ignored. If the document has
a link to a TOC page, there is no reason why the TOC page cannot be popped
up in a subordinate page. This mechanism also avoids the need for
duplicating content across pages. Further, it allows the author to indicate
the content type of the linked page.

The second mechanism is the <div> tag. This could again be used to
partition the page. The author could in fact indicate the content within
the tag, which could then be used *by the browser* to decide if a separate
partition ought to be created for the content type. CSS could be used to
define new content types for which subordinate windows ought to be spawned.

The emphasis I place on the browser making these decisions is two fold:
1. The user may have the ability to control the browser's display.
2. The browser may in fact be an automated agent, in which case the
   semantic markup would be much more useful.

Sunil