Re: HTML3 body tag

Murray Maloney (murray@sco.com)
Tue, 7 Mar 1995 11:25:57 -0500 (EST)


Subject: Re: HTML3 body tag
To: Dave Raggett <dsr@hplb.hpl.hp.com>
Date: Tue, 7 Mar 1995 11:25:57 -0500 (EST)
From: Murray Maloney <murray@sco.com>
Cc: www-html@www10.w3.org, html-wg@oclc.org
In-Reply-To: <199503071543.AA046700981@dragget.hpl.hp.com> from "Dave Raggett" at Mar 7, 95 03:43:01 pm
Message-Id:  <9503071125.aa05993@dali.scocan.sco.COM>

> > 2) A more viable solution might be to introduce an <INCLUDE> element.
> 
> > <DL> ....
> > <DT> foobar
> > <DD> <INCLUDE ID=foobar NAME="foobar" HREF="http://foo.com/bar.html.inc"
> >               REL="subdoc" REV="glossary" 
> >             TITLE="The definition of foobar is...">
> 
> OK, but we would have to use <!ELEMENT BODY O O (%body.content) +(INCLUDE)>
> to ensure that the INCLUDE element can be placed at any point in an
> HTML document body.

Well, either that, or we determine that <INCLUDE> may only contain
elements above the text level.  That is, <INCLUDE> might only 
be allowed to contain block- and division-level containers.
Just a thought.
> 
> > P.S.  Looking at the HTML 3.0 spec, I am surprised to see the REV
> > attribute missing from both <A> and <LINK>.  Why is that?
> 
> Do we really need REV and METHOD ?

I can't speak for METHOD at the moment, but REV is certainly needed.

Putting aside the list of REL/REV values that is currently suggested,
The following example is an example which illustrates how REL and REV
can be used together to advantage, with one value describing a 
directional cue and the other describing a relationship:

	<LINK REL=prev   REV=sibling TITLE="My sibling"     HREF="...">
	<LINK REL=next   REV=parent  TITLE="My first child" HREF="...">

It also occurs to me that it should be possible to attach 
one to many link semantics to a document.  That is, it should 
be possible for REL="next" to be used in combination with 
multiple REV="value"s.  So, if I am browsing through chapters
of a book -- that is my application -- I could "next" from 
the top of one chapter to the top of the next, and so on.
However, at some point I may want to step through the book
(or a specific chapter) linearly.  

There is a lot of potential power available through the use
of the <LINK> element, especially with a good set of REL/REV
values.  I'd be happy to say more about this when I have a 
bit of time to write something up.  
> 
> I also left out URN, feeling that none of these attributes have seen
> widespread useage over the last few years, so perhaps they are just
> cluttering up the standard for little real benefit.
> 
> (They are still in the DTD - so I need to get the explanation 
>  and DTD into sync again).
> 
> Comments please.
> 
> > I'm glad to see a list of values for REL, but I wonder why 
> > each begins with a leading capital letter.  Since case is
> > significant for these attribute values, according to the DTD,
> > I would suggest that lower-case values be used.  In fact, I wonder
> > whether we need to discuss that as a design principle.
> 
> As they are SGML NAMES, aren't they case insensitve?

My mistake.
> 
> > Hmmm!  Come to think of it, I don't think that we (HTML WG)
> > have any articulated and documented design principles.
> 
> I have tried to do this and welcome suggestions for improvements.
> 
> -- Dave Raggett <dsr@w3.org> tel: +44 117 922 8046 fax: +44 117 922 8924
>   Hewlett Packard Laboratories, Filton Road, Bristol BS12 6QZ, United Kingdom