html & style

Torgeir Veimo (torgeir@ii.uib.no)
Tue, 25 Oct 1994 18:52:51 +0100


Message-Id: <27550.9410251752@brems.ii.uib.no>
To: Multiple recipients of list <www-html@www0.cern.ch>
Subject: html & style
Date: Tue, 25 Oct 1994 18:52:51 +0100
From: Torgeir Veimo <torgeir@ii.uib.no>


I would like to add some concrete proposals to the debate of adding layout
semantics to html. In my view, layout semantics will eventually become part of
html.

Introducing new markup tag specifying layout is bad. One can't just introduce
new tags, when the old ones don't support the wanted features. Introducing new
markup tags usually breaks current implementations. This means <center> <font
size=5> etc. (How did they come up with that braindead fontsize scheme
anyhow?) 

Using attributes with existing tags is good, since logical style is preserved,
allowing for user-defined overrides and browser capability adjustments
(braille etc.). This means <p align=center>, <h1 align=center>, <p size=12pt>,
<strong size="big"> etc.  

The important thing here is that there already exist default layout for
different markup tags in current implementations. Overriding these with
attributes would be simple. Adding stylesheets to the html specification would
only override the default layout, for those browsers that understand this.

What we need is a basic, well-defined set of logical markup elements (such as
html+ or something else), which is a superset of current html standards. 
_Then_ we can add attributes to specify layout. Style sheets could be added
later, when we have agreed on how to do it. It would not be difficult to
specify a style sheet format if we first agreed on a basic set of layout
attributes.

A basic set of attributes could probably defined very easily, if we first
specified character layout attributes, and the defined paragraph layout
attributes as a superset of this. [If you know FrameMaker's
paragraph/character designer you know what I mean.]

This would give us what we want: 

 o logical encoding 
 o explicit layout if wanted, otherwise implied layout.

Doesn't this make sense?

--
Torgeir @ http://www.ii.uib.no/~torgeir/