Re: Will behaviors be added to CSS?

> The real question is whether linking behavior should be part of the
> structure or the presentation, of course...

Perhaps it should be neither.

A static document is easily separated into structure and presentation
by the questions "what *is* it?" and "how does it *look*?".

There should be a third property too, the behavior, which answers the
question "what does it *do*?".

Javascript has been used to describe behavior, as have CSS (depending
on point of view, the ":hover" and ":active" pseudo-classes represent
dynamic change of presentation, as well as the presentation itself).

I can see links as both stucture and behavior.

A link is a reference to another document, so it is (meta-)structure
in the web of hypertext pages. It represents the linked page. It makes
sense to mark an attribute as containing a "link", i.e., a reference
to another ressource, because that is what it *is*.

What using a link *does*, is behavior. In browsers, clicking on an "a"
element replaces the current page with the linked page. The "link"
elements contain references to linked elements that are not displayed
in all browsers. The browser decides the behavior.

CSS has nothing to do with (hypertext) structure or behavior, so
linking should not be handled in CSS. Ofcourse, there is no generally
applicable "behavior language" to complement XML and CSS. Maybe there
should be one, or maybe the document "browser" should decide the
behavior in a non-document-controllable way.

That's my (minimal value coin of local currency)'s worth of opinion.
/L
-- 
Lasse Reichstein Nielsen - lrn@daimi.au.dk
'Faith without judgement merely degrades the spirit divine.'

Received on Saturday, 26 July 2003 08:52:08 UTC