W3C home > Mailing lists > Public > www-style@w3.org > October 1998

Re: Property suggestions

From: Ian Hickson <py8ieh@bath.ac.uk>
Date: Thu, 22 Oct 1998 15:28:59 +0100 (BST)
To: www-style <www-style@w3.org>
cc: Daniel Glazman <Daniel.Glazman@der.edf.fr>
Message-ID: <Pine.GSO.4.04.9810221510230.705-100000@amos.bath.ac.uk>
On Thu, 22 Oct 1998, Daniel Glazman wrote:
>> Content, Script, and Style are three different parts of a document,
>> and should be kept separately. It would be very sad for CSS to
>> aquire a "behaviour" property, but it would be great for CSS to
>> aquire a sister spec sharing a similar syntax, but which
>> concentrated on scripting while CSS concentrated on styling.
> In fact, Action Sheets don't use CSS general syntax. Action sheets
> applied to a selector syntax being the CSS syntax, a selection
> domain based on elements and attributes, and actions conformant to
> CSS declarations syntax give Cascading Action System (CAS). CAS is
> just one example of Action Sheets.
Sorry. I stand corrected.

> CAS and CSS look very similar because both based on the same general
> syntax.
Indeed. That, IMHO, is good.

FWIW, MS-like behaviours similarly have CSS+XML syntax. However, IMHO
their system unnecessarily pollutes CSS. See below.

> Even if I agree with the main lines of your very concise and
> remarquable comment Ian, I have to temperate your definitive opinion
> about the presence of a property defining behaviors (I did not write
> the MS "behavior" property...) in CSS.
Note, BTW, that in particular I dislike the idea of a "behaviour"
property. It would (a) end up being overloaded (having to cope with
every scripting thing involved!), (b) require an external function
declaration for every simple one line scripted bit (as I understand
it, anyway, cf Action Sheets which allow 'inline' scripting, as per
example below) and (c) cause huge problems with cascading, of a much
worse nature than the current CSS2 'counter' problems.

> If CSS and CAS interactions are well defined (order of application,
> circular references, ...), I clearly see nothing against
> 
> 	A > B[C] { onLoad : "initialize(this)";
> 	           color : black }
In that case, continuing that to it's logical conclusion, wouldn't you
rather have the following?

  <A> <B C="" onLoad="initialize(this)" style="color: black"/> </A> 

I believe the CSS, CAS and XML should be kept in three separate
documents. The "adding scripting to CSS", which is basically what
you're suggesting, is quite possibly the worst thing that could happen
to stylesheets. You basically have two totally separate beasts: a
system for specifying the look/sound/feel of a document, and a system
for specifying the behaviour of (parts of) the document. The
'behaviour' parts requires additional information such as scripting
language, function definitions, et al (see the Action Sheets proposal)
but CSS requires virtually none of those (basically, the character set
is all it needs worry about).

If you are suggesting CSS aquires the CAS ideas, then it is IMHO
unworkable, or will at least result in a major syntax overhaul (very
nasty!). If you are however suggesting that CAS aquires, in addition
to the event 'properties', the functions, et al, the entire CSS
property set, then that will IMHO result in a redundant and bloated
spec.

Better to keep them separate.

-- 
Ian Hickson
Received on Thursday, 22 October 1998 10:29:11 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:53:56 GMT