W3C home > Mailing lists > Public > www-html@w3.org > January 2003

Re: XHTML 2.0 considered harmful

From: Sampo Syreeni <decoy@iki.fi>
Date: Thu, 16 Jan 2003 00:07:24 +0200 (EET)
To: Petr Baudis <pasky@ucw.cz>
cc: www-html@w3.org
Message-ID: <Pine.SOL.4.51.0301152354280.5935@kruuna.Helsinki.FI>

On 2003-01-15, Petr Baudis uttered to Mikko Rantalainen:

>Well, now if you will imagine yourself in position of author of WYSIWYG
>editor which wants to be able to produce valid XHTML/2.0 and still be
>able to compete on the market, probably the only obvious pragmatic
>solution is to generate various random id names ("f97hacr5dt183a") and
>put the appropriate record for given id to the stylesheet.

The right solution would be to decouple style and content, even in the
editor end of the chain. This is something which isn't addressed by the
current WYSIWYG editors, true, but that isn't really a problem of XHTML or
CSS. It's a problem with the editors.

>I believe that encouraging of such practices is a bad idea.

Why, precisely?

>I can see no harm caused by _allowing_ the style attribute - then, you
>will address the possible problems, both current and future ones,
>satisfying the world.

You won't, actually. My basic reference point is, I always think what it
would be like to use a speech or plaintext browser. (I go with newest Lynx
and Mozilla versions, in about 50/50 proportions.) If you make *any*
semantic distinction based on classes, id's, whathaveyou which isn't
marked up by a shared semantic element, you'll be doing harm to those who
rely on (X)HTML alone. That sort of thing should clearly be discouraged.

I believe Mikko's point boils down to the fact that it isn't possible to
preserve all semantic distinctions with *all* current and future XHTML
applications if style becomes something that has to be preserved. I agree
with the proposition, and also think it has to be a primary objective for
XHTML development.

Style is style. Content is content. If style can be used to carry content,
such style simply shouldn't be taken into consideration in the context of
XHTML development until after we have a deterministic way of translating
said style into unambiguous formal semantics, usable by *any* XHTML user
agent whatsoever.
Sampo Syreeni, aka decoy - mailto:decoy@iki.fi, tel:+358-50-5756111
student/math+cs/helsinki university, http://www.iki.fi/~decoy/front
openpgp: 050985C2/025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2
Received on Wednesday, 15 January 2003 17:07:31 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 30 April 2020 16:20:48 UTC