RE: css as xml?

Err... Umm... Sorry?  :)

Please note that I am neither for nor against a CSS as XML implementation.
And I offer my apologies for my previous poorly worded comment.  I was not
suggesting that there would be *ONLY* 1 parser, rather I meant that 1 XML
parser could parse 2 XML files (one containing data, the other containing
presentational information) and the application using that parser would
handle the data as it saw fit.  By this, I mean that you would not need 1
file parser for XML files and another file parser for the CSS files (though
I agree, the application will still need something to handle the CSS values
that it obtained from the XML parsing).  Is this useful?  From an authoring
perspective, this arguement has no relevance.  From a UA development
perspective, I would guess this might be slightly useful... but I've never
developed a UA, so I really can't say.  Just speculating.

As far as authoring goes, what *would* be the benefits of a CSS as XML
implementation?  I suspect the arguements for and against such a thing can
probably be found in the archives, and I have no real ambition to try to
think of any, so I'm happy not to go there.  Like I said, I am neither for
nor against it.  :)

Regards,
Peter Foti


> -----Original Message-----
> From: Tantek Çelik [mailto:tantek@cs.stanford.edu]
> Sent: Wednesday, February 19, 2003 6:20 PM
> To: Ian Hickson; Peter Foti (PeterF)
> Cc: www-style@w3.org
> Subject: Re: css as xml?
> 
> 
> On 2/19/03 12:55 PM, "Ian Hickson" <ian@hixie.ch> wrote:
> 
> > 
> > On Wed, 19 Feb 2003, Peter Foti (PeterF) wrote:
> >> 
> >> I haven't been following this thread, but I can think of 1 
> benefit of an CSS
> >> as XML... XML User Agents would only need a single parser, 
> vs. having both
> >> an XML parser and a CSS parser.
> > 
> > This isn't true. They'd still need an HTTP parser, a URI 
> parser, a CSS
> > value parser, a JavaScript parser, a PNG parser, etc. In 
> any case, the
> > parser is by _far_ the _easiest_ part of CSS to implement, 
> so this isn't
> > really an advantage.
> 
> I completely agree with Ian.
> 
> I can't believe people are *still* trotting out the tired and 
> completely
> irrelevant argument of "only need a single parser".  Not only is this
> completely false (as Ian handily demonstrated), but is about 
> equivalent to
> saying you "only need a single for loop".  It really is that dumb.
> 
> Of the dozens (millions?) of things you could try to optimize a
> data/programming/development model for, why the **** is there such an
> obsession with optimizing for a "single parser"?
> 
> Are people being paid per angle bracket or what?
> 
> Tantek (who is seriously wondering if there is some sort of 
> underground
> "angle bracket racket" (which would make a cool name for a band...))
>

Received on Wednesday, 19 February 2003 18:46:34 UTC