W3C home > Mailing lists > Public > www-svg@w3.org > July 1999

Re: SVG requires a parser in a parser?

From: Chris Lilley <chris@w3.org>
Date: Fri, 23 Jul 1999 14:14:31 +0200
Message-ID: <37985CA7.9B436297@w3.org>
To: Paul Prescod <paul@prescod.net>
CC: www-svg@w3.org, Rainer Prosi <rprosi@hdpp.de>


Paul Prescod wrote:
> 
> Chris Lilley wrote:
> >
> >
> > Recall that style can be specified on parent elements, in a style
> > element, or in an external style sheet, and will cascade in accordance
> > with the normal CSS rules. This is much easier to accomplish when using
> > the normal DOM 2 CSS Object Model, with a single style attribute which
> > is just one part of the complete cascade.
> 
> I do not know why you say this. 

Well, it's not just to annoy you (although I know that it does have that
effect, because we seem to have very different mental models).

> I asked in a thread months ago but did not
> hear from anything from you. You keep accusing me of re-starting this
> debate but then we always get to the point where you almost understand
> what I am proposing and then quit.

Thats because, after some exchanges, we typically get to the point of
understanding what each other is proposing and then I say "I prefer my
model" and you say "I prefer *my* model" and there isn't much further we
can go from there.

> CSS's inheritance and cascade rules are part of its *semantics*, not its
> *syntax*. In fact, one wouldn't know anything about inheritance from the
> BNF *at all*.

Theoretically true; in practice, multiple variant syntaxes for the same
thing are not a great idea. Remember JASS, which had identical semantics
to CSS1 (or an early working draft therof) but very different syntax.
Recall the difficulty people have with working with RDF, with its
multiple ways of saying the same thing. And this is why, for example,
the metadata folks come up with schemas so that they can say, by
"author" I mean precisely this... and while theoretically I could dream
up a lisp-like variant xml syntax and have exactly the same semantics as
any XML instance, in practice it would not do me the slightest bit of
good.

> Inheritance works perfectly well in languages that map
> properties to attributes. xml:lang is an example (even if its *semantics*
> are not implemented in the DOM or some other particular API).

I think that counts as an own goal. Yes, theoretically xml:lang is
inherited and in practice, any program that wants to act upon it has to
implement that inheritance themselves on a case by case basis.

> NOTE-XSL-and-CSS also depends upon semantic inheritance with properties as
> attributes. 

That Note has no status - it was merely a discussion point, and with
hindsight was more confusing than helpful. It can't really be compared
with a widely implemented and stable Recommendation like CSS.

--
Chris
Received on Friday, 23 July 1999 08:14:17 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:17 GMT