Re: Status of XSL and CSS?

Bill Wadley wrote:
> Hello!
> I'm wondering what the current thinking is regarding XSL and CSS. I've read
> some archived messages by Paul Prescod and Simon St.Laurent
> (, but can't tell in
> which direction the SVG-WG is looking.
> However, CSS (at least CSS1) was created before XML (and therefore slightly
> kludgy when it comes to XML?),

No, it works fine with XML, but CSS2 works better because the selectors
are more flexible. The actual apparatus of both CSS1 and CSS2 works fine
with XML, and there are several CSS1 implementations that style XML

> while XSL was designed for XML and is more
> powerful. Why is there a debate? Isn't using XSL self-evident? (Obviously
> not, but I'm wondering why :)

Because CSS decorates the source tree, and inherits down the source
tree, wheras XSL creates a new result tree and inherits down that.

For SVG, inheritance down the source tree is what we want. XSL
transformations could be used for generating or rearranging some XSL,
but for XSL to be used for styling it would need a set of vector
graphics formatting objects (which would exactly duplicate what SVG
does) which had specialised layout rules. Both CSS and XSL use a nested
box or area contraint layout model; graphics, in general, uses and
explicit placement and overlap model.

> Also, how do the "fill*" and "stroke*" properties
> ( relate to CSS? 

They use the CSS syntax and can be parsed by existing CSS parsers using
the forward-compatible parsing rules in CSS1 and CSS2. Some properties
are exactly the same as in CSS1 or CSS2; some are new, and of those,
some will become part of CSS in general and some will be SVG-specific.
There will be clearly distinguished in the specification.

> I've read
> (, and it says, `In general, CSS
> properties were used for the following:
>    Parameters which are clearly visual in nature and thus lend themselves
>    to styling. Examples include all attributes that define how an object is
>    "painted" such as fill and stroke colors, linewidths and dash styles.'
> Are we hoping to get these properties included in CSS (or XSL)?

Some will find more general applicability and some will be specific to
SVG (and so noted in the SVG spec).

> Is it
> possible we will split the property list between a style language and our
> drawing language depending on which primitives are supported where?

I need to know what you mean in more detail in order to answer that.


Received on Thursday, 8 April 1999 09:03:21 UTC