Lists Home |
Date Index |
David Rosenborg wrote:
> David Carlisle wrote:
>>but the main point of svg is to specify lines and curves and stuff. It
>>makes sense to css style them. Similarly xhtml specifies headings and
>>lists and other structural things, it makes sense to css style that.
> I don't think your argument holds here. Just because the SVG
> vocabulary is more elaborated in terms of how many
> specific elements it contains doesn't make it more fit for CSS than XSL-FO.
Well,actually we have:
- the CSS syntax for writing down several presentational properties
- the sematic of several presentational properties.
XSLFO needs to specify several presentational properties. It reuses
most of the semantics from CSS2, with some refinements and a few
incompatible changes. It does not use CSS syntax to express the
properties, instead it uses XML attributes.
SVG also needs presentational properties for it's elements, and they
have choosen to use CSS syntax, one reason being that XSLFO was
developed later and slower, the other was that SVG is supposed to be
transported over the network, which profits from the less verbose
CSS syntax. Note that XSLFO was supposed to be produced almost always
in situ in the very process where it is consumed, therefore avoiding
verboseness of the syntax was not considered to be of the same
importance as it is for SVG.
So what's the problem? Some people like CSS syntax, because they know
it well, they have experience and tools, and because it can be slim.
Other people like the XSLFO resp. XSLT+XSLFO apporach more, because it
is more of pure XML, and can be completely handled with XML tools.
Do you notice we are about to fight just another religious war about
"the best syntax"?
IMHO the correct approach would have been:
1. Decouple syntax from semantics.
2. Describe the semantics of the relevant presentational properties in
abstract terms, including the "inheritance" concept and whatever
seems to be necessary.
3. Define a CSS grammar. Define how the abstract presentational property
definitions are mapped into CSS syntax.
4. Describe a mapping of the abstract presentational property definitions
into an XML attribute syntax (and something about inheritance), and
perhaps a generic way to use it with any other vocabulary which needs
to express presentational properties.
5. (optional) Describe an XML vocabulary which can express presentational
property definitions, and a generic way to use it with any other
vocabulary which needs to express presentational properties.