Re: The concept of cascading

> Not so difficult. But then, you skipped the hairy code, which --
> depending on the type of hair involved -- some might consider obscene.

I'll take that as a joke. The hairy code was of course just meant to 
demonstrate that parameterizations can be as simple or complex as you

> No, I think that there would need to be a great deal of forethought put
> into the design of a primary stylesheet as well as secondary ones, or
> it isn't going to work very well. I thought that was your opinion also.

Well, it IS possible that you could design a stylesheet with a brain-deaad
font and I could improve it in CSS by overriding all of the elements that
mention fonts explicitly and choose a better one. This would be harder in

> With care, though, CSS could help keep documents consistent throughout
> an organization without creating style chaos. As could a parameterized
> style mechanism. Is DSSSL-O a good mechanism? Your examples look good,
> but when I look at the spec it seems way too focused in some areas and
> totally blind in others, and I have a hard time imagining widespread
> acceptance.

Well, the spec is an ISO spec. It contains all of Scheme, for instance, for
those that need an escape hatch to something programmatic. Sometimes something
that seems major and complicated can be specified very compactly, so they 
don't put in reams of explanatory information to deomonstrate how major the
feature is. As far as blindness: that may be just your perception (the 
DSSSL-O spec is, after all, only a subset of DSSSL, not a rewrite), or it
may be the fact that DSSSL-O is not a Web standard: it is an online
presentation standard and has no particular attachment to the Web. So you'll
find it has a fairly large "blind spot" when it comes to URLs, for instance.

That is why Webifying DSSSL is the next major project for the XML Working 
Group. If you can specify what you think it is missing, we would certainly
look at it in the working group.

 Paul Prescod

Follow-Ups: References: