W3C home > Mailing lists > Public > www-style@w3.org > April 1997

Re: From CSS to DSSSL

From: Paul Prescod <papresco@calum.csclub.uwaterloo.ca>
Date: Sun, 20 Apr 1997 18:24:11 -0400
Message-ID: <335A978B.F5269BA@calum.csclub.uwaterloo.ca>
To: www-style@w3.org
Chris Lilley wrote:
> > Let us presume that we are starting with a blank slate DTD.
> 
> We do not have that luxury. 

I could almost buy that. If we're talking about HTML, then obviously CSS
will have the advantage of "implying" the right formatting for
everything except the things that you care about. But XML is right
around the corner. It isn't for vanity home page creators but I think
that it is for serious webmasters.

You also mention XML here (after some interesting stuff I deleted):

> Given the buzz at WWW6, I would expect an overlapping third phase of
> simple XML documents together with CSS. I would expect to see some
> browsers offering DSSSL support at that point, since there will be
> automatically generated documents that use DSSSL. That DSSSL will be
> generated by a skilled Scheme programmer and possibly parametrised
> slightly by the document authors.

Why? What does DSSSL have to do with Scheme? That sounds strange, but it
is a relevant question. DSSSL is a stylesheet language. It can be used
completely declaratively as I demonstrated at
http://itrc.uwaterloo.ca/~papresco/dsssl/hard.html . It happens to also
be extensible through an expression language based on Scheme in the same
way that Netscape HTML is extensible "JavaScript". I'll admit that the
relationship is closer: DSSSL's notation is based on Scheme, whereas
HTML does not "look like" Javascript. But you can use HTML (and even
make Kewl web pages) without learning JavaScript and you can use DSSSL
without learning Scheme.

> There will be some excellent looking XML documents on the Web during
> this phase, that will use DSSSL; however it won't be hand edited
> nor would I expect it to be editable by a WYSIWYG editor. I would
> also expect to see such  browsers implementing CSS internally as a
> conversion to DSSSL, since I am assured that DSSSL can represent
> everything in CSS.
> 
> I would not expect to see any significant proportion (ie, more
> than 0.1% ) of HTML document authors producing DSSSL by hand
> in the forseeable future. Ditto for authors of any other type of
> SGML document intended for use on the Web.

So what are these people who are not using DSSSL going to use? Is CSS
going to add features for describing images, tables, columns,
line-breaks ...? By the time you are done, it will be more complex than
the declarative subset of DSSSL, because it will have been created ad
hoc, as the market demanded particular features.
 
> Making selections from the limited parameterisation of a DSSSL
> stylesheet programmed by someone else, yes. Causing a DSSSL stylesheet
> to be generated together with an XML document from some automated
> conversion of database information, yes.
> 
> Writing by hand, from scratch, for arbitrary DTDs - no.

Why not? Because you've heard that DSSSL is based on Scheme and you've
heard that Scheme is hard? I can teach someone to make declarative DSSSL
stylesheets for arbitrary DTDs in about the same time that it would take
to teach them how to make CSS stylesheets for those DTDs. When they are
done, though, they will have taken the first steps down a path towards
something substantially more powerful and extensible.

 Paul Prescod
Received on Sunday, 20 April 1997 18:18:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:53:49 GMT