W3C home > Mailing lists > Public > www-style@w3.org > April 1996

Re: CSS and SGML document formatting -Reply

From: Wolfgang Rieger <rieger@bse.de>
Date: Fri, 26 Apr 1996 16:11:49 +0200
To: Gavin Nicol <gtn@ebt.com>
Cc: www-style@w3.org
Message-Id: <3180ce38.12423266@mail.space.net>
On Thu, 25 Apr 1996 04:55:52 +0200, you wrote:

>>IMO, wether CSS is good enough for general SGML usage cannot be
>>decided by a committee. If CSS is extended to support arbitrary DTDs
>>and people who are using CSS style sheets for their HTML documents
>>start using CSS for SGML formatting and do not start using DSSSL-O,
>>then the users will have decided.
>I agree that end users make the final decision, but on the other hand,
>I can tell you right now that CSS, as currently defined, simply cannot
>be used to format arbitrary SGML, and the changes required to make it
>even capable of doing so (which is different to being useful if it is
>possible to do so), would be incompatible with almost everything
>written so far.
>>For me "extended CSS" may be good enough to specify formatting for
>>conversion of SGML to
>>- RTF
>>- WinHelp RTF
>>- paginated formatted line printer output (like nroff)
>>- text display output
>>- TeX
>>and similar formats. And I do not think that very much has to be done.
>>But I may be wrong.
>You are wrong.
I am sorry to hear that. And I agree with you, as far as _arbitrary_
SGML is concerned. Of course "extended" CSS would only be applicable
to a limited but important set of DTDs. I did not make that class
explicit, but I wrote:

>>The first thing to do is to enable CSS to express everything necessary
>>to process HTML documents without knowing anything about HTML.

As soon as this has been done, you have a set of document types to
which "extended CSS"-formatting may be applied. Such DTDs are
supersets of HTML and DTDs which use (parts of) HTML to represent
generic text, and - in general - SGML documents who need no reordering
(i.e. no STTP) and contain the information to be presented already in
explicit form. This is a limited, but important and common class.

To do so, the element semantics implicit when formatting HTML have to
made explicit (i.e. "H1 is a heading", "IMG is an image", etc.).

>>I was tempted to implement CSS1, I was not tempted to impliment DSSSL.
>>I was rather intimidated by DSSSL :-) And it is hard to pronounce,
>This is part of the problem. People don't understand DSSSL, or are
>intimidated by it, because they don't understand the problem it's
>trying to solve to the level of detail the designers of DSSSL do. 

This discussion reminds very much of the discussion of HTML in
comp.text.sgml. Whatever that may imply ;-)

>I think that CSS and HTML will be used for a while, but eventually, I
>think that they *must* become as complex as the other solutions.

I hope not.

>CSS is small, it's easy. Anybody can work with, or design something
>like it. The one thing is is *not* is suitable as a general-purpose
>stylesheet language for large scale publishing of structured

May be. And may be not. But I insist that there should be something
like CSS+ (whatever you call it), which can be used to format
supersets of HTML and documents based on HTML-ish DTDs.


Buero fuer Software-Entwicklung           Email: rieger@bse.de
                                          WWW  : http://www.bse.de/
Rosenheimer Str. 214                      Phone: +49 89 497738	
81669 Munich, Germany                     Fax  : +49 89 497738
Received on Friday, 26 April 1996 10:09:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:53:44 GMT