Re: text transformations in CSS?

Scott E. Preece writes:
>    From: Glenn Adams <glenn@stonehand.com>
> |       Date: Fri, 8 Dec 1995 10:13:00 -0600
> |       From: preece@predator.urbana.mcd.mot.com (Scott E. Preece)
> |
> |       Has the group thought about including more general textual
> |       transformations in CSS?  Something like
> |
> |	   P.abstract	{text-edit: "<B>Abstract</B> #value"}
> |
> |   This is a fundamentally bad idea, one which has surfaced from time
> |   to time and quickly dismissed (at least in the form you have presented).
> |   Namely, HTML and style sheets are designed to be dependent of each
> |   other; one should be able to parse an HTML document independently of
> |   its style sheet and vice versa.  If the style sheet were permitted
> |   to generate arbitrary content including markup, then this separation
> |   would no longer be possible.
> ---
> I guess this makes me wonder if the HTML/stylesheet model is pitched at
> too limited a level of functionality.  When you say "stylesheet" to me
> it covers a lot of things that do involve exactly this kind of
> transformation - things like order of elements, presence or absence of
> structural headings, ordering and punctuation of elements in
> bibliographic citations, etc.
> The CSS proposal seems much more at the level that word processing
> programs tend to use the word "stylesheet" - just typographic style
> control.  This has always seemed to me to be a key shortcoming of those
> word processing programs, and one of the reasons I thought SGML, which
> supports real structural markup, was a far better way for the future.
> I think I would be inclined to move my own authoring towards SGML,
> rather than HTML, assuming I can find good tools.

  For truly complex things like this, I really think that DSSSL and/or
DSSSL-lite are the way to go.  I've been reading the DSSSL spec and doing a
partial implementation for the Emacs-W3 browser (in my oh-so-copious
amounts of spare time :) and it is much closer to what you want than CSS.

  I don't think CSS should try to do _everything_ that DSSSL does.
Remember, it took 7 years to make DSSSL. :)

-Bill P.