[admin] Re: [Clump] Re: Rebel With A Cause

[Klaus Weidner]

On Mon, Nov 03, 2003 at 11:24:49AM -0800, Thomas F. Burdick wrote:
> I think the basic idea for cl-typesetting is to have three layers:
> 
>   input formats  ==>  typesetting engine  ==>  ouput formats
>     sexp                cl-typsetting            cl-pdf
>     tex-like                                     postscript
>     ...                                          HTML
>                                                  ...
> 
> So if something is written in a cl-typesetting input format, it should
> be possible to get it in any of the output formats.  Obviously it's
> currently pre-alpha, but certainly what's in the list above is planned.

I think I see what you mean - it would be useful to have an engine that
could for example create an HTML document containing an automatically
generated table of contents, index, cross-references and so on, and to
use CSS to provide an approximation of the intended print layout.

I'm currently doing something similar for documents that I'm maintaining;
I'm editing them in Perl's POD format (which is pleasant to edit due to
containing only a minimal amout of markup, i.e. making something italic
takes 3 characters I<foo>, compared to HTML's 7 characters <i>foo</i> and
LaTeX's 9 characters \textit{foo}). Since the pod2html converter doesn't
do a good job with section numbering and the table of contents, I'm using
pod2latex as an intermediate step and then latex2html to get the output I
want.

This is duplicated code since both LaTeX and latex2html need to parse the
TeX macros to generate the information they need. In a Lisp solution it
would be much easier to abstract out the intermediate interface here.

The layer to do that doesn't exist in cl-typesetting yet, and I'd see it
as a fairly separate thing that should be abstracted out instead of being
commingled as it is in LaTeX:

input formats ==> logical markup     ==> typesetting   ==> output formats
  sexp             table of contents     pdf:               pdf
  tex-like         index                   pagination
  docbook          section numbers         page# xrefs
  html             image/table numbers     graphical output
  POD              footnotes             html:              HTML
  ...              ...                     split into files
                                           internal hyperlinks
					   style sheet
					   text output
					                    plain text
							    ...

It's a matter of terminology where you draw the border between these
stages.

I'd love to take a crack at the "logical markup" phase if I can find some
time in my schedule (or a clone)...

-Klaus