[texhax] PlainTeX/LaTeX 'Best Practices' help request

Charley Bay charleyb123 at yahoo.com
Thu Feb 26 04:06:40 CET 2004

I'm working on a content management system which
currently publishes HTML.  I'd like to also natively
publish PlainTeX, and maybe LaTeX (it may be easier to
use the higher granularity of PlainTeX, as the tool
already has access to substantial metadata that one
might otherwise rely on LaTeX to generate).

The tool has "internal string defaults" that it uses
to markup content, but you can override anything with
a simple text configuration file (for example, you
can specify 'pre' and 'post' tags, like for SGML
output).  Now, I'm working on getting good
configuration defaults for PlainTeX (and LaTeX).

I'm currently developing on Win32 using MikTeX,
but the code is cross-platform and I'm porting it
back to a number of Unix/Linux variants.  At the
simplest level, it's merely a command-line 'data
compiler' like any command-line compiler for 
programming:  text files (and resources like images)
are read in, transformed, and written out (as HTML,
TeX, DocBook, etc.)

I've bought copies of most of the TeX and LaTeX
books (Knuth, Lamport, Kopka, both Goossens books,
Seroul, Eijkhout, etc.) and have read much of the
freely available TeX intro documentation available on
the web.  However, I'm most interested in "best
practices" for robust (and beautiful) cross-platform
TeX authoring.  These will become internal 'defaults',
and users can override with whatever dialect or
customization they want.  Rather than learning these
through experimentation, I thought it best that I
find an experienced TeX author (I'd need 'review'
help anyway).

I have some 70+ 'regression tests' that I run, most
fairly small (a few paragraphs, an itemized list, 
an enumerated list, a nested section, etc.)  I'd like
to send my HTML output to a TeXpert who would send me
back a "best practices" translation in PlainTeX (or
LaTeX).  I'll modify my tool until it has resource
granularity to generate the same TeX source code
(including spacing, comments, and indenting).  Thus,
I *don't* merely want a TeX file that "works" or
which looks nice; I want a "best possible expression
in TeX" file.  (Yes, that can be subjective, which
is why users can override internal defaults.)

As this may seem quite similar to 'real work', it's
only fair that I state this is not currently an
open source project.  And, current plans are to 
eventually release it commercially.

As far as compensation, we'd have to work something
out.  Aside from 'thanks', we could look at access
to the software and documentation, and possibly
some modest dollars or financial contribution to TUG.
This is a small private effort, so we don't have a
Microsoft-sized budget.

Perhaps we could also look to summarizing our findings
in a 'best practices' document of example TeX
publishing scenarios.  I'd be interested in using
that, if anyone made it available.

Anyone interested in playing?

charleyb123 at yahoo.com

Do you Yahoo!?
Get better spam protection with Yahoo! Mail.

More information about the texhax mailing list