mailing-list for TeXmacs Users

Text archives Help


Literate programming in TeXmacs and other wishes


Chronological Thread 
  • From: address@hidden
  • To: texmacs-users <address@hidden>
  • Subject: Literate programming in TeXmacs and other wishes
  • Date: Sun, 15 Dec 2002 08:50:38 -0500

Hi all,

I was reading about literate programming recently. In some way literate
programming remembers me the way I (and I think most of people) do when
use TeXmacs and CAS simultaneosuly in producing documents. We start
taking a problem, dicuss the problem, the solution, and then we insert
the code of a CAS session wich help to solve the problem. The document
is intended for human readability and then the source code for a CAS is
presented. I think that this concept can be extended in TeXmacs not just
for small CAS (or scheme) sessions but for another languages or more
long code.

I am thinking something like that:

1. To have the posibility to insert the code tag with a particular
grammar especified. Something like the present "Text -> Enviroment ->
Code" but adding a programming language sublist of Codes, for example
"Scheme, SQL, Perl, etc", a lá vim or emacs.

2. When the code grammar is specified we can use some visual helps like
higtligthing for a particular grammar. The Higtligthing can be specified
in some style file a is visually similar to the preamble mode with marks
in diferent colors for reserved code words.

3. In the same way that we can specify the computational engine for a
CAS, when we specify the programming language for a code whe can
specify the "engine" for the language. Bindings for every "engine" must
be constructed, in similar way that we use the bindings for CAS. For
example if we specify "Text -> environment -> Code -> Scheme ", we will
use Guile as a Engine to interprete that code, by default, but some
other people will be interested in using DrScheme as engine for
pedagogical reasons. People could change the "engine" for every
language.

4. Because we are now not only using interactive sessions, but also
writing a more long code, we would "execute" some parts of the document.
May be selecting them and then choosing and "execute marked parts"
option or something like that. Every marked part is passed to the
respective language "engine" and results of execution are shown. Because
of the estructured nature of TeXmacs may be it can be done generating
temporaly files of each Code Section and send them to the specified
language "engine".

5. There is also the "tangling" and "weaving" process. I really dont
have a clue/sugestion about how to made it.

What do you think. Is a worth work to be done? What part do you see more
easily implemented in the short future?

Offray.

pdt: Another wish is the posibility to have a "Document -> Style -> Bug
Report" in the menu. So the detected bug can be filled as a TeXmacs
document and send as a atached file to the proper e-mail.




Archive powered by MHonArc 2.6.19.

Top of page