Plans for future versions of noweb
Plans for noweb 2.10 or 3:
- make Lee Wittenberg's `nocond' filter for conditional tangling
part of the standard distribution.
- revamp the distribution. I want to minimize the `official' ftp
distribution via CTAN; most options will leave the official
distribution and be distributed only through the WWW.
Plans for future versions beyond 3:
- better support for browsing multi-file documents and having
automatic links between documents.
- exploit compiler cross-reference info where available
Noweb wish list
Here's a set of things I'd like to see added to noweb one
day.
I may eventually do some of these myself, but for most of them I
expect to rely on the generous contributions of noweb users.
- Some modifications to the nuweb front end (numarkup) that
account for the different treatment of trailing newlines by nuweb
and noweb.
- A back end like unmarkup, except one that converts to nuweb source
form. This conversion involves some newline fiddling that's a bit
trickier because it requires global knowledge.
- The ability to change noweb's markup conventions (e.g.,
<<...>>
) by means of command-line options.
- Adaptation of a lexer generator like
re2c
, so
noweb can more easily implement language-dependent
functions like macro processing and prettyprinting.
- A macro processor based on Lee Wittenberg's suggestion of using
quoted code in chunk names to define and instantiate macros. Such a
processor would require at least a minimally smart lexer, so as to
avoid substituting in the middle of quoted strings, comments, and
such.
- An adaptation of the prettyprinter generator from Spidery WEB.
- A version of noweb with an embedded command language, and versions
of all the filters written in C, so that noweb could be implemented
without any Unix shell scripts or Unix tools. I'm still looking for
a good embedded language; Lua
seems the best option so
far. If this adaptation is ever made, it will be version 3; see the
Noweb 3 white paper for more information about
plans for version 3.