Date: Wed, 23 Feb 2011 23:02:00 -0500
From: Greyson Fischer <greyson@foosoft.us>
To: nr@cs.tufts.edu
Subject: noweb + interpreter line = noscript


Dear Norman,

I have been a literate programmer for a few years now. Although I must 
limit my use of it most of the time due to corporate pressure, it comes 
in extremely handy for particularly new or challenging tasks. Of course 
I use noweb for many of my literate programs, preferring it over even 
cweb for C and C++ (sure, 'int' isn't bold, but at least it's indented 
the way I like to read).

I found myself repeating a pattern when it came to writing in 
interpreted languages; specifically those that use a shebang to specify 
their interpreter. I write the literate script, tangle it, and then copy 
the result off to be used.  The problem with this approach comes when I 
want, or need, to edit the script again later. Although I know it came 
from noweb (usually because of the complete lack of comments) I couldn't 
always track down the original source in a timely manner, leading me to 
make changes to the derived script, rather than the source.

So, I came up with a simple fix. noscript tangles a document on the fly 
(assuming the first line has %!) and executes it inline with the 
specified interpreter.

To use it, take a noweb document (for example: myscript.nw) which 
tangles <<*>> into, for example, a shell script.  Add a "%!/bin/sh" at 
the top of the file. Run 'noscript myscript.nw'. Done, the script has 
been executed.

I've attached my version 0.1 in case your interested (along with some 
trivial test documents). It's made it quite a bit easier for me to keep 
the document and the script together. Perhaps you, or one of your users, 
might find it of use.

Cheers,
Greyson Fischer