Permuted Indexes

[Sandy Harris <sharris -at- dkl -dot- com>]

John Lynch wrote:

> Subject: Help... Permutated Indexes in MS Word?
 
> Is this possible, if so, can you help me learn how to do it?

Did this produce any response?
For that matter, how do you do a permuted index in Frame?

If anyone needs it, my code to produce a permuted index from HTML
headers in a group of files is in the FreeS/WAN distrubtion:

http://www.xs4all.nl/~freeswan

For this code, grab the current snapshot or release 1.1 (Real Soon Now).
It is not in 1.0.

It is under the Gnu Public License, free to distribute, use, and modify.
I'd appreciate a copy of any interesting modifications.

WARNING: I wrote this to solve my own problem. It has not been designed
to handle all HTML files, or tested on any other than my own. It is
almost certain to misbehave somehow if applied to a large collection of
files I didn't write. On the other hand, you get the code, so whatever
problems appear can almost certainly be fixed.

For anyone unfamilar with the term, a permuted index takes a bunch of
one-line items like the simple command descriptions in Unix man pages

grep -- search for patterns in text

and turns them into a series of lines something like this:

            search for   patterns    in text                grep(1)
                         search      for patterns in text   grep(1)
search for patterns in   text                               grep(1)

where grep(1) refers to the page titled "grep" in section 1 of the
manual. Every "content word" in the input gives one line. Words
like "for", "in", "of" and "the" are skipped.

This becomes interesting when you do it for every command and combine
the outputs, sorting on the central column so a piece of your index
is:

search for patterns in   text                               grep(1)
                         text       processing language     awk(1)
                         text       editor                  ed(1)

Now a user with some text-related problem can find all the utilities
that do something with text.

This certainly isn't a replacement for a good index, table of contents
or cross-references, but I consider it a useful addition and would be
quite embarassed to deliver any substantial set of docs without a
permuted index.