searchable archives

[Brian Behlendorf <br...@apache.org>]

On Sun, 17 Oct 1999, Bill Jones wrote:
> The Apache search engine doesn't find the mail lists on the dev server.

That's because messages aren't individually addressable through the web
interface at apache.org, only as big mbox files, so there's not much point
yet.

I have used hypermail in the past to archive messages, but found it to be
too buggy and too much of a resource hog - and I'm not really fond of the
idea of a separate filesystem file for every message, with the HTML
interface fixed into the file and thus unchangeable.  It could be
hypermail 2 fixes some of these issues, but I've not looked at it.

I've got a specification (and some simple proof of concept perl code
written) for a module or handler that can read standard mbox files plus a
corresponding DBM file which indexes message attributes, like "the author
of message 14", and produce a web interface to browsing that mbox like any
webmail-style system out there, with a configurable interface.  

Added to that, a plug-in to swish-e to read the mbox archives and be able
to map individual messages to their corresponding independent URLs, as if
they were separate files.

This is not a weekender project, necessarily, but it would be extremely
beneficial to our project and others.  If anyone is looking for a
well-defined hacking project that would really benefit Apache, let me
know.

	Brian