Re: Proposal: Mailing list archive requirement(s)

[George Michaelson <ggm at dstc.edu.au>]

The BSD archives are read by a similar sized body of people to IETF I think.

(ok probably not, but its within an order of magnitude)

FreeBSD uses mailindex, which is Perl5. The URL is:

	http://www.freebsd.org/~wosch/mailindex/

an example is

	http://docs.freebsd.org/mail/current/cvs-all.html

The ip-filter list which is a nice archive is a different example.		
http://false.net/ipfilter is using Hypermail. It has a very crisp look
and (to me at least) is functional. Hypermail had bad bugs a while back
which freaked the server if MIME bodyparts were in there. Maybe its back
on the development track now.

Samba is another package using Hypermail. It seems to behave pretty well
for their level of usage too (the CVS lists take a *lot* of traffic from
the atomic cvs commit messages.)

We use MHonArc internally. I like it a *lot*. But then I'm an MH-bigot.
MHonARC has scaling problems for larger lists. Its not insurmountable but
presumes a good sweep/archive cycle. the surrounding clutter in MHonARC
is very tunable. This I like.

I cite UNIX-centric archival/index methods. Its not apparent to me that
any other platform is going to encompass the kinds of behaviours people
will need from this service.

cheers
	-George
--
George Michaelson         |  DSTC Pty Ltd
Email: ggm at dstc.edu.au    |  University of Qld 4072
Phone: +61 7 3365 4310    |  Australia
  Fax: +61 7 3365 4311    |  http://www.dstc.edu.au