If there is a concensus that effort in this area should focus providing an interface to external search tools, I think that it would probably be best left to other individuals. I would be happy to discuss the pros and cons of the various alternatives.

I would like to see some kind of indexing mechanism as part of the standard mailman distribution. My requirements for a search are such that only simple word/phrase searching is required, and something like ht:dig would be overkill.

Perhaps it would be useful to have a mailman standard search, and then leave hooks in to allow integration with external search tools, should people require that. Personally, I'd like to see the index updated at the same time as the pipermail archives are, and it makes sense to have an indexing package that "understands" mailman, even if its an external module.

My inclination would be to provide some indexing and search mechanism within the existing framework. There are several reasons for this, not the least of which is that it fits well with my motivation for doing this in the first place.

Then I vote you do it, and I look forward to seeing what you come up with. You should fulfill your own requirements, even if other peoples' are different. If you leave it sufficiently open that other people can implement their own searches, then we end up with a stronger, more flexible and more open product.

Mailman's at a stange now where confidence in it is growing. Many people I speak to are looking at it with interest, and choosing to deploy it. By making is a more rounded tool, it can only further increase its appeal as a complete mailing list solution, rather than just a another mailing list manager.

Stu.