A while ago, when discussing what ought to be on conlang.org, I
more-or-less promised to look into web-based list archives, since
egroups isn't doing all we'd like it to do. (BTW, is it more useable
if you are registered? *grmbl*) Furthermore, I'm attending classes on IR
(Information Retrieval) this fall, so I have an idea of the techniques
behind search-engines, and what would suit this list. Here's my
conclusions.

The myriad of archivers[*] out there, like hypermail and mhonarc, doesn't
have inbuilt search-mechanisms. You can view messages by thread, author,
subject, date etc., but if you're looking for something specific, a
phrase or similar, you're out of luck. (And with the subject-lines of
postings here on conlang-l tending to stay the same no matter what,
browsing subjects doesn't do much good either.)

Looking at it from an IR perspective, our list isn't very well suited
for search-engines as they are today, because the list is multi-lingual,
and the level of expertise in English varies wildly. Suffice to say,
ir-systems likes the orthography to be consistent and the total amount
of possible words to be less than infinite :)

So, building a system ourselves seems best. Volunteers? :)

On the subject of threading: the listserv at Brown seems not to preserve
the Message-Id and References headers, which makes true threading
impossible. Unless there is someone who has everything archived, -with-
those headers intact, or there is some other way of restoring them,
threading will be an imperfect kludge of an affair.

-----------------
[*] Hypermail, mhonarc etc.


tal., whose grammar ought to be taken out and shot
--
"Better living through conlanging"