[plug] Re: Argh, plug at plug.org.au

[James Devenish]

Well, this PLUG stuff is becoming too difficult. I'd like to have saved
about 300 people from refiling today's e-mail and updating their list
filters, but for a number of reasons (including still not being able to
log in to the new server -- long story), there's not really much I can
do. And with odd things happening off-list, it's just been a bad week
all round. I'm giving up on a number of list-related items, but I am
writing them here so that others can pick them up (I'd have rather done
them myself and taken credit for them, but I guess I don't have the
time). Of the ideas I can currently remember, the ones for the website
& mailing list were:
 - Find a way to overcome the following problem: under the old regime,
   mailing lists were under mail.plug.linux.org.au. Because of this, it
   was possible to search the mailing list by restricting yourself to
   the mail.plug domain, or search the website by restricting yourself
   to the www.plug domain. Someone changed the search to plug.org.au so
   that the mailing list swamped out the website. Now the mailing list
   is at www.plug, so Google is indexing the mailing list and website
   together. Fortunately, Google seems to have settled itself and
   searching works fairly well. However, I'd still like to work out how
   to re-separate the 'website search' from the 'mailing list' search,
   and then update the PLUG website forms accordingly.
 - Update website design's copyright notice for 2005.
 - Originally, Google was giving me better results than Lurker so I kept
   using the Mailman archives. However, because the list has moved
   around so much, Google's indexes have become fractured. Fortunately,
   Google should now be indexing the Lurker archives fully (hmm...so
   much for saving disk space with Lurker). I still want to put the
   Mailman archives back online because they offer 'by date', 'by
   author' and 'by thread' listings. When I last looked, Lurker only
   offers its 'dynamic' listing. After letting the Mailman archives
   stablise, I wanted to update the website to provide links to
   Mailman's flat listings. Personally, I make use of the Date listing
   to check up on gaps or pauses in delivery of list messages (and to
   read the list if my e-mail is down).
 - Fix the mail<->news gateway. In particular, there are differences
   in the interpretation of Message-IDs in NNTP versus SMTP (e.g.
   Mailman can happily accept the same Message-ID repeatedly due to
   cross-posting or duplicate posting). The aim of the gateway would be
   to cope with cross-posts, repair bad Message-IDs (to prevent messages
   being lost by the NNTP service), and ensure the following headers are
   correctly munged for each service: Newsgroups, Followup-To, Reply-To,
   To, CC, Mail-Followup-To, News-Followup-To (dealing with cross-posts
   is a particular snag, but I had conceived of a way of reconciling
   Mailman with Leafnode).
 - List Guidelines and Newcomers' Welcome:
   + Convert the Guidelines to HTML and add them to the website.
   + Send the newcomers' message and guidelines to new subscribers (by
   adding the text to the Mailman interface -- don't forge to include
   the website URL).
   + Post the guidelines to the list ("New List Guidelines (attn: all
   subscribers)"), possibly repeating this every one or two months.
   + Update the Mailman listinfo page to refer to the Lurker and Google
   web searches, and linking to the list guidelines.
   + Update the footer for non-digest and digest recipients so that it
   includes whatever extra links are appropriate.