[Mailman-Developers] Pipermail replacement? Zest!

[Barry A. Warsaw]

>>>>> "JRM" == Jason R Mastaler
>>>>> <jason-list-mailman-developers@mastaler.com> writes:

    JRM> After looking at the Zest sample output, I'd have to say I
    JRM> still like pipermail's output better.

The thing that's really cool about Zest is that you can follow the
conversation almost completely by skimming the left frame.  Although
we won't know until it's in production, my gut tells me that it'll be
/much/ easier with Zest to zero in on exactly the message that you
care about.

The unfortunate thing about Zest is the frame layout.  Personally, I
hate frames primarily because of the bookmarking problem.  That might
be alleviated with a "bookmark this message" link.

    JRM> I think pipermail is fine, it just lacks search capability
    JRM> which is the only reason I still use Wilma to archive my
    JRM> lists (see http://mla.libertine.org/tmda-users for a sample).

Pipermail is basically a jumble of unmaintainable code.  So far, no
one's stepped up to the plate to fix the major glaring problems[*] and
I'm not going to do it.  If people get fired up about Zest aside from
Mailman, I'll be happy to bundle it with Mailman, or at least make it
really really easy to hook in.  If someone wants to build a better
Pipermail, I'm all for that too.

-Barry

[*] I think I've fixed the problem with the lack of MIME knowledge.
Now we need to fix things like:

- the huge memory consumption when rebuilding an archive from scratch

- the lack of predictable unchanging url for a message
  (i.e. regenerate from scratch and it's possible all your bookmarked
  urls will break).

- the ability to ask the archiver "what's the url for the message with
  this Message-ID:"

- the ability to handle multiple charsets in a single message

- performance problems for on-the-fly re-indexing when adding a
  message

- searching <wink>

- the ability for threads to span the artificial by-date index
  boundaries