Re: mutt development status
On Tue, Jul 12, 2005 at 11:43:52AM -0400 I heard the voice of
Derek Martin, and lo! it spake thus:
>
> The only difference should be that mutt should treat folders with
> new mail consistenly across folder formats, which it currently does
> not do.
In theory, that consistent-UI pattern is of course dead on. But in
practice, it's somewhat icky. There's a lot of useful things that are
trivial with maildir, but really really really expensive with mbox.
No amount of "it would be better if" can wish that away. Similar
differences reach across to POP and IMAP as well, for instance. I
found more than I expected when I started writing a mail library.
One could, in theory, get full consistency just by eating the overhead
of doing very expensive things in each format (though which things are
expensive will differ). But I sure wouldn't want to use a mail client
that did that; it's be too farkin' slow. I think that's a bad, bad
idea. Heck, part of the joy of having different storage formats
available is that you CAN optimize for different things. It sure does
suck marking as read first message from a mbox, but reading a
compressed 800,000 message maildir is no fun either.
The only other way around it is to store a bunch of external state
(header caching is a step in this direction) for mailboxes. As well,
you'd have to store a bunch for mutt itself (your 'ring' of mailboxes,
for instance, would require cramming a bunch of state into a file
somewhere). That's what most other mail clients do, TTBOMK, and it's
a path that (for loudly debatable reasons ;) mutt hasn't taken.
I'd sure want a way to turn it off with a vengeance if it appeared,
but that's just where I sit. It's a tough discussion for me to really
get that much into, since I'm one of those boring people whose desires
for a mail client mutt already fills to 99% or better.
--
Matthew Fuller (MF4839) | fullermd@xxxxxxxxxxxxxxx
Systems/Network Administrator | http://www.over-yonder.net/~fullermd/
On the Internet, nobody can hear you scream.