On Thu, May 05, 2005 at 10:29:56AM -0600, Charles Cazabon wrote:
> Yes, it happens, but it is a truly rare event in any case. There
> are also lots of other maildir-aware programs that can be used to
> read mail from maildirs , many programs that will read a maildir and
> write an mbox file (maildir2mbox is one), and the user can always
> fall back to `less Maildir/{new,cur}/*`. Defaulting to maildir
> would not significantly raise the bar.
You're still missing the point entirely. Switching a site over to
maildir is not difficult -- that's not the issue. The issue is that,
in production Unix environments, the default normally is already mbox,
and for various reasons changing it simply is NOT ALLOWED. This is
generally a policy decision that is out of the hands of the user, and
often even out of the hands of certain system administrators who would
like to change it.
Not that I'm one of those... Locking issues aside, I find that mbox
simply outperforms maildir in the common case, based on the way I use
e-mail. When locking issues aren't a worry, I'd choose mbox over
maildir in most cases, especially with very large mailboxes. mbox
opens these much faster than maildir, generally by a very significant
margin, as has been demonstrated in the past by various metrics people
have run and posted on this very list (or possibly mutt-users).
Caching can be used to improve maildir's performance in this regard,
but so is it true of mbox. Multiple lseek() calls are inherently much
faster than multiple open() calls.
One of the only reasons I've used maildir folders in the past is
because mutt's default method of detecting new mail in mbox folders
breaks when mail is being backed up by Amanda... Unfortunately Amanda
updates time stamps in a way which leads to mutt not realizing there
are new messages in a folder until the next time mail is delivered
into that folder after the back-up has run. If you're filtering
critical mail into a particular folder, but such mail is received
relatively infrequently, this can lead to important messages going
unnoticed and potentially causing a serious problem...
[Arguments against critical messages being sent by e-mail
notwithstanding... Again, often we have no choice in the matter about
how things are done in the real world, even when we can demonstrate
how stupid the current procedure is.]
--
Derek D. Martin http://www.pizzashack.org/ GPG Key ID: 0xDFBEAD02
-=-=-=-=-
This message is posted from an invalid address. Replying to it will result in
undeliverable mail. Sorry for the inconvenience. Thank the spammers.
Attachment:
pgpCNbqrvAiVC.pgp
Description: PGP signature