<<< Date Index >>>     <<< Thread Index >>>

Re: The future of mailboxes?



I'm not a very smart guy (forwarded message below) ;-/

On 8/11/06, Seth Arnold <seth.arnold@xxxxxxx> wrote:
On Fri, Aug 11, 2006 at 10:40:03PM +0000, Dave wrote:

Incidently, this mail went only to me; should it have gone to the whole
list? If so, feel free to bounce this reponse, or add mutt-dev back to
the Cc: for any responses..

> I actually built something called mboxfs on FUSE a couple of years ago
> (after I got really annoyed with every new version of Mutt breaking
> something else in my config).  The idea was rather simple: provide a
> safe and simple way for anybody/everybody to do basic operations on an
> mbox without a mail client.  It even worked fairly well.  My server's
> motherboard died on its way across the Atlantic back in November,
> though, so I no longer have it (and I don't think it was good enough
> to be worth wasting time hooking the drive up to another SCSI
> controller to DL the stuff - LOL).  I'm planning on building a new

Oh bother ;) I've lost a bunch of code to apathy before ;) but a dead
machine, that's a hassle..

It sounds a bit like 'mh', which always interested me at an abstract
level, but .. first elm, now mutt, was just too much of a draw.

> version of the same thing based around a MySQL database.  (Yeah, yeah,
> databases are bad for email, blah blah blah ... I think I've got
> enough servers to throw at the problem, and nearly a terabyte of RAID.
> I'm far more concerned with separation of the service from the
> filesystem, and SQL is the ideal abstraction interface ... and
> wickedly fast to boot.)

I'm not at all convinced by that article that a database engine is the
wrong choice for a MUA backing store. Database-backed mail clients like
Outlook, Evolution, Thunderbird, etc, seem to do 'ok' -- once you set
aside their glaring problems :) -- but the database end of those systems
appears to be fairly well designed for large mail loads.

I thought about doing a postgres backend for mutt myself, but my first
coding experience in mutt showed me that perhaps an experienced mutt
programmer would make more headway. :)