On 1/4/2011 10:00 AM, Adam Tauno Williams wrote:
>
>>>> Many people care about storage format.
>>> And they are misguided in doing so. Details of message storage is an
>>> internal [server's] problem.
>> So how do you suggest solving that problem when it is in fact a problem?
>
> You're missing the point of my objection.
The point doesn't matter. Solving the problem does. And you probably
can't solve it without knowing how things work.
> That the server is slow or difficult to manage [which includes backup]
> is the issue - not that it uses MBOX [regardless of that its use of MBOX
> is the root of the *server's* issue].
>
> Thinking of it in terms of messages-storage-format is misguided [I only
> mention it because I see the MBOX/MH/Maildir/Maildir++/etc... debate
> frequently]. This represents, IMO, a flawed approach to the problem.
Yes, you could solve it by ignoring the related physics and throwing
infinite resources at it - if you have infinite amounts of money. Or
you could do a sysadmin's job and understand the physical constraints
and optimize the results you can get from them.
> Migration to a new solutions [an IMAP server] that provides better
> performance / management is the fix. Incidently, it will almost
> certainly not use MBOX.
Odd that you would mention that, just after saying you shouldn't care...
And in fact, some servers (e.g. dovecot) may handle more than one
storage format, leaving it up to the admin to choose which is best.
> If you are 'manually' crawling around in your message-store [the *only*
> case where you'd actually care much about storage format] indicates
> something else it wrong [as well].
Your backup system will most certainly be crawling around your message
store.
--
Les Mikesell
lesmikesell at gmail.com