From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <4c95e398035a59475b2d48d243fec3ea@vitanuova.com>
Date: Tue, 25 Jul 2006 14:38:27 +0100
From: rog@vitanuova.com
To: 9fans@cse.psu.edu
Subject: Re: Re: Re: [9fans] Re: 9fans Digest, Vol 27, Issue 52
In-Reply-To: <3e1162e60607241628s50e42262re44d206c4bd9c847@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Topicbox-Message-UUID: 8c5e1436-ead1-11e9-9d60-3106f5b1d025

> > have you found a upas limit?
> >
>
> I've not... I was just expressing my skepticism of it's ability to
> scale.  Perhaps it's not as bad as I thought.

well, it scales a certain amount but not indefinitely. i've been putting off pruning my mailbox
for a little while and it's currently got over 3000 messages (about 27MB), which makes for a upas/fs
taking about 57MB. that's not great.

some time ago i dived into making it index-based, but as with most of my little side projects
i got too ambitious and it never got completed. my aim was to enable tagging of mail messages
and quick keyword searching. selection was to have been done by cloning a mailbox directory
and writing a filter string to a ctl file, resulting in a directory with only the selected messages.

i'd like to get rid of the need to delete or archive any messages at all. the tagging should
get rid of the need to guess which messages are new or unread (or unreplied to, or whatever).

[for the record, the bit i stumbled over was the underlying tag/keyword store. so many
trade-offs in b-tree algorithms, and was that really what i wanted anyway?]