No one mentions portability, but I believe it is a wonderful property of email - I have more than 15 years of email, moved across various servers and clients on various operating systems and it is still available and functional. What other personal communication channel is that resilient to technological change ?
Whitfield Diffie (yes, of Diffie-Hellman), just posted to the IP list an email he pulled up from 1976(!) in response to a crappy article about VA Shiva Ayyadura "inventing" email. I'm sure he's got even older emails, but the fact that he had it, was able to easily find it, and then resend it with a descendant of itself 38 years later is pretty amazing to me!
"
A quick look at my files finds messages from from 1976 that use that format:
29-SEP-76 1935 FTP:CERF at USC-ISI Data Communication Conference/ Sept 1977
Date: 29 SEP 1976 1802-PDT
From: CERF at USC-ISI
Subject: Data Communication Conference/ Sept 1977
To: walker
cc: cerf, wd at SU-AI
22-AUG-76 0857 FTP:LEFAIVRE at RUTGERS-10 MORE ON THE LISP COMPILER
Date: 22 Aug 1976 (Sunday) 1155-Est
From: LEFAIVRE at RUTGERS-10
Subject: MORE ON THE LISP COMPILER
To: DIFFIE at SU-AI
True. But on the other hand, its still (against all reason) too hard to actually do anything with such massive archives unless you really prepare for it .. like you, I have more than 15 years of email collected over the decades, but actually being able to do something productive with it depends on just how much time I want to spend setting up a system that can handle the mbox files of the 80's, the Eudora archives of the 90's, the Mail.app sql-database of the 21st Century, and so on. It has persisted, dear mail-format, but oh what a mess it still is after all these years. I'd kill for a solution that just gives me my own full-text search engine to query/manipulate this archive, without first having to configure stream-plugins to do text conversions and so on. I guess this is one of those cases where I should just set up a dedicated Linux machine (finally, a use for an rPi!) with all the years mails, and promptly set a text search tool on the problem .. trouble is, can I do that without having to train the thing to handle SMTP headers first? Intuitively I'd say "probably this is a done deal out there somewhere" but a part of me still cowers at the idea ..
(And before anyone mentions it: no, google is not my solution. I want to do offline manipulation of decades of mail archives .. not just turn it over to some Entity, Inc. to solve..)
You might want to have a look at notmuch (http://notmuchmail.org/). Said abruptly, it's a wrapper on top of Xapian (the search engine) that you feed with your email, after which you can query using mail-related queries:
$ notmuch search from:linus crap
and, most usefully, tag:
$ notmuch tag +rant from:linus crap
Tags work exactly how you would expect them to:
$ notmuch count tag:rant
There are myriads of frontends (even a web-based one [0]) if you want to go further than cli.
You're gonna have to transform everything to maildir or mh though (unless you can somehow iterate over your mails from the formats you have), but I guess that's not too unreasonable to do anyway.
And the reality is that there's a vanishingly small percentage of that archived email that is of any use to you or anyone else. I don't have anything before about 2001 for reasons of isolated proprietary mail systems (e.g. minicomputer-based), disk failures, or just not making the effort to archive. I do have about the next 10 years saved but it's sitting in Outlook on a retired Windows system I keep around for such purposes--but there's no easy way AFAIK to translate it to another format and it's not something I'm inspired to devote great energy to.
Since I always use IMAP, the local archive format is irrelevant to me - my mail entirely lives on an IMAP server.
I did have to migrate twice: once when switching from Eudora using POP3 to using IMAP, I moved the mail from the local Eudora archive to the IMAP server. Then other times when switching servers - either by moving the maildir or using IMAP synchronization.
I subscribe to the old archiving adage: if it is not spinning, it is dead !
