[Mailman-Developers] Improving the archives

[Barry Warsaw]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Jul 4, 2007, at 1:16 PM, Dale Newfield wrote:

> Barry Warsaw wrote:
>> Maybe a way to think about this is that the canonical url is based on
>> the message-id, but then there's some way to distill even this down
>> to a tinyurl or simple integer that would be stable in the face of
>> full archive regenerations.
>
> The resistance to basing this on message-id has always been that  
> there's
> no guarantee of uniqueness...
> ...but I believe each list has some sort of counter for how many
> messages it's seen, so we could add another header with that  
> number, and
> use as a unique id the two concatenated together...
> (That way the archiver can know from the content of the header exactly
> how to generate the same unique id as mailman, which would allow  
> for the
> url-in-the-footer to happen w/o first hitting the archiver.)

I'm not crazy about this idea for a couple of reasons.  First, it  
means that someone who has a copy of the message that didn't come  
from the list (e.g. one of the two you will get of this message),  
cannot calculate this unique ID.  Second, things can happen to a list  
that might cause this sequence number to get corrupted.  Maybe a list  
will get deleted and then recreated.  Maybe it will get moved and the  
sequence number will get reset in the move.  Maybe the list will be  
upgraded to a new version of Mailman.

I think we can do just as well by using Message-ID + Date and get  
very low collision rates.

- -Barry

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iQCVAwUBRqCobXEjvBPtnXfVAQIHFQP/Sz6WVqyFmo0lraw0hyyP5x4AhgBPDQmA
/rFfSBRGbdORLXA2Ss0YdhI5cy8n7LMSsLawgtSt+JA7F5IEiC6Hk5C1M8C+Oe09
4ICYEuuL+gcXPPVc4aYtxp33HvPBFCzPJkGBS2PHaqCQkYIKdWHCtDZ8iLWCOxjc
b674lsQk9tM=
=a09C
-----END PGP SIGNATURE-----