[Squeakfoundation] Harvesting infrastrucure (was: Don't kill Bert's workstation)

[Bert Freudenberg]

Am Donnerstag, 05.06.03 um 19:26 Uhr schrieb Brent Vukmer:
>> Please remember these are not UIDs, just a count from 1 to
>> number-of-messages-in-archive. If something (someone?) (me??)
>> messes up
>> my mbox file, these numbers will change.
>>
>
> Right, I know they're not actually UIDs.  And in the scenario I 
> propose,
> I guess they *still* won't be UIDs -- the process will parse and export
> individual email .txt files, numbered in the order they're extracted.
> But there is no safeguard against a hiccup in the parsing process.
> Ideally someone should be able to download the mbox file from
> lists.squeakfoundation.org, parse it, and get the same output as my
> (theoretical) server.
>
> Anybody got suggestions for a good algorithm generating the UID?  How
> about a hash just using the subject line and the date?

Why don't you make the individual numbered .txt files the main 
repository? Easily served by Apache, easily updated by a procmail 
script or ftp upload or whatever. KISS ;-)

-- Bert