Open Codex HISTORICAL entry

2007 Jan 23 | Broken lists

I'm presently cursing whoever changed the configuration/names of Wikipedia lists. Identifying emails in archives is sadly a difficult problem, it really need not be, but fortunately the good folks at the aimsgroup MARC also archive the lists and associate the unique identifier of every message with a persistent and unique URL, as I wrote about previously. But when Wikipedia moved its lists from "" to "" it not only broke email filters across the land, it broke the MARC archives evidently. No message is available in the MARC archive since the change, on January 6. Now, Wikipedians are realizing that many of the links from the Wikis to email messages (e.g., referencing a message on the Wikimedia Foundation list) are broken.

My backlog of email messages to scrutinize is growing as I hope Hank Leininger and the other volunteers at MARC find the time and means to address the problem. What would be great is if Wikipedia and other users of archive software (i.e., mailman) pressed for stable references to messages as a priority feature!

Posted by mako at Tue Jan 23 10:36:38 2007
For mailman lists, like the WM/WP ones, you can use the "List-Id:" header to identify messages from the list. A simple regex should work or you update a procmail recipe and use "formail" to redeliver.

Posted by Joseph Reagle at Wed Jan 24 08:50:23 2007
Awesome, Hank has updated the archives!

Posted by Mark Bergsma at Wed May 23 15:11:09 2007
(I just stumbled across this blog by coincidence)

I was the one who made that change in January. The reason we did that is because the old names had been used since the old days when Wikipedia was small and everything was running on a small multipurpose e-mail server. With our growth this was scalability causing problems. Therefore we just bit the bullet.

Before I made the change I informed all list admins with a detailed explanation, asking them to inform the list's users and other parties where deemed necessary. Unfortunately, few if any at all actually did.

Sorry for the trouble, we hope never to have to change the URL again and maintain stable archive links from now on. Although I have to say, Mailman does quite suck in that regard...

Posted by Joseph Reagle at Wed May 23 15:35:30 2007
Thanks for the background Mark!

Open Communities, Media, Source, and Standards

by Joseph Reagle