rebuilding our archives is difficult and fraught with peril

[gosko]

Currently it is difficult to make changes to our mailing list archives, for a number of reasons:

• we have many mailing lists (~2500), with many messages (~1.5 million)
• some of these archives are 20+ years old and were last rebuilt using obsolete software
• some of our lists use custom patched versions of our archive software

Because the stability of the URIs in our archives is very important, this makes us reluctant to deploy changes to our archive software at the risk of breaking something. (This is part of the reason our archives have failed to keep up with the times.)

It would be nice if we had a way to rebuild our archives with confidence that nothing has been broken, possibly by generating them in a scratch space and comparing the results from before and after to make sure messages didn't get moved around.

To fully modernize our archives we will probably want to convert them to HTML5 and possibly iterate over several designs, so it would probably be worth figuring this out sooner rather than later. This would allow for improvements to be deployed more quickly in the future as well.

[gosko]

As part of recent work to deploy a new version of hypermail we invested a lot of effort in cleaning up old list archives, and now have a script to rebuild them all at once. So this should be much less of an issue in the future.