Adaptive Cooldown Times

[Tookmund]

(Copied from http://pqxx.org/development/swapspace/ticket/14)
Recognize pattern of consistent policy reversals too close to timeout—or too soon after decision is made. Lengthen or shorten cooldown periods accordingly.

Don't go overboard with adaptiveness though; when adaptive parameters start to interfere, there is a risk of going into a feedback cycle or other pathological state.

[jtv]

Hi @Tookmund — thanks so much for running this fork. I found your email long, long after you sent it and I doubt you ever got my reply.

I just had a little disaster on pqxx.org and had to rewrite the development sites as static pages. The main page for Swapspace now links to your project.

Unfortunately, the links to those tickets on the old site are now all broken. My apologies!

[Tookmund]

Yeah, email is certainly not the most reliable thing sometimes.
That's unfortunate. Crazy to think how much you had to do originally to host your projects.
Glad I duplicated the issues when I did, though I think swapspace is unlikely to see any big changes from here on out. It just works, and I'd like to keep it that way.

[jtv]

And that is really good to hear. :-) Thanks again.

[pragnesh]

original site ticket can be accessed using archive.org https://web.archive.org/web/20100821101550/http://pqxx.org/development/swapspace/report/6