[21:34:50] <aclark> j00bar: so i poked at redis and designed two failed database schemas for vanity.aclark.net. keys would be perfect e.g. Pillow:18900, except you aren't supposed to use them in your app code (they're good because i can call incr on them everytime someone requests a package)
[21:35:19] <aclark> j00bar: and my current #fail using strings which is in production :-)
[21:35:26] <aclark> you said to use sorted lists i think
[21:36:38] <aclark> but i can't picture it. i need: a queue to track packages as they come in; a counter to track the number of times a package has been requested (most vain list)
[21:40:38] <j00bar> a moderate sized machine can probably to 80k ops/minute
[21:41:02] <aclark> i mean incr <package>:<count> might do everything i need
[21:41:22] <j00bar> so if you can have redis do more lifting versus python doing parsing, you'll win 9 times out of 10
[21:41:43] <aclark> no wait, there was another prob… because i want to replicate the current behaviour which moves the selected package back to the top (via my filter func)
[21:50:25] <TheCowboy> i keep hearing a lot of good things about redis, something else i have to learn about i guess
[21:50:46] <aclark> TheCowboy: it's awesome so far