[03:05] <mhall119> first we have to point everybody at UDS to popey's server running Node.js, and if he doesn't have to call the fire department then we're good
[03:06]  * mhall119 is late to the topic, ^^ was about Etherpad Lite
[03:08] <paultag> mhall119: my firehose runs nodejs
[03:09] <paultag> mhall119: since it's inturupt driven and stores datastructs in memory (not per-instance such as django or php), it uses way less CPU and RAM then other stuff
[03:09] <paultag> unless you're an idiot about it
[03:10] <paultag> not to mention v8 is pretty damn awesome
[03:11] <paultag> longpoll + nodejs = awesome
[03:37] <mhall119> inturupt driven?  Isn't that how twisted works too?
[03:38] <mhall119> paultag: I'm seriously thinking of dropping my Queue datastructure and just making is a list
[03:38] <mhall119> but still calling it a queue
[03:39] <mhall119> I'm doing some major data structure refactoring on RF
[03:39] <mhall119> the last major item on my list before I open it up for testing
[03:40] <paultag> mhall119: thank christ for that
[03:40] <paultag> mhall119: I'm stoked for you, I love seeing stuff built
[03:41] <mhall119> I love building stuff
[03:41] <mhall119> I've already reached my initial goal for this project
[03:42] <mhall119> which, in and of itself, is a pretty big win
[03:42] <mhall119> btw, is the "thank christ" for me dropping the queue structure, or for me going into beta?
[03:42] <paultag> mhall119: dropping queue
[03:42] <mhall119> heh
[03:42] <paultag> mhall119: :)
[03:43] <mhall119> I just need to figure out an efficient way to let users point to the same feeds and articles
[03:43] <mhall119> which still isolating one user's data from another (until I add sharing)
[03:44] <mhall119> but I think I've got that figured, thanks to some Django helper functions, which will reduce the number of DB calls the additional layers will cause
[03:45] <paultag> hammer the db and set up memcached
[03:45] <paultag> if you get big enough where that's not enough you should rewrite anyway
[03:46] <paultag> just get it pushed and feature complete
[03:47] <mhall119> it's not so much about hammering the DB, it's about making 50 nearly identical calls when you could get all the data in one only mildly complex call
[03:47] <paultag> buggy but working 1.0 is better then a never-working 0.1
[03:47] <mhall119> I'm already at 1.0, this is 2.x now
[03:47] <paultag> mhall119: you know what I mean.
[03:47] <mhall119> but migrating the data is going to be a pain, and I really want to do it before I start bringing in a bunch of users
[03:49] <mhall119> also, I found a django-invitations app that'll give me google-style invites with very little effort
[03:49] <mhall119> which I'm happy about
[03:50] <mhall119> anyway, it's nearly midnight, and I'm laying down flooring in the little girl's room tomorrow, so I'm off to be
[03:50] <mhall119> d
[03:50] <paultag> mhall119: night
[03:50] <mhall119> night