ruote log_2010-04-06

2010-04-06 08:07:58 utc jmettraux tosch_le : hi !
2010-04-06 08:16:34 utc tosch_le hi john!
2010-04-06 08:16:42 utc jmettraux how are you doing ?
2010-04-06 08:16:43 utc tosch_le did you have nice easter holidays?
2010-04-06 08:16:49 utc tosch_le fine, thanks. and you?
2010-04-06 08:17:09 utc jmettraux eh, no easter holidays here :)
2010-04-06 08:17:19 utc jmettraux doing fine
2010-04-06 08:29:11 utc jmettraux there was someone named "vagmi" yesterday, perhaps he will come back today http://ruote.rubyforge.org/irclogs/ruote_2010-04-05.txt
2010-04-06 08:29:26 utc jmettraux he was looking for some ruote-kit + sinatra help
2010-04-06 08:29:38 utc jmettraux please help him if you have the time (and if he comes back)
2010-04-06 08:31:28 utc tosch_le sure i will do so.
2010-04-06 10:46:48 utc wayneeseguin jmettraux Thank you again :)
2010-04-06 11:46:33 utc jmettraux tosch_le : many thanks for the ruote-kit reply on the ML :)
2010-04-06 11:53:49 utc tosch_le it's always a pleasure. and i'm learning a lot, didn't think about the fact the rk is used as rest interface to ruote...
2010-04-06 12:07:05 utc jmettraux :) it surfaces sometimes, like in http://twitter.com/enricobianco/status/11484093014
2010-04-06 12:09:15 utc tosch_le nice tweet conversation
2010-04-06 12:23:14 utc jmettraux had to add a quick note to the mail thread, "one more way to shoot yourself in the foot"
2010-04-06 12:29:19 utc tosch_le that may not only hit the foot but also the whole leg...
2010-04-06 12:30:01 utc jmettraux yeah, but sorry, I hate when people complain about "imaginary limitations"
2010-04-06 12:31:02 utc tosch_le the more unfriendly but nonetheless righteous answer would have been: "read the f****** source code"
2010-04-06 12:31:40 utc jmettraux :) people can barely read an error message ;)
2010-04-06 12:33:38 utc tosch_le but i can understand asier, reading the source code is some work if you don't know ruby (although it's much more pain in the ass with other languages)
2010-04-06 12:33:53 utc jmettraux +1
2010-04-06 12:35:26 utc tosch_le what annoys me more is that he didn't write why he needs dynamic participant registration. that way, one needs a crytal ball to give an usable answer
2010-04-06 12:35:44 utc tosch_le s/crytal/crystal/
2010-04-06 12:35:54 utc jmettraux same thing here
2010-04-06 12:37:28 utc tosch_le but hey, he got some spots to look on. i'm happy -- the sun is shining :-)
2010-04-06 12:37:42 utc jmettraux how is the job ?
2010-04-06 12:39:12 utc tosch_le some ugly work on a php based app last week, but it's getting better: doing a rails app now and i'm considering a schema-less document store
2010-04-06 12:39:30 utc jmettraux oh cool, which one are you evaluating ?
2010-04-06 12:39:56 utc tosch_le wdyt? couchdb? mongodb? or redis? wayne was so enthusiastic...
2010-04-06 12:40:22 utc jmettraux depends on speed/accessibilty/volume of data
2010-04-06 12:40:44 utc jmettraux redis requires you to have all the data in-memory
2010-04-06 12:40:46 utc tosch_le there are only a few requirements
2010-04-06 12:41:05 utc tosch_le and the data volume will be rather low.
2010-04-06 12:41:34 utc jmettraux couchdb gives you futon, a ready-to-use web interface
2010-04-06 12:41:43 utc jmettraux take your time
2010-04-06 12:42:21 utc tosch_le it will be a resource planning calendar for an indepent cinema / event location
2010-04-06 12:42:46 utc tosch_le they need some flexibible attributes on the event model
2010-04-06 12:43:10 utc jmettraux sounds like you'll have fun
2010-04-06 12:43:40 utc tosch_le yeah, definitely! the first time i use cucumber and rspec for real :-)
2010-04-06 12:44:30 utc tosch_le how is /your/ job?
2010-04-06 12:44:35 utc jmettraux :)
2010-04-06 12:45:02 utc jmettraux doing fine, the requirements for the motherboard tracking system are slowly evolving
2010-04-06 12:45:15 utc jmettraux now have 3+ process definitions (tinier)
2010-04-06 12:45:32 utc jmettraux and some of them are tracking batches of boards instead of individual boards
2010-04-06 12:46:46 utc jmettraux speaking of redis : http://github.com/jmettraux/ruote-redis
2010-04-06 12:47:47 utc wayneeseguin jmettraux: Redis has a VM system now, so it only needs to store keys in memory and can store values on disk.
2010-04-06 12:47:51 utc wayneeseguin It's also persistent
2010-04-06 12:48:08 utc jmettraux wayneseguin: yes, looking forward to it
2010-04-06 12:49:15 utc tosch_le i've seen ruote-redis in my github dashboard. did you do some benchmarking already? (not that i would need a really fast ruote...)
2010-04-06 12:49:30 utc jmettraux ok
2010-04-06 12:50:01 utc jmettraux ruby test/functional/test.rb
2010-04-06 12:51:40 utc jmettraux ruote-hash : 63s
2010-04-06 12:52:59 utc jmettraux ruote-dm : 154s
2010-04-06 12:53:29 utc jmettraux ruote-fs : 91s
2010-04-06 12:55:51 utc jmettraux ruote-redis : 90s (but tested with ruby 1.9.1 while the other results are 1.8.7)
2010-04-06 12:56:23 utc jmettraux couch is... slow
2010-04-06 12:57:03 utc tosch_le "relax"!
2010-04-06 12:57:28 utc tosch_le or better: "time to relax" ;-)
2010-04-06 12:57:31 utc jmettraux ACTION is running --couch test
2010-04-06 12:57:35 utc jmettraux :)
2010-04-06 13:06:37 utc jmettraux ruote-couch : 569s
2010-04-06 13:07:08 utc jmettraux the frustrating thing is that I measured that ~40% of this time is spent in setup and teardown
2010-04-06 13:08:31 utc jmettraux but I'm working on two rails app ruote-kit + ruote-couch and the performance is acceptable (the slowness is not perceptible at all)
2010-04-06 13:10:05 utc tosch_le i suppose realtime responsiveness is no key requirement for ruote
2010-04-06 13:10:11 utc tosch_le ;-)
2010-04-06 13:10:30 utc jmettraux :)
2010-04-06 13:10:44 utc tosch_le i'm still perfectly happy with ruote-fs
2010-04-06 13:11:26 utc jmettraux JSON everywhere ftw