2011-01-29T00:30:58  *** amstan_ <amstan_!~alex@206-248-153-79.dsl.teksavvy.com> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T00:55:46  *** Allied_Envy <Allied_Envy!~AlliedEnv@adsl-68-88-234-130.dsl.hstntx.swbell.net> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T01:33:50  *** amstan <amstan!~amstan@206-248-153-79.dsl.teksavvy.com> has quit IRC (Ping timeout: 250 seconds)
2011-01-29T02:18:52  *** chris___0076 <chris___0076!~Chris_007@adsl-190-49-78.asm.bellsouth.net> has quit IRC (Ping timeout: 255 seconds)
2011-01-29T02:32:09  *** chris___0076 <chris___0076!~Chris_007@adsl-176-227-147.asm.bellsouth.net> has joined #aichallenge
2011-01-29T02:58:28  *** AlliedEnvy <AlliedEnvy!~AlliedEnv@adsl-68-89-170-33.dsl.hstntx.swbell.net> has joined #aichallenge
2011-01-29T04:41:59  *** AlliedEnvy_ <AlliedEnvy_!~AlliedEnv@adsl-66-141-23-177.dsl.hstntx.swbell.net> has joined #aichallenge
2011-01-29T04:44:12  *** AlliedEnvy <AlliedEnvy!~AlliedEnv@adsl-68-89-170-33.dsl.hstntx.swbell.net> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T04:46:01  *** Allied_Envy <Allied_Envy!~AlliedEnv@adsl-68-89-170-117.dsl.hstntx.swbell.net> has joined #aichallenge
2011-01-29T04:47:00  *** AlliedEnvy_ <AlliedEnvy_!~AlliedEnv@adsl-66-141-23-177.dsl.hstntx.swbell.net> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T04:58:35  *** delt0r_ <delt0r_!~delt0r@188-22-19-88.adsl.highway.telekom.at> has joined #aichallenge
2011-01-29T04:59:48  *** delt0r___ <delt0r___!~delt0r@188-22-160-160.adsl.highway.telekom.at> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T05:22:58  *** medrimonia <medrimonia!~medrimoni@APoitiers-258-1-125-244.w90-50.abo.wanadoo.fr> has joined #aichallenge
2011-01-29T05:41:42  *** AlliedEnvy_ <AlliedEnvy_!~AlliedEnv@adsl-68-89-175-47.dsl.hstntx.swbell.net> has joined #aichallenge
2011-01-29T05:44:13  *** Allied_Envy <Allied_Envy!~AlliedEnv@adsl-68-89-170-117.dsl.hstntx.swbell.net> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T06:14:35  *** Kingpin13 <Kingpin13!~Kingpin13@Wikipedia/Kingpin13> has joined #aichallenge
2011-01-29T06:17:56  *** mceier <mceier!~mceier@89-77-208-118.dynamic.chello.pl> has joined #aichallenge
2011-01-29T06:41:32  *** smellyhippy <smellyhippy!~asmodean@about/csharp/regular/smellyhippy> has joined #aichallenge
2011-01-29T06:49:51  *** Allied_Envy <Allied_Envy!~AlliedEnv@adsl-68-90-45-160.dsl.hstntx.swbell.net> has joined #aichallenge
2011-01-29T06:51:34  *** AlliedEnvy_ <AlliedEnvy_!~AlliedEnv@adsl-68-89-175-47.dsl.hstntx.swbell.net> has quit IRC (Ping timeout: 260 seconds)
2011-01-29T07:12:18  *** Meatkat <Meatkat!~bovard@217.113.66.228> has quit IRC (Ping timeout: 255 seconds)
2011-01-29T08:08:19  *** sigh <sigh!~sigh@28.250-50-210.dynamic.dsl.mel.iprimus.net.au> has quit IRC (Remote host closed the connection)
2011-01-29T08:11:20  *** Naktibalda <Naktibalda!~quassel@cpc5-heme9-2-0-cust144.9-1.cable.virginmedia.com> has joined #aichallenge
2011-01-29T08:41:15  *** medrimonia <medrimonia!~medrimoni@APoitiers-258-1-125-244.w90-50.abo.wanadoo.fr> has quit IRC (Quit: Leaving.)
2011-01-29T10:37:01  *** Accoun <Accoun!~1@static-213-88-2-78.chebnet.ru> has quit IRC (Ping timeout: 276 seconds)
2011-01-29T10:37:10  *** Accoun <Accoun!~1@static-213-88-2-78.chebnet.ru> has joined #aichallenge
2011-01-29T10:38:00  <delt0r_> wibble
2011-01-29T10:38:10  <delt0r_> any word on the new contest yet?
2011-01-29T10:46:12  <jmcarthur> ah, technical seminar over. i can now get back to more fun projects and such, yay!
2011-01-29T10:46:24  <jmcarthur> i have heard no word, but i've been distracted
2011-01-29T11:01:43  *** choas <choas!~lars@p578F6859.dip.t-dialin.net> has joined #aichallenge
2011-01-29T11:37:16  *** Allied_Envy <Allied_Envy!~AlliedEnv@adsl-68-90-45-160.dsl.hstntx.swbell.net> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T11:49:55  *** Mathnerd314 <Mathnerd314!~mathnerd3@dsl.6.28.53.206.cos.dyn.pcisys.net> has quit IRC (Quit: ChatZilla 0.9.86-rdmsoft [XULRunner 1.9.2.12/20101026210630])
2011-01-29T12:27:01  *** dr`away is now known as dr-
2011-01-29T12:29:54  *** mega1 <mega1!~quassel@adsl-63-195-37-158.dsl.snfc21.pacbell.net> has joined #aichallenge
2011-01-29T12:56:12  *** mega1 <mega1!~quassel@adsl-63-195-37-158.dsl.snfc21.pacbell.net> has quit IRC (Ping timeout: 260 seconds)
2011-01-29T12:59:09  *** delt0r___ <delt0r___!~delt0r@188-22-166-9.adsl.highway.telekom.at> has joined #aichallenge
2011-01-29T13:00:32  *** delt0r_ <delt0r_!~delt0r@188-22-19-88.adsl.highway.telekom.at> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T13:20:38  *** mega1 <mega1!~quassel@2001:470:1f05:548:215:58ff:fe7d:773b> has joined #aichallenge
2011-01-29T13:38:57  *** davidjliu <davidjliu!473a6d0e@gateway/web/freenode/ip.71.58.109.14> has joined #aichallenge
2011-01-29T13:40:59  *** amstan <amstan!~amstan@206-248-153-79.dsl.teksavvy.com> has joined #aichallenge
2011-01-29T13:40:59  *** ChanServ sets mode: +o amstan
2011-01-29T14:04:31  *** Accoun <Accoun!~1@static-213-88-2-78.chebnet.ru> has quit IRC ()
2011-01-29T14:16:27  *** Accoun <Accoun!~1@static-213-88-2-78.chebnet.ru> has joined #aichallenge
2011-01-29T14:36:33  *** Naktibalda <Naktibalda!~quassel@cpc5-heme9-2-0-cust144.9-1.cable.virginmedia.com> has quit IRC (Remote host closed the connection)
2011-01-29T15:04:29  *** mceier <mceier!~mceier@89-77-208-118.dynamic.chello.pl> has quit IRC (Quit: leaving)
2011-01-29T16:15:02  *** dr- is now known as dr`away
2011-01-29T16:55:54  *** javagamer <javagamer!~javagamer@ool-4356dfec.dyn.optonline.net> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T17:02:27  *** Migi32 <Migi32!~migi@178-117-104-154.access.telenet.be> has joined #aichallenge
2011-01-29T17:03:30  *** j3camero <j3camero!~jeff@207.219.39.71> has joined #aichallenge
2011-01-29T17:07:48  <j3camero> amstan you around? Or anyone who can help me because I am a git n00b.
2011-01-29T17:08:37  <j3camero> I wanna lay down some code so that an entire contest instance can be trivially deployed and launched inside a VM. That would allow me to see my changes live in my own private contest instance with almost zero work.
2011-01-29T17:08:50  <j3camero> It's something that would really quicken my development cycle.
2011-01-29T17:09:36  <jbroman> j3camero: So what's your git problem?
2011-01-29T17:09:52  <j3camero> Oh more of a github best-practice question.
2011-01-29T17:10:55  *** javagamer <javagamer!~javagamer@ool-4356dfec.dyn.optonline.net> has joined #aichallenge
2011-01-29T17:11:01  <j3camero> The main code repo for the contest is now at https://github.com/amstan/aichallenge/tree/master. I'm wondering what's the best-practice way for me to pull it down, make some changes, then commit my changes. I know there's a few ways, and I still don't quite grok which is the 'neat and tidy' way to do it with Git.
2011-01-29T17:12:17  <jbroman> j3camero: "git clone" to create your local repo, "git add" and "git commit" to commit them, "git pull" to pull upstream changes, "git push" to push changes to origin.
2011-01-29T17:12:54  <jbroman> As far as trivially launching a VM, looked into Vagrant?
2011-01-29T17:14:38  <j3camero> Oh I haven't seen Vagrant. I was just picturing installing Ubuntu Server on a qemu VM, "savevm"-ing it, then "loadvm"-ing it every time I want to deploy the contest to test things out. I will check out Vagrant.
2011-01-29T17:15:31  <j3camero> jbroman: so I really should git push my changes back to the repo? I wasn't sure if that was a bit pushy (no pun intended) or if it would be better to request that my changes be pulled so that they could be reviewed before getting committed or somesuch.
2011-01-29T17:16:35  <jbroman> j3camero: Depends if you're treating it as "your repo" or as someone else's. (This is more project policy than anything else.) If there's a maintainer that authorizes changes, pushing to your own github repository and then issuing a pull request is kosher.
2011-01-29T17:17:13  <j3camero> Ah that makes sense. Thanks.
2011-01-29T17:17:59  <j3camero> jbroman: things are kind of a mess right now due to much hurried work last term, but are you interested in laying down some code or maybe helping with the live deployment for the next contest?
2011-01-29T17:18:29  <j3camero> My plan is to get a bunch of servers on EC2 like we did last time and have everything nice and distributed and slick.
2011-01-29T17:18:53  <jbroman> j3camero: Possibly. Is this being coordinates on forums, mailing lists, or somewhere I can keep an eye on?
2011-01-29T17:19:39  <j3camero> Yes. http://ai-contest.com/forums/ under the "Behind the Scenes" section.
2011-01-29T17:19:56  <j3camero> We used to use a mailing list but switched to this so that everyone could see the discussion.
2011-01-29T17:20:07  <j3camero> Hang on I'll enable you to post in that section.
2011-01-29T17:25:12  <j3camero> Okay you're in.
2011-01-29T17:25:59  <j3camero> Things are somewhat stalled at this point, and that's mostly my fault. I really have to pick a specific game and an approximate release date to give people direction.
2011-01-29T17:26:47  <j3camero> I am thinking multiplayer ants (aka zombies) to release within one or two months?
2011-01-29T17:27:39  <jbroman> j3camero: I see you've mentioned the in-browser coding idea again. If you're still keen on that, it would probably worth doing a survey or something to determine how many people would use it.
2011-01-29T17:28:02  <j3camero> I think I actually know the result of that survey: 0 :-)
2011-01-29T17:28:22  <j3camero> It would definitely be for the newer users, not the seasoned ones.
2011-01-29T17:28:53  <j3camero> People with absolute zero programming experience but who "have always thought it would be neat to maybe try"
2011-01-29T17:29:30  <jbroman> j3camero: I don't think the IDE is the barrier to entry, really. Installing the IDE is the easy part – getting past "Hello, world" is the trickier part.
2011-01-29T17:30:18  <j3camero> I think you're probably right. I don't attach a lot of weight to the in-browser IDE idea right now. I think that there are higher barriers to entry that still remain, as you say.
2011-01-29T17:30:54  <Zannick> j3camero: if you're really concerned about changes being accepted before being pulled, you could make a branch
2011-01-29T17:31:39  <j3camero> Zannick: by the way thanks for being awesome and contributing changes. You inspire me to actually start doing work.
2011-01-29T17:31:51  <Zannick> j3camero: heh, thanks
2011-01-29T17:32:25  <j3camero> Zannick: okay I may start using branches for my changes if that's the general way to do things. I've just never been in a large project that uses git before, and am a general git n00b.
2011-01-29T17:32:31  <Zannick> i was mainly working on the next next thing
2011-01-29T17:32:56  <j3camero> Haha yeah. "Zeta"
2011-01-29T17:33:01  <Zannick> j3camero: i'm mostly a git noob, i just use some docs often
2011-01-29T17:33:25  <jbroman> j3camero: How you use git branches is up to you. Local branches are essentially free. Most people seem to develop small changes on master, large changes on feature branches (in which case you should run "git rebase master" periodically to remain up-to-date, and do "git merge" when you're done)
2011-01-29T17:33:34  <amstan> j3camero: yo
2011-01-29T17:33:38  <jbroman> And I think the set of non-git-noobs is {Linus Torvalds}.
2011-01-29T17:33:45  <Zannick> haha
2011-01-29T17:33:57  <j3camero> Yeah I get that feeling haha.
2011-01-29T17:34:22  <j3camero> I used to use darcs with b4taylor and that was great. So much less scary than git!
2011-01-29T17:34:23  <amstan> j3camero: i was talking to sayed, he didn't recomend the vm idea at all
2011-01-29T17:34:37  <j3camero> amstan: oh? Why not?
2011-01-29T17:34:42  <amstan> j3camero: apparently it takes too long to load it
2011-01-29T17:34:49  <amstan> even if you do fix that networking problem
2011-01-29T17:35:08  <amstan> j3camero: also.. can we please get this going already?
2011-01-29T17:35:13  <j3camero> Oh other sayed. Yeah, using VMs for security is a fucking bottomless hole.
2011-01-29T17:35:14  <jbroman> amstan: If you're referring to the conversation I overheard, I think he said that using a VM for each bot was a bad idea, not the idea of a VM for running the contest environment in general.
2011-01-29T17:35:15  <Zannick> i was actually looking around for a container architecture to replace the crazy system used in user_sadbox
2011-01-29T17:35:20  <Zannick> but it was very daunting
2011-01-29T17:35:57  <jbroman> I think using the basic limits.conf stuff to restrict CPU/mem/etc, plus locking out networking (via iptables UID filtering) is probably enough.
2011-01-29T17:36:15  <Zannick> ah, i may not be familiar with that
2011-01-29T17:36:16  <j3camero> Zannick: if you build a better way to do sandboxing, you would be my hero. I sank weeks into that problem before Galcon.
2011-01-29T17:37:07  <jbroman> Ultimately, if people try really hard to be shady, they can be. I don't think Fort Knox need be the objective.
2011-01-29T17:37:15  <j3camero> But yes, jbroman makes a good point. The approach we used in Galcon (running the bots as an unpriviledged user with limits.conf to restrict what they can do) actualyl worked pretty well. It surprised me.
2011-01-29T17:37:55  <amstan> j3camero: the limits+good security practices+backups will probably be enough
2011-01-29T17:38:05  <j3camero> I think we should just go ahead and use that for the next contest. If it aint broke dont fix it.
2011-01-29T17:38:21  <Zannick> yeah, i was looking at os-level virtualization like openvz or lxc
2011-01-29T17:38:27  <Zannick> probably overkill
2011-01-29T17:38:34  <Zannick> and they seem to require kernel patching
2011-01-29T17:38:49  <j3camero> Yeah so many of these things require kernel patches, which is pretty much out for us.
2011-01-29T17:39:03  <j3camero> It makes automated deployment of game-playing servers so hard.
2011-01-29T17:39:10  <Zannick> yeah
2011-01-29T17:39:17  <Zannick> makes testing hard, too
2011-01-29T17:39:24  <j3camero> amstan: yes I agree. Let's get this thing going already.
2011-01-29T17:40:25  <j3camero> amstan: I volunteer to start writing a Zombies engine if people are cool with that.
2011-01-29T17:40:25  <Zannick> hm, then i guess i know my next steps for the zeta backend
2011-01-29T17:40:38  <amstan> j3camero: what's zombies?
2011-01-29T17:40:39  <janzert> a chroot environment to put the bots into would be nice although someone needs to do some work to set it up properly
2011-01-29T17:41:30  <j3camero> janzert: not sure about this, but I believe the problem with chroot is that it's not feasible to set up a new one every time we launch a bot.
2011-01-29T17:41:34  <Zannick> that's a lot of work
2011-01-29T17:42:03  <j3camero> amstan: zombies is just the ants idea that I developed with you.
2011-01-29T17:42:18  <j3camero> But someone suggested that we rebrand it as zombies and I agree :-)
2011-01-29T17:42:26  <jbroman> I wouldn't even bother with chroot. I don't think it gets us much that distinct Unix users don't already get us.
2011-01-29T17:42:27  <Zannick> with proper permissions, we can lock the users out of /home and they can't get in to anyone else's
2011-01-29T17:42:33  <j3camero> As soon as we rename it from ants to Zombies, participation will double :-)
2011-01-29T17:42:47  <jbroman> j3camero: Rename it to sex.
2011-01-29T17:43:04  <j3camero> Google Sex Challenge. We have a winner.
2011-01-29T17:43:06  <janzert> so long as the permissions are properly locked down should just need to clear out the home directory like we currently do, not completely recreate it
2011-01-29T17:44:18  <janzert> which reminds me, the code from the last contest does *not* clear out the home directory correctly
2011-01-29T17:44:23  <j3camero> janzert: yes you're right.
2011-01-29T17:44:25  <janzert> that should be fixed for sure
2011-01-29T17:45:44  <janzert> currently .* files are left untouched, including any added rogue ssh keys the bot might add
2011-01-29T17:46:22  <janzert> deleting the directory itself and replacing it with a pristine copy is probably the best thing to do
2011-01-29T17:46:31  <Zannick> janzert: i was about to suggest that
2011-01-29T17:46:44  <j3camero> So in order to kick start development, I think it would be good to create a script that deploys an entire working contest instance. That includes apache, database, DB schema, and a single game worker, all on the same machine. That way it would take only a moment to kick off an entire instance inside a VM. That would be a huge mental help for me in developing features and such, I don't know about other people.
2011-01-29T17:47:21  <Zannick> yes, it's a good idea
2011-01-29T17:47:47  <jbroman> j3camero: It should also be possible to deploy Web-only and worker-only machines, as this separation makes sense from a security perspective.
2011-01-29T17:47:52  <j3camero> Okay I may try that now. I have a few hours before I have to go out tonight, so I'll take a crack at it.
2011-01-29T17:48:26  <j3camero> jbroman: for sure. Those would also be handy when we're managing the actual production deployment.
2011-01-29T17:48:38  <Zannick> just make sure you don't check any passwords into git :)
2011-01-29T17:48:51  <j3camero> There is already such a script for game-only machines. I believe danielvf created it.
2011-01-29T17:51:25  <janzert> I haven't looked into it much, but it may be interesting to look at something like Fabric for deployment http://docs.fabfile.org/0.9.3/
2011-01-29T17:52:21  <janzert> anyway have fun with development :) I gotta run
2011-01-29T17:53:34  <amstan> ugh..
2011-01-29T17:53:47  <amstan> internets are now capped at 25GB in ontario
2011-01-29T17:53:51  <j3camero> That Fabric looks pretty sexy.
2011-01-29T17:53:53  <amstan> for bell dsl
2011-01-29T17:54:04  <jbroman> Looks like Capistrano for Python. Which is good.
2011-01-29T17:54:30  <j3camero> jbroman do you have an interest in automated deployment?
2011-01-29T17:54:51  <j3camero> If you want to take this one, I will start on the actual engine. Otherwise, I will take it no worries.
2011-01-29T17:55:07  <jbroman> j3camero: It's an interest. Ultimately, whatever we come up with should be rolled into a nice AMI so that we can bring up n instances with epsilon effort.
2011-01-29T17:56:30  <j3camero> Oh for sure.
2011-01-29T17:57:04  <j3camero> A game server AMI and a web server AMI, for example. So we can launch n instances of either easily.
2011-01-29T17:57:14  <j3camero> That's how we did the game servers last term actually.
2011-01-29T17:57:55  <j3camero> There was a game server AMI which would wake up and execute "wget startup_sript; ./startup_script".
2011-01-29T18:00:49  <amstan> Zannick: i was thinking having the backend work with the django models as well
2011-01-29T18:01:17  <Zannick> amstan: i'm not sure what you mean?
2011-01-29T18:01:39  <Zannick> i've been designing the backend so it's self-contained
2011-01-29T18:01:53  <Zannick> but so things can call into it
2011-01-29T18:02:13  <amstan> Zannick: yeah, but the backend will use the same db and schema as the front end
2011-01-29T18:02:25  <Zannick> basically so it can be run without running the whole game server, which last time needed server_info and MySQLdb
2011-01-29T18:02:40  <Zannick> so i was going back and forth between remarking out those lines to test the scripts
2011-01-29T18:02:40  <amstan> Zannick: so we could write the models for django and use them for the backend as well
2011-01-29T18:03:11  <Zannick> amstan: in short, i haven't excluded that, just moved it external to the backend scripts :)
2011-01-29T18:03:31  <amstan> well.. i was thinking about having the models, as common code, both for frontend and backend
2011-01-29T18:04:01  <amstan> Zannick: https://github.com/amstan/aichallenge/blob/master/django/contest/models.py
2011-01-29T18:04:32  <j3camero> I think the suggestion is that we kind of do the backend in Django, or have some Django code in the backend. Not sure if that's a good idea.
2011-01-29T18:05:33  <amstan> it's according to this: http://docs.djangoproject.com/en/1.2/misc/design-philosophies/#dry
2011-01-29T18:07:32  <Zannick> it's something that could be looked into
2011-01-29T18:08:16  *** pyro- <pyro-!~pyro@unaffiliated/purplepanda> has joined #aichallenge
2011-01-29T18:08:44  <j3camero> I don't think that's what DRY really means.
2011-01-29T18:08:47  <Zannick> again, i'm trying to make it so that it doesn't rely on the existence of other parts of the contest
2011-01-29T18:08:51  <amstan> having everything based on the django models is also beneficial if we decide to switch from mysql
2011-01-29T18:09:04  <j3camero> The big negative is that our backend is all tied up with a particular web framework.
2011-01-29T18:10:19  <Zannick> ah, so "don't repeat yourself" is good, but i've learned through experience that trying very hard to not repeat yourself can also be bad
2011-01-29T18:10:31  <Zannick> eg. uniting two different things because they do similar things
2011-01-29T18:11:05  <j3camero> I think that Dont Repeat Yourself just refers to DB normalization.
2011-01-29T18:11:39  <Zannick> it's probably better to have a middleman between the frontend and backend
2011-01-29T18:13:18  <jbroman> Why does the backend need to know anything about MySQL?
2011-01-29T18:13:32  <jbroman> Why not just send it messages that say "here, run this and this and report back"?
2011-01-29T18:13:57  *** amstan_ <amstan_!~alex@206-248-153-79.dsl.teksavvy.com> has joined #aichallenge
2011-01-29T18:13:57  *** ChanServ sets mode: +o amstan_
2011-01-29T18:14:02  <Zannick> that's my philosophy :)
2011-01-29T18:14:37  <amstan> jbroman: who sends it such messages?
2011-01-29T18:14:48  <amstan> jbroman: that sender still has to know about db stuff
2011-01-29T18:15:18  <Zannick> the middleman or the frontend
2011-01-29T18:15:41  <amstan> you can't have the frontend send messages such as play these 2 ppl
2011-01-29T18:15:53  <amstan> and the middleman.. if it does that.. it's pretty much the backend
2011-01-29T18:16:25  <Zannick> i think we're talking about different backend/frontend definitions here
2011-01-29T18:16:26  <jbroman> amstan: Makes the backend layer thinner and cleaner, IMO.
2011-01-29T18:17:13  <amstan> well, ok, let's define things so we're not getting confused
2011-01-29T18:17:24  <Zannick> there are probably actually three separate parts around the database
2011-01-29T18:17:39  <amstan> frontend is anything related to the site
2011-01-29T18:18:02  <amstan> then you have the backend.. which is sandbox, compiling engine and so on
2011-01-29T18:18:12  <Zannick> the website, the contest engine (which i was calling the frontend, sorry), and what runs the games and compilation and whatnot (which is what i called the backend)
2011-01-29T18:18:15  <amstan> where does the tournament engine sit in here? doesn't it need to have db access?
2011-01-29T18:19:07  <Zannick> so you're right; what you were calling the backend needs to know about the database, but that part is a layer above the actual game engines and compiler which doesn't
2011-01-29T18:19:43  <Zannick> sorry for the confusion :/
2011-01-29T18:20:15  <amstan> Zannick: yes
2011-01-29T18:20:16  <jbroman> The terms backend/frontend are confusing. I think we all agree that there are three entities: the website, the tournament engine (or whatever it is that decides who plays whom), and the worker. The former two need a communication channel (which is probably the database), and the latter two need a communication channel (hopefully not the database).
2011-01-29T18:20:38  <amstan> +1 for jbroman's terms
2011-01-29T18:20:50  <Zannick> ah, yes, worker.
2011-01-29T18:20:51  <Zannick> +1
2011-01-29T18:21:11  <amstan> so my suggestion is that the tournament engine layer, in order to communicate easier to the database, should use the django model definitions as well, and django itself to connect to the db
2011-01-29T18:21:20  <j3camero> Well hang on. The code that picks matches and answers calls from game servers who are ready to play games is now actually part of the website. So it's actually totally legit to use Django for that.
2011-01-29T18:22:15  <jbroman> Tournament engine using Django ORM sounds fine. The model code is probably thin enough that it wouldn't really matter if we used other ORM, but since it's Python, Django ORM makes sense.
2011-01-29T18:23:16  <amstan> for anyone who asks ORM=Object-relational mapping
2011-01-29T18:26:31  <Zannick> i will likely move that backend folder i made to worker, then
2011-01-29T18:26:54  <Zannick> i have to duck out for a bit, be back later
2011-01-29T18:28:14  <j3camero> Sounds like a plan to me.
2011-01-29T18:28:48  <j3camero> Then game workers don't need to know anything about the DB. When they want to play a game, they just hit the relevant URL on the main site, which hands them back the players for the game.
2011-01-29T18:29:33  <j3camero> Do we want to go with a single web server this time? That has to be decided before we make this design decision.
2011-01-29T18:30:08  <j3camero> If we're going to have more than one web server, then we have to disentangle the DB from the web server.
2011-01-29T18:30:21  <amstan> j3camero: that's easy though
2011-01-29T18:30:34  <amstan> it's not like the main site reads the db files directly from disk
2011-01-29T18:30:49  <j3camero> True enough yeah.
2011-01-29T18:30:58  <amstan> relevant: http://stackoverflow.com/questions/886221/does-django-scale
2011-01-29T18:33:29  <amstan> oh lol, bitbucket runs django
2011-01-29T18:33:42  <j3camero> I'm sure it's better than what we have now.
2011-01-29T18:35:18  <jmcarthur> so wait, is the idea we're discussing to have the matchup algorithm be exposed as an http api?
2011-01-29T18:35:29  <j3camero> Right.
2011-01-29T18:35:34  <j3camero> That's how it works now actually.
2011-01-29T18:35:38  <jmcarthur> and also the game results?
2011-01-29T18:35:45  <jmcarthur> POST, i mean?
2011-01-29T18:36:00  <jmcarthur> ah, i wasn't aware of that
2011-01-29T18:36:04  <jmcarthur> well that's cool
2011-01-29T18:36:06  <j3camero> It would probably be a good idea for the game results as well. That is also how it works now by the way.
2011-01-29T18:36:49  <j3camero> Although the best thing would be to create a torrent for the daily game playback data. That will stop crawlers from hammering the website.
2011-01-29T18:36:56  <jmcarthur> totally
2011-01-29T18:37:02  <jmcarthur> but that's not what i meant
2011-01-29T18:37:10  <j3camero> Woah, a legitimate use for BitTorrent.
2011-01-29T18:37:21  <jmcarthur> i was talking about the tournament manager. is the idea to push game results to the db via the web service?
2011-01-29T18:37:30  <j3camero> Oh yes.
2011-01-29T18:37:35  *** choas <choas!~lars@p578F6859.dip.t-dialin.net> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T18:37:40  <j3camero> However, you need an API key to do so, so it's cool.
2011-01-29T18:37:43  <amstan> j3camero: i don't think there's enough interest for torrents to stay alive
2011-01-29T18:37:45  <jmcarthur> right right
2011-01-29T18:37:56  <jmcarthur> amstan: i think there would be enough interest for a weekly dump
2011-01-29T18:37:59  <jmcarthur> more than enough
2011-01-29T18:38:12  <amstan> weekly might be too rare for the people that need it
2011-01-29T18:38:15  <jmcarthur> i and several others wanted it
2011-01-29T18:38:20  <jmcarthur> i don't think so
2011-01-29T18:38:31  <jmcarthur> things aren't likely to change quickly enough that daily would be necessary
2011-01-29T18:38:38  <j3camero> Weekly may be enough to fend off the crawlers.
2011-01-29T18:39:13  <j3camero> And we could encourage people to grab it on the IRC channel so that the torrent doesn't die.
2011-01-29T18:39:22  <jmcarthur> i would gladly seed
2011-01-29T18:39:24  <j3camero> Just to take the burden off the main web server.
2011-01-29T18:39:37  <j3camero> I would seed as well. I have infinity fibre internets.
2011-01-29T18:39:46  *** chris__0076 <chris__0076!~Chris_007@adsl-176-224-170.asm.bellsouth.net> has joined #aichallenge
2011-01-29T18:40:16  <amstan> cool
2011-01-29T18:40:31  <amstan> j3camero: so what about the contest that we're about to launch?
2011-01-29T18:40:50  <amstan> j3camero: what game, what do we need to change in the old system for more stuff?
2011-01-29T18:42:19  *** chris___0076 <chris___0076!~Chris_007@adsl-176-227-147.asm.bellsouth.net> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T18:42:49  <j3camero> I am thinking we just want to pick a game and go for it. It seems like things are stuck on the choice of game. So let's just say it's the simplified ants variant that I created, branded as zombies.
2011-01-29T18:43:10  <j3camero> Multiplayer and all. I think we have enough devs that we can pull it off in a reasonable amount of time.
2011-01-29T18:43:31  <amstan> multiplayer? lol
2011-01-29T18:43:43  <amstan> that changes a lot of stuff though
2011-01-29T18:46:59  <j3camero> Yep.
2011-01-29T18:48:07  <j3camero> dmj claims to have a trueskill implementation that may work for zombies.
2011-01-29T18:59:24  <j3camero> Heading out to the gym, see y'all later.
2011-01-29T19:00:10  *** sigh <sigh!~sigh@28.250-50-210.dynamic.dsl.mel.iprimus.net.au> has joined #aichallenge
2011-01-29T19:05:54  <CIA-37> aichallenge: Benjamin S Wolf master * ra5b1112 / (6 files in 2 dirs): Rename backend directory to worker/. - http://bit.ly/fS4xtP
2011-01-29T19:05:55  <CIA-37> aichallenge: Benjamin S Wolf master * r15689ee / worker/compiler.py : Some previously uncommitted comment changes. - http://bit.ly/es7Zxs
2011-01-29T19:16:00  *** Migi32 <Migi32!~migi@178-117-104-154.access.telenet.be> has quit IRC (Quit: boinc)
2011-01-29T19:16:58  <jbroman> amstan: Any particular reason the Django project is called "django"? I can't even start it when I pull it, and I think name conflicts may be involved.
2011-01-29T19:17:36  <jbroman> (Where "think" = "I renamed it notdjango and it started")
2011-01-29T19:17:44  <amstan> hmm
2011-01-29T19:17:52  <amstan> let me try, i haven't touched it in a week or so
2011-01-29T19:19:14  <jbroman> Django projects have a namespace that starts with their name, so using "django" as the project name seems like a BadIdea™. But  I'm hardly a Django guru.
2011-01-29T19:19:48  <amstan> jbroman: i probably pushed after i played with it
2011-01-29T19:21:27  <amstan> jbroman: you're right.. k, i need a semantic name for this django related stuff
2011-01-29T19:22:10  <jbroman> aichallenge_website or similar?
2011-01-29T19:22:22  <amstan> or frontend
2011-01-29T19:22:34  <amstan> yeah, we'll say frontend for now
2011-01-29T19:26:12  <CIA-37> aichallenge: Alexandru Stan master * rd5fd60b / (20 files in 4 dirs): moved django to frontend, some name conflicts was making it impossible to start it, thanks jbroman for pointing it out - http://bit.ly/hiSIcE
2011-01-29T19:28:27  <amstan> why did Zannick call it the "zeta backend"? is that the name for the next gen contest(one after next)?
2011-01-29T19:29:38  <jbroman> That was the title someone (j3camero, maybe?) used in the forums.
2011-01-29T19:29:58  <Zannick> j3camero decided to call the far future "zeta"
2011-01-29T19:30:08  <amstan> oh.. ok, decent
2011-01-29T19:30:13  <Zannick> future contest, that is
2011-01-29T19:30:34  <janzert> wasn't there just a conversation about the ambiguity of using the terms frontend/backend :P
2011-01-29T19:31:03  <Zannick> that was referencing a comment from before that conversation
2011-01-29T19:31:04  <amstan> janzert: yeah, i'm about to put this in a readme file
2011-01-29T19:31:53  <janzert> ahh, I thought the result was to call it website/tournament manager/worker.
2011-01-29T19:32:11  <janzert> You can ignore me again, I'm only half paying attention :}
2011-01-29T19:32:19  <amstan> it's frontend - backend - worker
2011-01-29T19:32:22  <Zannick> janzert: that's why i moved the folder i had previously named backend/ to worker/
2011-01-29T19:32:35  <amstan> or.. tournament manager? lol
2011-01-29T19:32:46  <amstan> well.. the important this is that we have them defined
2011-01-29T19:32:46  <Zannick> website, manager, worker
2011-01-29T19:33:50  <amstan> ok
2011-01-29T19:33:52  <jbroman> Yeah, I like those names.
2011-01-29T19:35:52  <jbroman> As I said before, I think the terms "frontend" and "backend" are ambiguous.
2011-01-29T19:37:10  <Zannick> hm, so if the website and the manager both use django, then the django directory has to be above both?
2011-01-29T19:37:13  <Zannick> or should be?
2011-01-29T19:38:32  <amstan> Zannick: we'll move it later
2011-01-29T19:38:38  <Zannick> yeah
2011-01-29T19:38:40  <janzert> manager will just have a dependency on the website
2011-01-29T19:38:42  <amstan> Zannick: but you're right.. there's some common stuff for both the website and the manager
2011-01-29T19:39:21  <jbroman> Ideally, I would think that the models "belong" to the manager. But either way, you can probably bring the models from one into the other via appropriate imports.
2011-01-29T19:39:26  <Zannick> i'm not sure how django works and i'm not going to claim to know how that section should be organized
2011-01-29T19:39:28  <CIA-37> aichallenge: Alexandru Stan master * rd40b326 / (20 files in 4 dirs): added readme to explain the parts of the repo, renamed frontend/ to website/ - http://bit.ly/fPmnJK
2011-01-29T19:39:29  <amstan> there we go
2011-01-29T19:40:25  <amstan> fuck.. ugh..
2011-01-29T19:40:43  <CIA-37> aichallenge: Alexandru Stan master * r3ad4b60 / (README manager/.gitignore): really added readme - http://bit.ly/f7IQj1
2011-01-29T19:41:21  <Zannick> :)
2011-01-29T19:44:42  <amstan> anyway.. so there's not much in the django stuff yet
2011-01-29T19:45:09  <amstan> i was just getting a small thing going
2011-01-29T19:51:13  <amstan> j3camero: what about the wiki?
2011-01-29T19:51:27  <amstan> j3camero: interested in moving pages over?
2011-01-29T20:02:52  <amstan> actually.. hmm
2011-01-29T20:02:55  <amstan> they're in svn on there
2011-01-29T20:02:59  <amstan> and they're in git on the new one
2011-01-29T20:03:03  * amstan is converting them now
2011-01-29T20:10:08  <amstan> well.. they're in there, with proper histories: https://github.com/amstan/aichallenge/wiki/_pages
2011-01-29T20:35:08  *** Mathnerd314 <Mathnerd314!~mathnerd3@dsl.6.28.53.206.cos.dyn.pcisys.net> has joined #aichallenge
2011-01-29T20:49:16  <CIA-37> aichallenge: Alexandru Stan master * r932cdc7 / (README.md README): apparently the readme file needs to have a .md at the end to display nicely in github - http://bit.ly/fuqkGz
2011-01-29T20:57:55  *** AlliedEnvy <AlliedEnvy!~AlliedEnv@adsl-68-90-63-136.dsl.hstntx.swbell.net> has joined #aichallenge
2011-01-29T20:59:16  *** delt0r_ <delt0r_!~delt0r@188-22-164-235.adsl.highway.telekom.at> has joined #aichallenge
2011-01-29T21:00:32  *** delt0r___ <delt0r___!~delt0r@188-22-166-9.adsl.highway.telekom.at> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T21:03:01  <amstan> jbroman: do you want commit access?
2011-01-29T22:00:19  *** Kingpin13 <Kingpin13!~Kingpin13@Wikipedia/Kingpin13> has quit IRC (Quit: You're a kitty!)
2011-01-29T22:25:59  <jbroman> amstan: Sure, why not? jeremyroman on github
2011-01-29T22:26:37  <amstan> you can now merge your changes
2011-01-29T22:28:21  *** smellyhippy <smellyhippy!~asmodean@about/csharp/regular/smellyhippy> has quit IRC (Ping timeout: 240 seconds)
2011-01-29T22:30:45  <CIA-37> aichallenge: Jeremy Roman master * rab6189d / (4 files in 2 dirs): adjust readme, create root __init__.py so that the repo is a Python package, fix sample settings.py, adjust .gitignore - http://bit.ly/ezO2QI
2011-01-29T22:33:34  *** smellyhippy <smellyhippy!~asmodean@cust240-dsl50.idnet.net> has joined #aichallenge
2011-01-29T22:33:44  *** smellyhippy <smellyhippy!~asmodean@about/csharp/regular/smellyhippy> has joined #aichallenge
2011-01-29T23:04:54  <amstan> jbroman: you should get a gravatar
2011-01-29T23:05:23  <jbroman> amstan: I suppose.
2011-01-29T23:21:46  *** mega1 <mega1!~quassel@2001:470:1f05:548:215:58ff:fe7d:773b> has quit IRC (Read error: Operation timed out)