Friday, 2008-03-28

*** tpb has joined #tp00:00
*** ChanServ sets mode: +o tpb00:00
mithroI have a very important announcement00:00
bddebianUh oh00:00
vi1985andrei: thanks. Would be interesting to look at its reasoning, although modern top-notch engines must use different tools.00:00
vi1985*meant to say, reasoning tools00:01
vi1985*drums?*00:01
andreivi1985, I'm not so sure. Chess kind of stopped being an area of a lot of research once the AI community got to its senses.00:01
mithroCan you please all add "functional milestones" - these should be a single sentences which say something like "The AI *will* issue move orders" or "The client *will* be able to display planets in a 3d fashion" - The key word here is ****will**** it's assertive and they should have dates associated with them00:02
Nulerenmithro is a master of suspense ;)00:02
mithrothey should all be testable too00:02
andreimithro, Are mine ok now? :)00:02
mithroIE "The framework will be documented" != "functional milestones"00:02
mithroandrei: they are getting there - I think they still need to be more assertive and more functional IE "Find me a path to go to this location, though strategies have not yet been implemented and this would be exceedingly slow." -> "The AI will be able to find a safest path between two systems when there are multiple paths"00:05
mithroor something along those lines00:05
andreimithro, I am weary of making the safest assertion00:06
mithroall dates must be in ISO format IE 2007-01-1200:06
vi1985mithro: Did you guys have problems in past years with wishy-washy timeframes, and unfulfilled promises? :)00:06
andreimithro, Heh, seriously? :P00:06
SmokingRopehow about something like 'star map finished'?00:06
mithroandrei: well it' makes it easier to look at00:06
mithroSmokingRope: be more specific and include what the player can see00:06
andreimithro, I can't promise it'll the safest path. I can promise it'll find a good path; for some definition of good. But don't turst anyone that makes bold claims like 'the safest'00:06
mithroandrei: if you include a definition of safest you'll have no problems00:07
andreimithro, Because safest can mean too many things. In different contexts. And exactly what it means for this AI will vary widely depending on which modules are active00:07
vi1985andrei: I agree that it's hard to talk about very specific things at this stage.00:07
* llnz wanders home00:08
*** llnz has quit IRC00:08
mithroit's understandable that it's hard to talk about specific things, but being able to talk about specifics shows that you are well researched00:08
mithroand have a good understanding of the task00:08
andreimithro, The idea is to implement A* to expand the tree and look for safe paths employing other modules to determine what areas are safe or not. I'd have great difficulty saying exactly what safe means at any point to the AI. For example if a module is active that talks about fuel, safest changes radically. Etc.00:09
mithroSmokingRope: IE "star map is finished" -> "The starmap will display system as glowing objects in the correct locations"00:10
mithros/system/systems/00:10
vi1985mithro: that is definitely true. I can see that from your side that is what you want to see in good applicants. On my side, even though I might have the steps in my head, I'm reluctant to state minute-by-minute progress. I'll do my best.00:10
andreimithro, The path algorithm refers to the fact that A* exists and can use different measures of safety. That's the hard and important bit. Measuring safety in a lot of different ways is easy.00:10
mithrowe understand that the goals will change and the dates will change00:11
mithroandrei: "The AI will be able to find the shortest path using an A* algorithm only considering distance between two stars"00:11
SmokingRopemithro: that works for me, ty :)00:12
mithroSmokingRope: "The client will allow a user to submit a move order by clicking on the starmap"00:12
mithroSmokingRope: etc00:12
andreimithro, Ah, that's not quite how A* works though :) because you can use any weight algorithm for it. So sure, there will be one algorithm that talks about neighbour systems, but that's just part of the weight algorithm for A* :)00:12
vi1985mithro: ok, I'm glad to hear that. I'll elaborate much more when morning comes to Canada :)00:13
mithroandrei: choose a specific testible case and state the expected result00:13
andreimithro, Is it ok that the testable case is a lowerbound on what the AI can do?00:13
mithroandrei: yes - but your lowerbound must be sufficient that we are satisfied with the result00:14
andreimithro, Okies :)00:14
brennan_hello all00:15
* vi1985 *quietly wanders off to dream about TP AI enslaving the human race*00:15
*** vi1985 has quit IRC00:15
mithroIE "The AI will be able to plot a path between to systems which avoids enemies"00:15
mithroIE "The AI will adjust a plotted path to continue to avoid enemies based on predicted enemies positions"00:16
mithroassertive people :)00:16
mithrohave confidence in your own abilities00:17
mithrohey brennan_00:17
brennan_hi mithro00:17
mithrobrennan_: recommend reading the scrollback :)00:19
mithroI might put together an email and send it out00:20
mithrofor those on a less pleasent timezome00:20
brennan_lol00:23
brennan_thats alot of chat lol00:24
Ohmargh, I have no page-up key on this keyboard00:24
Ohmah, I do have logs though00:24
brennan_Mithro: did you get my pm from the forums? i know llnz looked at my proposal on the wiki, have you?00:25
mithrobrennan_: yeah I did00:25
mithroOhm: so do we00:25
mithrobrennan_: you didn't link to it from your GSoC application?00:26
brennan_no, i guess i should00:26
brennan_lol00:26
brennan_ok, i linked it, and copied over the newer version00:30
mithrobrennan_: okay you still need some type of timeline with dates00:31
brennan_ok00:31
mithroyou also need to include the milestones mentioned above00:31
mithrobrennan_: take a look at the timeline section of http://csclub.uwaterloo.ca/~abarbu/soc/thousand-parsec00:33
tpb<http://ln-s.net/1iXj> (at csclub.uwaterloo.ca)00:33
brennan_ok00:34
brennan_i actually had another idea, because last night i was reading on here about a mobile client idea, do you think a version of the metaserver for mobile browsers would be a good addition to my proposal?00:36
andreimithro, In the date format you want, is it dd-mm or mm-dd (I'm in the US and which format someone uses is a big problem here)00:37
mithroandrei: I want ISO YYYY-MM-DD00:37
andreiExcellent :)00:37
andreiThanks00:38
*** f0ster has quit IRC00:38
brennan_mithro: should i drop my schedule from the proposal and just have the timeline?00:39
andreimithro, Done :) (and I expanded upon the goals in the schedule00:42
andrei)00:42
mithroandrei: now I have some more specific questions00:43
mithroI have no idea what this means00:43
mithro    1. You should be able to pick 3 different AI sessions, with00:43
mithro     different settings (or have them be different AIs entirely) and00:43
mithro     just run './go' and see how they fared.00:43
*** bostonvaulte1 has joined #tp00:43
andreimidorikid, Ah, as in, you want to test some AIs out. There will be a script to do so easily and to see how well they did against eachother00:46
andreierm00:46
andreimithro,  that was for you00:46
mithroandrei: can you rewrite it to say that :P00:46
andreiSure :)00:46
andreiDone :)00:48
*** bddebian has quit IRC00:50
*** Erroneous has quit IRC00:53
*** Nuleren has quit IRC00:54
mithroandrei: "Move validation, no wrong moves should occur" negatives are hard to test :P00:55
mithroxdotx: ping?00:57
andreimithro, Not really. There are only so many moves that are possible. I plan to just exaustivly try all of them and make the client crash spectacularly if any happen :)00:57
mithroandrei: you do realise we have free-form space in many rulesets00:59
andreimithro, Actual client moves are only done by a few select very low level modules. So those are easy to try out (since they involve no real logic)00:59
andreimithro, Yup. I can't prove that incorrect moves can't happen. I can show that they will be extremely unlikely by trying out enough scenarios01:00
mithrofrom the email I am writing01:00
andreimithro, There is actually a way to show it by hypothesis testing.. but I haven't the time to implement that01:00
mithroNegative milestones are hard to test, avoid them if possible. For example "The AI will not issue wrong orders" is almost impossible to prove and not all that useful.01:01
andreimithro, My plan will essentially guarantee that wrong moves will not occur. I can change that to 'should' if you prefer01:04
andreimithro, or remove it01:04
mithroandrei: it's not a milestone01:05
mithroit's a goal01:05
andreimithro, Heh, okies01:05
andreiI removed it :)01:06
mithrohow many characters does google limit the detailed description too?01:14
andrei7.5k I think01:16
brennan_yeah01:16
brennan_abstract is 2.5k, detailed is 7.501:16
andreiI'm really curious to see what posessed them to put those limits01:21
andreiThey're tiny.. 7.5k is half my proposal01:21
brennan_lol01:27
mithroI guess quality is better then quantity :)01:27
andreimithro, What are you trying to say? :P01:27
andreiHeh01:27
mithrounlike many things in the world they are not mutually exclusive01:28
andreiI know, I'm just kidding :)01:28
mithroemail sent to tp-devel for a whole bunch of tips01:28
brennan_cool, i added 2 more things to my proposal that i can work on and my timelime is almost done01:30
brennan_which is good cus its nearing 2am here01:31
brennan_its a good thing these proposals arent fully done till monday, gives me more time to tweak it01:32
andreiI wouldn't count on this, but if previous years are any indication it won't be done by monday01:35
mithroandrei: I would definately not count on it01:36
mithrohello bostonvaulte101:37
mithroI don't think I have seen you around here before01:37
mithrohello Arc01:37
mithroI don't think I have seen you around here before either01:37
brennan_well, i didnt mean it will be done, but isnt the detailed description part of the proposal locked on monday?01:37
mithrobrennan_: possibly01:38
mithros/possible/probably/01:38
brennan_lol01:38
brennan_k01:39
brennan_the writing is a bit unclear on the google app guide01:39
brennan_actually, yeah, i see, its editable if a mentor makes a public comment01:39
brennan_after the monday deadline01:40
brennan_w/e the case i plan to be working on it as long as there are flaws in it01:40
brennan_lol01:41
*** xdotx has quit IRC01:44
*** xdotx has joined #tp01:45
mithroxdotx: poke poke01:46
xdotxyo01:46
*** JLafont has quit IRC01:48
*** nash has quit IRC01:55
brennan_Mithro:Ok i changed and added a timeline to my proposal01:57
mithroPSF is up to 37 apps01:57
*** greywhind_ has joined #tp01:58
*** greywhind has quit IRC01:58
mithrobrennan_: I think you still need more details, some proof that you have read and understood the current metaserver would be good too02:01
brennan_ok02:01
mithroa mock up of the matchmaking service search page02:01
JLPgood morning all02:03
brennan_JLP: lol, as im about to get ready to say good night XD02:03
brennan_even after all these years of having an connection to the net, i still get amazed by that sort of stuff02:04
brennan_mithro: okies, ill work on it some more later on today, but i think i should turn in now, 2am local time here02:06
mithrobrennan_: that is fine02:06
mithrobrennan_: it's looking much better02:06
brennan_mithro: thanks, i guess just getting prodded from you guys has def helped02:07
brennan_okies, night all02:11
*** brennan_ has quit IRC02:11
*** andrei has quit IRC02:14
* JLP goes reading the article from draginol02:20
mithrodraginol?02:23
JLPmithro: galcov developer - http://forums.galciv2.com/30351202:26
tpbTitle: Piracy & PC Gaming ยป Forum Post by Draginol (at forums.galciv2.com)02:26
mithroJLP: ahh that is old :)02:27
JLPmithro: well ia haven't read it yet02:28
JLPnice 5 more applications02:31
*** ryan__ has joined #tp02:32
mithrohello ryan__02:34
mithroI don't think I have seen you around before02:34
BZA_mithro: so I went through my Project deliverables list and for each item, expanded a list of no less than f things the user can expect form it.  lemme know what you think please =)02:37
BZA_mithro: where f = 5 =/02:37
mithroBZA_: url? and did you get a chance to read the tips?02:38
BZA_mithro: http://www.thousandparsec.net/wiki/Google_Summer_of_Code/BZA_Proposal, and yeah, you do you mean your announcement above,or something on the Wiki?02:39
tpb<http://ln-s.net/1jy_> (at www.thousandparsec.net)02:39
mithroI mean the email to the mailing list02:41
mithrohttp://www.thousandparsec.net/tp/pipermail.php/tp-devel/2008-March/000862.html02:41
tpb<http://ln-s.net/1jyc> (at www.thousandparsec.net)02:41
BZA_no then,I didn't think to get on the mailing list =/02:42
mithroBZA_: we have mailing list archieves02:42
mithrohttp://dir.gmane.org/gmane.comp.games.tp.devel02:44
tpbTitle: Gmane -- Mail To News And Back Again (at dir.gmane.org)02:44
BZA_mithro: so, right now my functional milestones don't have dates associated with them, so I can go through and put that on the schedule.  I have not yet submitted my app to Google, so I will get on that tomorrow, and it seems like most of my functional milestones are accomplishable in a 4 week period, but I don't know enough yet to really assert that, so I will try and get more of the mentors...02:53
BZA_...to look it over and give feedback02:53
mithroBZA_: I would recommend submitting your app to Google now02:53
mithroyou will have the ability to change it02:53
BZA_yeah, ok02:54
*** zzorn_laptop has joined #tp03:18
*** midorikid has quit IRC03:26
*** llnz has joined #tp03:28
JLPllnz: ahoy03:29
llnzhi JLP03:29
mithrowell heading home now03:29
mithrosee ya!03:29
JLPdamn the sorting on google webapp really sux, is it random or what03:30
*** midorikid has joined #tp03:37
llnzJLP: consistent non-ordering03:37
JLP:)03:38
JLPmidorikid: ahoy03:38
midorikidHiyah!03:41
* JLP is now checking out BZA_'s proposal03:42
BZA_JLP: thanks =)03:43
*** mithro has quit IRC03:48
BZA_general question, for the Detailed Description section of the GSoC app, should I put the contents of my proposal page, or the contents of the design doc?03:50
JLPBZA_: id say proposal page03:57
BZA_JLP: mkay, thanks03:58
JLPBZA_: i think you could combine delivarables and schedule into one section03:58
BZA_JLP: I'll consider that when I make my next edit, thanks.  My only concern is that the one combined section becomes too cluttered, like one bullet point needs a paragraph explanation.  I do however intend on putting the smaller functional milestone into the schedule though04:01
JLPBZA_: also mention how you will inform general public about ypur progress (blog ro something)04:01
BZA_JLP: okay, I will give that some thought too04:02
JLPotherwise i think is is very good so after these edit you can just put it into google webapp and we can polish it from there04:03
JLPBZA_: oh yeah another thing, write a paragraph ro so about the possible problems and risks that could happen and how you plan to prevent that or minimize the effect of them04:04
BZA_JLP: good idea, will do04:05
* BZA_ collapses in bed to sleep04:16
JLPbtw people, if any of you are on last.fm you can join the TP group - http://www.last.fm/group/Thousand+Parsec04:16
tpbTitle: Thousand Parsec Groups at Last.fm (at www.last.fm)04:16
*** BZA_ has left #tp04:16
*** mithro has joined #tp04:21
llnzmidorikid: Nice proposal04:41
midorikidllnz: Hey, thanks!  (sorry I was AFK for a bit)04:46
llnzno problem04:46
*** bostonvaulte1 has left #tp04:57
*** midorikid has quit IRC05:40
* llnz wanders off05:55
llnzlater all05:55
*** llnz has quit IRC05:55
*** greywhind has joined #tp05:58
*** greywhind_ has quit IRC05:58
* JLP is writing to the maintainer of Stars! FAQ trying to make him replace the link to Stellar Legacy (which is dead) with a link to Thousand Parsec06:38
mithroJLP: he he06:57
JLPhm the email bounced, let's try with the e-mail from whois08:03
*** bddebian has joined #tp09:40
bddebianHowdy09:41
mithrohey bddebian09:42
mithrobeen a bit quiet in here tonight09:42
bddebianHi mithro09:43
bddebianThat's amazing, it's been pretty busy here lately :-)09:43
*** jphr has joined #tp09:50
jphrmithro: ping?09:50
mithrojphr: pong!09:50
jphrmithro: good, you're here. So you were wondering about my two nearly identical apps?09:51
mithroyeah - good for our numbers but a little confusing :)09:51
jphrmithro: remember when we talked about Risk turn structure in TP, how it would be very awkward to implement09:51
mithrojphr: I think I have had that discussion with multiple people now09:52
jphrI figured, if llnz was a glutton for punishment, or REALLY wanted traditional Risk, I would be willing to do it. The modified ruleset proposal is the 4X compromise, which I frankly feel will be a better game anyways09:52
jphrlol @ the discussions you have had09:53
mithroBZA_ has also put in a Risk application09:54
mithroit's going to be fscking hard to choose this year :(09:54
jphroh noes!09:54
jphrwhat are your numbers like?09:54
mithrowe have 18 apps so far09:54
jphrmany more sitting on the fence still?09:54
mithroI'm guessing so09:55
jphrDoes BZA_ have any thing on the wiki about his risk proposal? (gulp*)09:55
mithroyes he does09:56
mithrocheck the logs09:56
jphrand as an aside: if you really don't like my original proposal, you can ignore it. My modified proposal is the one I prefer09:56
jphrwill do09:56
jphryeesh, its pretty good09:58
*** greywhind has quit IRC09:58
*** greywhind has joined #tp09:58
jphrmithro: for functional milestones I have come up with a basic game without proper map, and the full game with map. Is that enough? or should I break it down more?10:07
mithrobeing more specific is better10:08
jphrok10:08
mithroa couple of days worth of work if a good milestone10:08
jphrshould i scrap the roadmap? or have the functional milestones complement the more general roadmap?10:09
mithrothey sould be part of the road map10:09
mithrothey are not tasks in them selves10:09
mithrothey are the results of the task10:09
jphrok. and while I am busy with non-coding task, those aren't functional?10:10
mithrothe results could be10:11
jphryou mentioned they should be testable when i looked over the logs. I would think only code would be testable, or am I wrong?10:12
jphrgotta run to class10:17
*** jphr has left #tp10:18
*** andrei has joined #tp10:31
*** protohack has joined #tp10:38
protohackhey, its Kevin Cherry from last night10:39
protohackI got a notice saying I needed a lot more info for my application. Maybe it is just me, but Google doesn't give a lot of info for what should be in the detailed section of the application.10:40
protohackFrom their webpage: Detailed Description: Enter the text of your application here. Text only, 7500 characters maximum.  Doesn't really tell you much.10:41
protohackIs anyone around?10:43
*** nuleren has joined #tp10:48
*** nuleren has joined #tp10:49
JLPprotohack: i'm here now10:49
protohackah cool10:49
protohackyeah, is it just me or does Google not give enough info for what to put in the detailed description area10:50
mithroprotohack: there is a lot of information linked from the Google site and our google page10:50
mithrocheck the AdviceForStudents web page10:50
protohackwhere at?10:50
protohackwhat is the url of the AdviceForStudents web page10:51
JLPprotohack: i think it is in the FAQ10:51
nulerenjust migrated my proposal to http://www.thousandparsec.net/wiki/Google_Summer_of_Code/Nuleren's_Propsal10:53
tpb<http://ln-s.net/1k69> (at www.thousandparsec.net)10:53
nulerenwould appreciate any feedback ;)10:53
nulerenand it looks like the link got screwed up... maybe a shouldn't put an apostrophe in there10:53
protohacklink seems to work for me10:54
nulerenok, nm then10:54
protohackAlthough for some reason I had to click it like 10 times but I think that is just icechat messing up10:54
protohackAccording to your proposal there is a lot more than what I thought goes into it. There might be information on TP's website but I swear I just don't see it and I looked at Google's student application help section and didn't get much help. I will keep looking on TP's website for more info10:56
JLPnuleren: checking it out10:56
nulerenJLP: thanks :)10:57
mithroprotohack: it's effecitively a proposal to get $4500 USD10:57
protohackIt sucks cause I got a message about GSoC from the undergrad advisor here at LSU on Monday when the application were being accepted. I have a midterm exam today that I had to worry about. And this weekend is kind of busy. But enough excuses, I will still try to make one10:57
mithronuleren: it works using this http://www.thousandparsec.net/wiki/Google_Summer_of_Code/Nuleren%27s_Propsal10:58
tpb<http://ln-s.net/1k6D> (at www.thousandparsec.net)10:58
nulerenmithro: good call, thanks10:59
mithrooh nuleren, your the Tiger and Euphrates guy!10:59
nulerenmithro: yep!10:59
protohackmithro - true it is just frustrating to not know about this in advance and have to worry about an important test and other stuff while writing it.10:59
mithronuleren: are you on the mailing list?10:59
mithroprotohack: I know the frustration, I applied to be a student twice11:00
nulerenmithro: I just submitted my email address to subscribe to it a few minutes ago... waiting on the confirmation11:00
protohackAnd the thing is, this summer I will be free for the most part. So I know I will have plenty of time to get a project done.11:01
mithroJLP: think I should post the tips email to the forums?11:01
mithrohttp://news.gmane.org/gmane.comp.games.tp.devel11:02
tpbTitle: Gmane Loom (at news.gmane.org)11:02
nulerenmithro: ignore my last message... I just got the confirmation email11:02
JLPmithro: sure, why not11:02
mithroJLP: could you install a favicon?11:02
mithro(on the forums)11:03
protohackmithro: yeah these tips help out a lot.11:03
JLPmithro: i thought i did, work fine here11:03
mithrodoesn't seem to work with epiphany :/11:03
protohackon a side note: it's weird, for some reason on icechat you have to double click the link for it to open. I just realized that after clicking it a bunch of times.11:03
JLPmithro: works here in konqueror 4, let me check with firefox11:05
protohackWell, I'm bout to go. I'm actually at work now. I'll be on later.11:06
JLPmithro: also works11:06
*** protohack has quit IRC11:06
mithroJLP: weird11:07
mithrohttp://forums.thousandparsec.com/viewtopic.php?f=4&t=10111:19
tpb<http://ln-s.net/1k6h> (at forums.thousandparsec.com)11:19
*** Lukstr has joined #tp11:31
*** vi1985 has joined #tp11:41
*** greywhind_ has joined #tp11:58
*** greywhind has quit IRC11:58
*** zzorn_laptop has quit IRC12:10
*** DTRemenak has quit IRC12:39
*** DTRemenak has joined #tp12:40
*** nuleren has quit IRC12:44
*** Lukstr has quit IRC13:13
*** Erroneous has joined #tp13:20
*** andrei has quit IRC13:30
*** mithro has quit IRC13:31
*** greywhind has joined #tp13:58
*** greywhind_ has quit IRC13:58
*** Lukstr has joined #tp14:16
*** andrei has joined #tp14:22
*** brennan_ has joined #tp14:26
*** ryan__ has quit IRC14:28
vi1985hey ppl, I sent a letter out in the tp-devel list. The news is a significantly enhanced "roadmap" section (as per mithro's suggestions!), and  a link to a visual document, which helps outline the core algorithm of my proposal. Check it out ;)14:44
* vi1985 *finally can get to do some school projects...!*14:45
JLPvi1985: schedule looks nice to me now14:59
*** andrei has quit IRC15:05
*** andrei has joined #tp15:05
vi1985JLP: thanks. I was actually surprised to see that by expanding it, it gave me a much better perspective as to the scope of the task relative o the timeframe!15:05
JLPvi1985: that's one of the ideas of it probably :)15:06
vi1985JLP: now I see the logic in it ;)15:07
SmokingRopeJLP: i've got another revision ready if you have time to do a read through15:08
SmokingRopehttp://easlnx01.eas.muohio.edu/~hannasm/TPProposal15:08
tpb<http://ln-s.net/1k+2> (at easlnx01.eas.muohio.edu)15:08
JLPandrei: i think it is ok and maybe even desired to discuss about things with others, i see it all as part of an open source way of working15:08
JLPSmokingRope: i'll add it into one of my tabs here :)15:09
SmokingRopeyay15:09
*** Epyon has joined #tp15:09
andreiJLP: Ok, Just want to make sure people won't consider me an asshole for it :P15:09
EpyonNow thats one hell of a party! -_-15:09
andreiJLP: People tend to get touchy when others poke at their ideas15:09
* Epyon is back among the living.15:09
* vi1985 *going to do school projects... for real now. so long, and thanks for all the fish*15:10
andreivi1985: So, when you say that your search space is 4D. What are you referring to?15:10
Epyonandrei, the whole point of OS development is to make "my" ideas into "our" :P15:10
JLPandrei: that thing happens quite often in open source where everyone can see what you do :)15:10
andreivi1985: It seems to me you're misusing the word dimension when referring to GAs15:10
vi1985andrei: ah... I'll get back to you later. I really need to do the project! :)15:10
andreivi1985: Heh, okies :)15:11
andreivi1985: And err, there are  few other things you should expand on. Other than the search space15:11
*** vi1985 has quit IRC15:11
andreiwell, that was too late15:11
*** vi1985 has joined #tp15:12
andreiEpyon: JLP I know.. but people still get really pissed off. And some projects don't like it when you question other people's ideas15:12
vi1985andrei: it's not a search space ;)15:12
*** vi1985 has quit IRC15:12
Epyonandrei, I almost got banned at #bzflag for that, so watch your step :P15:13
andreiHeh15:13
Epyonandrei, it's your ideas but the project is everyones15:13
andreiIt doesn't seem #tp is that way :P15:13
Epyonandrei, because he're we're building peripherals. If you'd start messing with the protocol itself, I'm sure people would get a lot more touchy -- and that would be understood.15:14
JLPas long as the discussion is civilized and tastefull there should be no ban15:14
andreiAgreed, but in this case he got something fundamental about GAs wrong.. which changes his idea quite a bit (and he didn't explain the critical things that make or break a GA algorithm (that you have to know before starting))15:14
andreiI've been debating for the past 2 days if I should say something or not..15:15
JLPandrei: i think i can't say anything about that as i have more or less no clue about AI15:15
andreiJLP: I hate it when things get uncivilized.. it's what I stopped answering questions on #C++...15:15
andreiJLP: I suspect most people here don't know how GAs work, that's why I felt I should bring it up :)15:15
JLPandrei: probably best to wait until both mithro and vi1985 are back and discuss this15:17
andreiyup :) I can do it tonight15:17
andreiI feel bad bashing his idea..15:17
JLPandrei: yeah i can imagine, but maybe if there is something wrong it will be seen and there is probably still time ti fix it all15:18
andreiJLP: Well, we'll see :) (I'm curious to see how much experience he has with GAs)15:21
andreiI always fear when people break them out anywhere. Even researchers with a lot of knowledge have issues writing them (and that's for simpler problems)15:21
andrei(same with neural networks, they sound great but getting good results is a dark art)15:22
*** vi1985 has joined #tp16:22
*** ryan__ has joined #tp16:22
*** ViciousM has joined #tp16:34
*** ViciousM has left #tp16:34
*** JLafont has joined #tp16:38
*** ryan__ has quit IRC16:40
*** ryan__ has joined #tp16:41
vi1985hi everyone, I saw there was some interest in the PDF i published. Just a note to those who read it, I made some minor changes here and there, to have it convey the info more clearly. The wiki now points to the new document. I have sent a message on the td-devel list.. just thought it wouldn't hurt mentioning it here as well ;)17:18
andreivi1985: Will you be on later?17:20
andreivi1985: I was told to ask my questions when mithro is around since he knows about the AI stuff more than the other devs17:20
vi1985andrei: I'm expecting guests in an hour, so probably no17:20
andreivi1985: well, I'll be around until 1-2am EST so (and tomorrow as well)17:21
vi1985andrei: ok, I've got no probs. I have a few questions of my own to you. I wouldn't have bothered, but if you insist :)17:22
andreivi1985: Go ahead :)17:22
andreivi1985: Sorry.. I don't want to seem like I'm picking on you..17:22
vi1985andrei: oh well... so much for being civilized...17:22
vi1985andrei: it's legitimate.17:22
vi1985andrei: i wouldn't have started it, but i understand.17:23
andreivi1985: Started? I don't want to start anything; anyone is more than welcome to ask anything at all that they want about my proposal17:23
andreiI really don't mind at all :)17:23
vi1985andrei: ok. btw, i didn't notice the "am" part in the time... I'll come again at around 11am EST, and stay till about 6 or so17:24
vi1985andrei: is that fine?17:25
andreivi1985: Okies :) I'll try to be around most of the day17:25
vi1985andrei: it's a date.17:25
andreivi1985: I'm usually close to a computer (with so many surrounding me it's hard not to be)17:25
vi1985:)17:25
andreiExcellent :)17:26
vi1985andrei: lol, I usually take it 1 computer at a time! :)17:26
andreiHeh, I'm usually using 2 at once (3 monitors in total) and I ssh into one other to run various things17:27
vi1985andrei: you're multicore? :)17:27
andreiI wish :P17:28
vi1985andrei: kk... c u tomorrow. Bye everyone! :)17:29
*** vi1985 has quit IRC17:29
*** javawizard2539 has joined #tp17:52
*** javawizard2539 has left #tp17:53
*** greywhind_ has joined #tp17:58
*** greywhind has quit IRC17:58
*** bddebian has quit IRC18:18
*** EmoDeathMachine has joined #tp18:21
*** Trollin has quit IRC18:37
*** andrei has quit IRC18:40
*** Iwanowitch has joined #tp18:49
*** ryan__ has quit IRC19:00
*** ryan__ has joined #tp19:01
*** ryan__ has joined #tp19:02
*** EmoDeathMachine has quit IRC19:25
*** andrei has joined #tp19:44
*** greywhind_ has quit IRC19:54
*** zzorn has quit IRC19:54
*** danderson has joined #tp20:03
dandersonhi all.20:03
andreiHello20:04
danderson(no, I'm not another student looking for SoC information, I'm a mentor for another org and got curious :)20:05
Ohmhey danderson20:08
dandersonhmm. In libtpproto2-py's protocol.xml, what is the use of the descstruct node?20:19
dandersonit seems to invariantly contain <structure />, which I find a little puzzling20:20
dandersonhmm, guess nobody knows :-)20:34
JLafontdanderson, you might have to wait for mithro to get back, I think he was the one that wrote most of the python code20:43
dandersonJLafont: thanks for the info.20:43
JLafontdanderson, np. I just happened to look at my screen at the time20:44
*** Erroneous has quit IRC21:06
*** Iwanowitch has quit IRC21:29
*** bostonvaulter has joined #tp21:52
*** llnz has joined #tp22:08
llnzhi all22:12
Epyonyo ;)22:13
*** bddebian has joined #tp22:18
bddebianHi folks22:19
llnzhi bddebian22:23
bddebianHello llnz22:24
*** mithro has joined #tp22:34
dandersonmithro: hey. I'm told you're the guy to question about libtpproto2-py?22:35
mithroyes22:35
mithrohold on a minute my gnome-panels are crashing22:36
dandersonk22:36
mithroback22:38
dandersonhi22:39
dandersonso, the context: I've found this awesome hammer that is pyparsing22:39
dandersonso I'm on a rampage looking for nails :P22:39
dandersonor things that remotely look naily.22:39
dandersonso I wanted to have a little fun and try making a protocol description format that is a little easier on the eyes than XML22:40
dandersonfor that lib.22:40
dandersonso, first, I don't know if that interests you at all, or if you're happy with XML22:41
dandersonmithro: ^22:41
mithroI'm fairly happy with XML - but if you are going to also improve other parts of libtpproto2-py it would be cool :)22:41
dandersonheh22:41
dandersonwell, why not :)22:42
mithroit does need some love22:42
mithrobut it's got some cool features22:42
mithrolike it uses generators to allow you to have multiple outstanding requests in different parts of the application22:43
dandersonnice22:44
* JLP fell asleep after coming from work and now wakes up at 4am22:44
JLPahoy all22:44
dandersonwell, once I understand where/how the generated stuff is used22:44
dandersonI'd be happy to mess around with it, see what I can come up with :)22:45
mithroit still has a few bugs - but in theory you should no longer have to keep polling the network interface22:45
mithroI would also recommend checking out libtpproto-py too22:45
dandersonI do have a few questions though; I've been staring at protocol.xml, trying to understand the definitions, and I'm confused about a few things22:46
dandersoncan I pick your brain?22:47
mithroshoot :)22:47
mithrollnz is our xml master really :)22:47
* llnz is here22:48
dandersonokay, so first thing I find puzzling is the parameterset nodes22:48
dandersonI'm unclear as to how they differ from the ordinary structures found in the packets22:49
llnzdanderson: some frames (namely orders and Objects) can have an unlimited number of parameters22:49
llnzwhich are described in description frames22:50
llnzthere is a parameterset for objects and a parameterset for orders22:50
dandersonokay. Why not in that case use a list of structs in the frame?22:51
llnzEpyon: re 3d client proposal, maybe push usability more22:51
llnzdanderson: because they have a strong association between the description and use frames22:52
llnzincluding order22:52
llnzthe order of the parameters, that is22:53
dandersonNot sure I understand :/22:54
llnzwhich list doesn't capture22:54
dandersonah, a list is unordered?22:54
llnzhumm....22:54
llnzok22:54
llnzhave you read the protocol documentation?22:54
llnzeither protocol3.php or protocolxml.php?22:54
dandersonaha, no. Let me do that before I ask more stupid questions :)22:55
dandersonthanks.22:55
llnzhttp://www.thousandparsec.net/tp/dev/documents/protocol3.php22:55
tpb<http://ln-s.net/Ix2> (at www.thousandparsec.net)22:55
llnzhttp://www.thousandparsec.net/tp/dev/documents/protocolxml.php22:55
tpb<http://ln-s.net/JXj> (at www.thousandparsec.net)22:55
llnzprotocolxml is created from the xml22:55
llnzand documents TP04, protocol3.php is TP0322:55
mithrodanderson: that making more sense?23:07
mithrowe welcome improvments to our documentation too! :P23:07
dandersonstill parsing23:08
danderson(4am, processor lag)23:08
dandersonI'm somewhat confused. I see parametersets appear in the XML description, but not in the corresponding documentation23:27
danderson(in TP04)23:28
mithrowe welcome improvments to our documentation too! ;)23:28
dandersonoh, wait, there it is23:29
dandersonmithro: to improve them, I must first understand what is going on23:29
dandersonelse I'll be documenting at random :P23:29
dandersonanyway, bed time for now. bbl23:30
mithrookay then23:30
*** mithro_ has joined #tp23:31
*** bostonvaulte1 has joined #tp23:32
*** mithro has quit IRC23:32
*** mithro_ is now known as mithro23:32
*** bostonvaulter has quit IRC23:38

Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!