[06:24] Join: fiveop joined #corewars [07:33] MSG: Ping timeout: 255 seconds [07:33] Join: fiveop joined #corewars [08:18] Join: Core_old joined #corewars [08:19] booooring moooorning [08:20] hey... [08:20] *yawn* [08:20] just back from work... [08:28] MSG: Read error: Operation timed out [08:29] Join: fiveop joined #corewars [08:31] oh well [08:31] better listening to unix-course than actually working [08:45] Join: Roy joined #corewars [08:45] hi all [08:59] MSG: Ping timeout: 255 seconds [08:59] Join: fiveop joined #corewars [09:07] h R [09:10] * Roy is pondering on a minimal bounding circle.. [09:10] another of the math-challenges? [09:11] the current [09:11] Recmath.. we need two more simple algortihms and then we are going to try and bruteforce it a little :) [09:15] unix-lab pretty much done for today.. [09:15] time to parse koen for a while [09:19] You are parsing koen? [09:19] What for..? [09:19] searching rubble for pearls [09:20] I usualy search for older warriors which are pretty high in the list [09:21] They usualy have old thoughts that got lost, or things we don't see much anymore [09:21] * Roy is thinking of making a new modern(ish) vampire [09:22] Too bad koen doesn't have a 'submitted' date in the table [09:22] I wonder if John has experimented with the code i dumped on him.. [09:23] reminded because you mentioned vampire [09:23] You thought of something with a vampire too? [09:24] i dont know how much i want to give out the idea for now [09:24] though it is not a firsttimer, John had once tested similiar before [09:26] Ah, a lot of ideas get lost in time (too bad) [09:26] I want something to do with enslaving the opponent [09:27] bah, maybe ill just blurt the idea out anyway [09:29] First - take a scanner, plain blur is enough for demonstration. Give it a zoom-trick. Replace the scan-loop with spybomber similiar to that of piltdown (eg. mov.ab #iamhere, ptr). [09:31] (heh, no need to guess what type of warrior Xerox0.1 is) [09:31] Hehe, yeah, piltdown is cool [09:32] John's warrior didnt zooom, and that might be good enough add-on to get the warrior on hill (or to get same score today as it got ages ago) [09:33] but i dont know if the loop can be made fast enough to actually be dangerous [10:13] I really like .75/.8 scanloops, been experimenting with those, but they are to large for a competitive blur scanner [10:15] ye [10:20] Also I do like the pre-scanning idea (like Pre74blabla and Leap in the Dark) [10:21] Prescanning + vampire + enslaving should be fun [10:22] Instead of trapping the processes in a pit with only spl's its easy to convert it into a spl-clear, or maybe even a vampire-clear (with jmps back the to the clear) [10:23] This lets the enemy whipe itself muaha, and then you can easily convert the clear into a datclear at the end, then finally killing the clear itself...done :) [10:27] prescan -> something found, drop a pit -> boot vamp [10:34] or prescan -> drop pit -> scanvamp into pit -> dclear [10:35] Probably to big :( [10:36] I still can't believe HSA doesn't whip every body's ass... its so fast, and doesn't even scan things that are bombed and not executed afterwards [10:37] because it is being optimized against [10:37] and it is not small either [10:38] sure, its torns 95% of papers apart, but thats it [10:43] Yeah, not very bomb-proof [10:47] its kinda same with Recon2 [10:48] But that warrior is even stronger (smaller) [10:49] And fixed length so less vulnarable to clears etc [10:50] because both give me gray hairs when trying to hand-optimize something [10:52] Hehe, I was just going to say: Why not try and make a zoom scanner that bombs like HSA... [10:52] But that IS HSA :-) [10:52] Well, not completely.. [10:52] kinda hard to zooom a scanner with non-ongoing scanner [10:53] *clear [10:53] While bombing you can do a extra check to see if you should jump back to zooming [10:53] But that would cost way more then you gain [10:56] of course you could use second loop for clearing, which would gain almost the same thing (with more instructions) [10:56] Also a nice thing would be HSA bombing but with a max-length of 8 or something, but that too is hard to implement without much instructions per loop > slower bombing [10:59] you would have to do something like add.ab #count, 2 | mov.i spl, >ptr | djn.b -1, #count | jmz.f -3, @ptr [10:59] waitaminute [10:59] now i obviously didnt think that through [11:00] even with jmn.f -3, @ptr it would still remain vulnerable to imps and clears [11:01] One way might be: jmn mov mov djn [11:01] mov mov djn ends after 8, and the jmz jumps if its zero [11:02] But the speed of that drops from .66 to .5 [11:02] mov.i spl, @1 seq.ab #0, #target | djn.f -2, -1 | something like this comes to mind, but too unoptimal [11:03] unless one uses double-mov [11:04] yeah, which HSA already does [11:05] you could implement a part in scan which notices two similiar instructions in a row, but i is worth the increase in size and resetting scan-pointers or whatever [11:06] most likely scenario of two similiar instructions in a row is a spl-carpet or an imp rolling [11:06] (or overwritten papers, or double-spl in a stone) [11:07] The biggest problem is stones, a good stone-detection would be cool :) [11:07] almost impossible [11:07] sure, you can ignore single bombs, but they still cause slowdown as loop extends with one [11:08] maybe you could use seqsne.i dat0, target and replace dat0 with bomb --- nah [11:09] If you could recognise a bomb you should start a clear quickly [11:09] or a paper (see argon) [11:10] JM+me once made a scanner that ignores decrements/increments, that was funny too (not very effective, but fun) [11:12] maybe you do a pre-scan for something, and after finding something you will use variable-lenght carpet, and use the lenght of carpet as size for fixed carpet [11:13] youd have to ignore qscans, though [11:13] but that wont help against bombs a bit [11:13] well, 6-8 cycles, but thats still ~10 cycles lost per found [11:20] and tactically you would keep reacting to the opponent during early fight, which theoretically isnt the best thing [11:21] hm [11:21] you talk about ontopic stuff [11:21] that's boring :/ [11:21] I used to test also a scanner with constantly increasing carpets, didnt work [11:21] depends on ontopic [11:21] ;) [11:22] Some ontopics are fun, and brainstorming is always usefull! [11:22] So, what offtopic thing would you like to chat about fiveop? [11:23] i feel like optimizing a little-better-than-beginner's warrior, though i doubt i will be finding such before i run out of inspiration [11:26] bbl, school is over [11:28] MSG: Ping timeout: 255 seconds [11:31] Roy: dunno [11:31] atm I feel like doing nothing [11:39] How come? [11:51] tiredness? :P [12:23] back, sort of [13:22] MSG: Quit: dat [16:24] Join: Nenad joined #corewars [16:27] hi N [16:27] hi [16:27] :) [16:27] Join: Core_old joined #corewars [16:28] my internet connection is catastrofic today, due to some repairs on the connections at the provider side, but at least I can still do this :) [16:28] Miz: will you participate in the upcoming tourney? [16:29] I will post more details to rgc once my connection becomes normal again [16:30] yes [16:30] all I can say now is that I've decided not to make it an elimination tournament [16:30] more of a league... so, everyone participates in all rounds [16:30] ok. I'll add your name to the list [16:30] 14 players so far :) [16:31] I won't participate, of course. it would be unfair [16:31] but no beginners, so far [16:31] I'd like to see some newbies entering the tourney, as well [16:33] how are you? doing anything fun today? [16:34] eating bacon+cheddar flavoured cheese curls [16:35] sounds tasty :) [16:35] bacon makes almost everything taste better [16:36] and your weight increases as well... bacon is great, but in small amounts [16:36] makes wonders to your blood veins [16:52] "and your weight increases as well" I doubt this in it's generality :P [16:54] generality? [16:54] I don't think you wanted to address Mizcu in particular with 'your'? [16:55] oh :) [16:55] of course not. [16:56] besides, I don't know how any of you look like :) [16:56] and even then I wouldn't go that far as to make such a remark [16:57] fiveop: how about you? anything interesting going on lately? [16:57] you haven't applied for the tourney yet. not interested? no time? or still thinking about it? :) [17:01] Join: Owner joined #corewars [17:01] Hi [17:01] hi [17:02] Greetings Nenad [17:06] Nick Change: Owner changed nick to John [17:24] I've got to go :( [17:24] * Nenad waves [17:24] * John waves [17:24] MSG: [17:31] hm now I can't awnser him [17:39] Hi Fiveop [17:40] How's brx now, haven't seen him here for a while [17:40] no idea [17:42] MSG: Quit: Lost terminal [17:42] Hi Mizcu [17:45] Paperidea sucks. Surprise-ratio: 2% [17:45] of course i could debug it again -_- [17:48] I've had 59 different visitors to this page today http://corewar.co.uk/scanner.htm [17:48] Not sure why [17:48] cant say anything myself [17:48] but there has been a newbie trying a scanner on the sal-b [17:48] and i think it was Roy who had been thinking some ideas too earlier [17:49] 59 different IP addresses, Roy + newbie = 2 [17:49] Since it seems so popular I stuck a google ad at the top! ;-) [17:50] hmm, this could use a little bruteforce.. [17:51] the paper is a 3-in-1.. (yay!) [17:52] 3-in-1? [17:53] ihad a talk with you about the 2-in-1 paper some time ago [17:54] however it socks enough not to be a hill-viable [17:57] Link my new scanners then :-/ [18:12] Link EQU Like [18:21] my code is crap today [18:21] :-( [18:22] what i find weird is that realigning the scan doesnt make any difference to the place is actually selftriggers on [18:22] or i am just blind [18:22] When something doesn't work for me I rewrite it, and usually pick up on bits I can improve in the process [18:29] MSG: Remote host closed the connection [18:30] MSG: Ping timeout: 255 seconds [18:33] Join: John joined #corewars [18:37] Hmmm... strange. [18:38] I have now had 76 visitors to the scanners page, all different IP addresses. The strange thing is 71 of them are using Firefox [18:38] * John wonders [18:39] various countries? [18:42] Not sure, don't know how to check that without doing each IP individually [18:42] well, various first subnets? [18:48] Yes, mostly different [18:55] i guess you dont see the Referrers in the stats? [18:55] No :-/ [18:56] that would explain things [18:57] maybe the page for some reason doesnt render well on firefox and it got put on some bugreport-page [18:58] Hmmm... [18:58] Actually you could be right [18:59] Yeah, the border on my

 preformatted text behaves differently when the browser is resized in Firefox / IE
[19:03]	 Not sure how to fix it :-(
[19:10]	Join: fiveop joined #corewars
[19:11]	 Hi Fiveop, welcome back
[19:14]	 hi
[19:14]	 tahnks
[19:14]	 played a game of chess
[19:16]	* John hasn't done that for a while
[19:50]	* John wves
[19:50]	MSG:  Quit: mov.i #1,1
[21:17]	MSG:  Quit: Lost terminal
[23:10]	Join: Brad_Garr joined #corewars
[23:10]	 hey
[23:10]	 is anyone here?
[23:10]	 I need some help
[23:13]	 Mizcu?
[23:13]	 pak21?
[23:13]	 multimode?
[23:13]	 log_guy?
[23:13]	Nick Change: Brad_Garr changed nick to Brad
[23:13]	 asw?
[23:13]	 datagram
[23:13]	 johnkw?
[23:13]	 brad?
[23:17]	 omg why isnt anyone ever active?
[23:25]	MSG: