00:20:24 -!- coppro has joined.
00:22:31 -!- pikhq has joined.
00:34:48 -!- pikhq has quit (Read error: 104 (Connection reset by peer)).
01:03:47 -!- sebbu2 has joined.
01:05:59 -!- Gracenotes has joined.
01:07:04 -!- sebbu has quit (Read error: 60 (Operation timed out)).
01:21:15 -!- Patashu has joined.
01:27:07 -!- pikhq has joined.
02:02:44 -!- coppro has quit (Remote closed the connection).
02:08:28 -!- sebbu has joined.
02:23:05 -!- sebbu2 has quit (Read error: 110 (Connection timed out)).
02:56:20 -!- coppro has joined.
03:01:57 -!- calamari has joined.
03:04:38 -!- Corun has quit.
03:10:55 -!- pikhq has quit (Read error: 54 (Connection reset by peer)).
03:21:30 -!- pikhq_ has joined.
04:01:32 -!- Slereah_ has quit (Read error: 110 (Connection timed out)).
04:07:52 -!- Slereah_ has joined.
04:09:42 -!- pikhq_ has changed nick to pikhq.
04:16:57 -!- pikhq has quit (Read error: 104 (Connection reset by peer)).
04:55:03 -!- pikhq has joined.
05:02:21 <pikhq> So, in theory, Unicode works.
05:03:00 <pikhq> «¡The composé key works!»
05:03:29 <pikhq> Amazing what having your local set right can do.
05:21:34 <pikhq> Unicode, I declare, is a great boon. What sayeþ ye, men and women and small fuzzy creatures, of all þat is Esolang?
05:42:38 -!- oerjan has quit ("leaving").
07:07:14 -!- FireFly has joined.
07:10:55 -!- FireFly has quit (Client Quit).
07:14:24 -!- calamari has quit (Read error: 104 (Connection reset by peer)).
07:20:33 -!- kar8nga has joined.
07:30:39 -!- olsner has joined.
07:44:27 -!- coppro has quit (Read error: 110 (Connection timed out)).
07:49:23 -!- kar8nga has quit (Remote closed the connection).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:25:39 -!- olsner has quit ("Leaving").
09:52:38 -!- GregorR-L has joined.
09:58:07 -!- Slereah has joined.
10:02:04 -!- impomatic has joined.
10:08:37 -!- Slereah_ has quit (Read error: 110 (Connection timed out)).
10:12:31 -!- Slereah_ has joined.
10:20:34 -!- impomatic has quit (Read error: 110 (Connection timed out)).
10:22:21 -!- Slereah has quit (Read error: 110 (Connection timed out)).
10:25:18 -!- Slereah has joined.
10:36:38 -!- Slereah_ has quit (Read error: 110 (Connection timed out)).
11:12:00 -!- jix has joined.
12:12:04 -!- lereah_ has joined.
12:13:24 -!- sebbu2 has joined.
12:28:09 -!- sebbu has quit (Read error: 110 (Connection timed out)).
12:50:20 -!- MizardX has quit ("What are you sinking about?").
13:04:07 -!- MizardX has joined.
13:05:16 -!- Slereah_ has joined.
13:05:49 <lereah_> http://may.2chan.net/27/res/117551.htm
13:06:15 -!- Slereah has quit (Read error: 60 (Operation timed out)).
13:30:01 -!- Corun has joined.
13:56:41 -!- Slereah has joined.
14:03:00 -!- Slereah_ has quit (Read error: 110 (Connection timed out)).
14:17:20 -!- Slereah has quit (Read error: 110 (Connection timed out)).
14:34:47 -!- ais523 has joined.
14:40:32 -!- Slereah has joined.
15:07:27 -!- Slereah_ has joined.
15:07:46 -!- Slereah has quit (Read error: 110 (Connection timed out)).
15:22:52 -!- Slereah_ has quit (Read error: 60 (Operation timed out)).
15:39:29 -!- lereah_ has quit ("Leaving").
15:42:04 -!- impomatic has joined.
15:43:59 * ais523 grumbles at the hill still apparently being upside-down
15:44:57 <ais523> there was meant to be a tournament running here
15:45:04 <ais523> but the current code eliminates all the good programs
15:45:06 <ais523> rather than all the bad ones
15:45:33 <ais523> well, suicide, the worst BF Joust program theoretically possible, is still there
15:45:49 <ais523> it loses every game, unless the opponent uses the same strategy, in which case it's a draw
15:46:17 -!- Slereah has joined.
15:46:28 <Patashu> do you have control over the code or is it someone else's bungle?
15:46:57 <ais523> it's in GregorR (or GregorR-L) 's code
15:49:07 <impomatic> I've seen someone get a KOTH upside down before so it must be pretty easy to do :-)
15:49:25 <ais523> well, something as simple as a reversed test could to it
15:49:28 <impomatic> Hopefully it'll be fixed soon. I assume the hill will then run forever?
15:49:49 <ais523> but in this case it's the fact that an array's the other way round to what GregorR thinks
15:49:53 <Patashu> are there any 'program vs program' languages besides corewar, fukyourbrane and BF joust?
15:49:54 <ais523> and yes, probably for ages, anyway
15:50:00 <ais523> after all, the hill is still up
15:50:08 <ais523> Patashu: quite a few, I think; but only corewar is really popular
15:50:27 <Patashu> I have to agree about BF joust with other people who've noted there's only one way to go about winning
15:50:40 <ais523> the revised version, there are at least three strategies
15:50:58 <ais523> I've done quite well with defence programs, which wait to be attacked then try to trick the opponent off the end of the tape
15:51:09 <ais523> by dropping the flag to 0 for one cycle, then back up again
15:51:12 -!- FireFly has joined.
15:51:20 -!- Gracenotes has quit (Success).
15:51:51 <ais523> I think that strategy is a) one of the best, but b) very very hard to pull off well
15:52:40 <impomatic> I had a defensive program too before the hill was changed. I've got a copy of the 20 warriors from the old hill so I can resubmit it when the hill has been fixed.
15:52:57 <Patashu> any plans for more opcodes?
15:53:14 <ais523> Patashu: it just wouldn't be BF then
15:53:27 <ais523> zzo38 has a version with input from the opponent's NOPs, although I don't see the point
15:53:59 <Patashu> input from the opponent's NOPs? what does that mean
15:54:32 <ais523> Patashu: well, . does output in BF
15:54:48 <impomatic> There's a description on the discussion page of the BF Joust wiki page.
15:54:57 <Patashu> since you can simulate a NOP with <><> or whatever
15:55:08 <ais523> well, a one-cycle NOP is useful for parity reasons
15:55:20 <ais523> and messing about with <> or >< or even +- can be very dangerous in BF Joust
15:55:27 <ais523> due to the end of hte tape
15:55:58 <Patashu> it doesn't sound like a very well-founded concept in any case
15:56:18 <Patashu> since you'd never WANT to tell your opponent anything
15:56:24 <Patashu> why make a command for it? :P
15:56:50 <impomatic> I needed to use a one-cycle NOP in one of my programs to get the timing right
15:58:32 <Patashu> maybe , could get the value of your opponent's pointer or the value under your opponent's pointer or something
15:59:02 <ais523> attack5 uses quite a lot of them
15:59:09 <ais523> it's a counter-defence attack program
15:59:23 <ais523> although it can confuse about half of genuine attack programs by setting decoy cells to -6
15:59:27 <ais523> which they aren't expecting
16:00:59 <impomatic> Patashu: see http://aiforge.net for a programming games forum and about 1500 links.
16:01:18 <Patashu> fighting programmable robot games
16:01:53 <impomatic> http://www.sumost.ca/steve/games is also a decent page
16:02:41 <impomatic> There are a few which don't involve robots, but most of the ones I've looked at are similar to Corewar
16:21:53 -!- Patashu has quit ("Patashu/SteampunkX - MSN = Patashu@hotmail.com , AIM = Patashu0 , YIM = Patashu2 , Googletalk = Patashu0@gmail.com .").
16:22:31 <ais523> impomatic: corewar question: how effective is using SPL 0, rather than a goto, to create a loop?
16:30:01 <impomatic> If you have a SPL 0 at the head of the loop and let the processes drop off the end, the loop will be 1 cycle slower.
16:30:10 <impomatic> Processes take 1 cycle to execute the DAT which kills them.
16:31:21 <ais523> if you're planning to spawn a load of processes anyway, though, it might be worth it
16:31:37 <impomatic> However, the loop can only be killed by destroying the SPL. Damaging any other instruction won't kill the loop, just damage it.
16:31:50 <impomatic> It's best to have a SPL 0 at the top and a JMP at the end :-)
16:32:27 <ais523> I was vaguely thinking of setting up some sort of huge number of imp spirals
16:32:41 <ais523> hitting thread limit, so that anti-paper strategies wouldn't work
16:32:47 <ais523> and hoping that at least one survived to overwrite the opponent
16:32:49 <impomatic> But occasionally the instructions are out of sync, but it doesn't really matter. Also it means the loop reverse!
16:34:06 <impomatic> It takes ages to hit the process limit. It's normally 8000 processes.
16:34:16 <ais523> I thought it varied by hill
16:34:20 <impomatic> An opponent can clear the core in that time.
16:34:32 <ais523> but at least, churning out your own threads will reduce the impact of a SPL-based bomb
16:34:54 <ais523> and an imp spiral swarm would tend to intrinsically dodge core-clears
16:35:01 <ais523> 2/3 chance it misses once you're down to one imp
16:35:25 * ais523 wonders what warriors do after the core-clear ends
16:35:40 <impomatic> max processes is normally the same as coresize, the the same applies to the other hills, e.g. nano hill = 80 cells, 80 cycles to clear it with d-clear, 80 maxprocesses.
16:36:13 <impomatic> They just keep clearing over and over, just in case they missed a mobile warrior first time round.
16:36:21 <impomatic> By the way, the #corewars channel has moved to freenode :-)
16:36:31 <ais523> so, is it corewar or corewars?
16:36:42 <impomatic> Unfortunately koth.org is closing down :-(
16:37:06 <impomatic> The official name of the game is Core War.
16:43:33 -!- Gracenotes_ has joined.
16:44:06 -!- Gracenotes_ has changed nick to Gracenotes.
16:50:20 <ehird> ais523: "I disappear yesterday therefore I return."
16:50:32 <ais523> it all depends on what returning means
16:53:53 * ais523 moans at the upside-down hill
16:56:46 -!- OoS has joined.
17:06:38 <kerlo> Hey, where'd my server go?
17:06:59 <kerlo> Looks like all of Slicehost is unusually sequestered.
17:07:00 -!- BeholdMyGlory has joined.
17:07:00 <ais523> kerlo: well, I can log into it
17:07:02 <ais523> if it's the one I think
17:07:07 <ais523> I'm online there right now
17:07:25 <kerlo> Can you get a login prompt at s2.normish.org? I can't.
17:07:35 <kerlo> Well, now I can. Never mind.
17:07:41 <AnMaster> strong static discharge from my chair!
17:08:12 -!- MizardX has quit ("from __future__ import skynet").
17:08:57 * kerlo successfully remembers half of his password.
17:09:09 <AnMaster> lot of static discharges here today... Strange.
17:09:14 <ais523> kerlo: do you want me to change your password, using my scammed root powers?
17:09:15 <AnMaster> from anything in metal at all.
17:09:35 <kerlo> ais523: you have scammed root powers on s2.normish.org?
17:09:45 <AnMaster> anyone has any idea for a reason?
17:09:50 <ais523> if it's a different server
17:09:53 <ais523> just on the main normish.org
17:10:13 <ehird> So kerlo's now paying $50/month so that he can develop a new normish separately on the live deployment environment?
17:10:26 <ehird> You mean like what I wanted to do originally, and was faced with "WHY DON'T YOU JUST DEVELOP IT LOCALLY FOO"
17:10:37 <ais523> I suspect the "new normish" isn't actually a nomic
17:10:48 <ais523> otherwise there'd be no point
17:11:05 <AnMaster> ehird, why not develop it locally? It is a good point IMO.
17:11:21 <ais523> well, I developed the old system locally
17:11:25 <ais523> and it even works, sort of
17:11:25 <ehird> AnMaster: I was pointing out some hypocrisy, and I'm not exactly in a mood to go into Talk To AnMaster Very Carefully And Slowly mode.
17:11:47 <ais523> AnMaster: this is an old flamewar lasting weeks from several months ago
17:11:48 <ehird> ais523: I've been watching kerlo ask things like how to avoid SQL injection with a PHP registration system in Sine. Cringeworthy that he administrates Normish >_<
17:11:49 <ais523> so you're missing contest
17:11:57 <ehird> not exactly flamewar
17:12:05 <ais523> well, it did cause me to shun you for weeks
17:12:07 <ehird> more strong disagreement with additional fire :)
17:12:09 <ais523> which is rather out of character for me
17:12:27 <kerlo> At least I now know how to avoid SQL injection, I think. :-P
17:13:01 <ehird> Next up: How to escape HTML! With ASP.NET -or- ColdFusion -or- PHP!
17:13:06 <ais523> yes, bad-but-works way = proper escaping, good-and-works-better-and-is-easier-way = parameterized or stored queries
17:13:07 <ehird> 10 Best Javascript Libraries
17:13:18 <ehird> 50 Ways To Monitor Your Site's Uptime!!!!
17:13:28 * ehird vomits all over the blagosphere
17:13:41 <kerlo> Parameterized queries. Sounds nice.
17:14:19 <ehird> ais523: he's using PHP and uses Windows, I'm pretty sure you're wasting your time :)
17:14:29 <ais523> I'm pretty sure even PHP can do that
17:14:51 <ehird> I'm just pointing out that maybe talking about reasonable best practices and clean design is better given to /dev/null
17:15:13 <kerlo> That sounds likely.
17:15:43 <ehird> i'm staying on Actual Normish, which at least has proper unix tools instead of PHP/MySQL user registration forms.
17:16:23 <kerlo> Hey, I'm planning to have proper Unix tools as well.
17:16:40 <ehird> kerlo: The user database is in MySQL, is it not?
17:16:49 <ehird> Therefore, not Unix accounts. Therefore, no, no you're not.
17:17:28 <kerlo> Is there some magical force preventing me from having each user in the database have a linked Unix account?
17:17:43 * ehird puts his head through a wrangler to forget what you just said
17:18:09 * pikhq tries to unhear things. Unsuccesfully.
17:18:13 <ais523> actually, the correct way to do that would be to write a mysql backend that stores data in /etc/passwd
17:18:26 <kerlo> That sounds kind of weird.
17:18:30 <ehird> ais523: That is correct for values of "DEAR GOD KILL ME".
17:18:39 <ehird> pikhq: We're doomed.
17:18:48 <ehird> I'm making an underground bunker
17:19:23 <ais523> ehird: I was trying to think up the sanest way to achieve what kerlo wanted; the fact that it's insane implies the original request was insane
17:19:55 <ehird> maybe we could clone the internet, put kerlo in it, wait 3 years, then destroy the cloned internet and let him on the real one
17:19:58 <ehird> it's our only hope
17:20:35 <pikhq> ais523: The sanest way involves a PAM module.
17:20:50 <ehird> ais523: the very existence of this insane "Normish s2" may collapse the whole internet
17:20:52 <ais523> err, that reads a MySQL database?
17:20:56 <ehird> or at least severely damage it
17:21:11 -!- impomatic has quit (Read error: 110 (Connection timed out)).
17:21:37 <pikhq> Not saying it's very sane, mind.
17:22:03 -!- OoS has changed nick to impomatic.
17:22:06 <ais523> pikhq: I don't think PAM's tied into the UNIX UID system for files
17:22:32 <ais523> as in, creating MySQL table rows wouldn't magically create users for any purpose other than logging in
17:23:24 <kerlo> Okay, people who are significantly intelligent and/or knowledgeable than me. Suppose I want people to have Unix accounts and I also want to track various bits of information about those people.
17:23:40 <kerlo> Is creating the accounts and then using YAML that references their usernames the best way to accomplish that?
17:23:53 <ehird> Use proper files, dammit.
17:23:57 -!- impomatic has left (?).
17:24:04 <ehird> Real, unix, plain text, lightweight-format FILES.
17:24:36 <kerlo> Is MySQL useful for its intended purpose? If so, perhaps I've forgotten what that purpose is.
17:24:48 <ehird> To be a bad version of PostgreSQL.
17:24:57 <ehird> PostgreSQL is a program whose purpose is not to be used as a replacement for plaintext files.
17:25:07 <kerlo> What is its purpose, then?
17:25:31 <ehird> Correct me if I'm wrong,
17:25:33 <ehird> but I just told you
17:26:36 <ais523> kerlo: MySQL is the C of databases
17:26:43 <ais523> you have to do everything by hand, but it can be very efficient as a result
17:26:47 <ehird> No, that's an utterly terrible comparison.
17:26:57 <ehird> It's nothing like C.
17:26:59 <kerlo> The sentence "PostgreQSL is a program whose purpose is not . . ." does not tell me what its purpose is.
17:27:03 <ais523> maybe it's the C++ of databases, as it also has a lot of useless features
17:27:08 <ehird> kerlo: whose purpose is (not ...)
17:27:21 <pikhq> Definitely the C++ of databases.
17:27:30 <pikhq> It does a lot of stuff, and does it poorly.
17:27:44 <kerlo> The thing about sarcastic statements is that they're not very useful when interpreted as the truth.
17:27:46 <ehird> ais523: C's lean, mean, efficient, it has a clean design, and does what it does completely well. MySQL has its legs the wrong way round. It's not even really fast. It tries to do things and does them all badly. It's built on top of a bad design.
17:27:52 <ehird> It keeps trying to do more things, and it keeps failing at them.
17:29:01 <kerlo> ehird, I suppose you're the kind of person who would stop using a system if new features were added and those features sucked.
17:29:37 <ehird> I suppose you, kerlo, are the kind of person who will jump to conclusions without basis and then state these matter-of-factly in a deliciously non-sequitur manner.
17:29:54 <ehird> Call me crazy, but.
17:30:48 <kerlo> That may be true, but I'm just saying that to pacify you, not because I really believe it, not that I don't believe that I believe it.
17:30:51 <pikhq> MySQL does most everything it does poorly.
17:35:31 <kerlo> Well, I refuse to use a database system whose name contains four consonants in a row.
17:35:52 <ehird> kerlo: "unix" does not have four consonants in a row.
17:36:12 <kerlo> And now for something not blatantly false:
17:36:39 <kerlo> Would you (collectively) recommend switching to PostgreSQL, then?
17:36:55 <ehird> I would recommend using Unix properly, dammit.
17:37:51 <kerlo> So, suppose hypothetically that I want to store people's usernames and passwords.
17:38:02 <ehird> You would use proper Unix accounts.
17:38:14 <kerlo> Okay, suppose I do that.
17:38:22 <pikhq> And you wouldn't store their passwords, you'd store the hash of their passwords.
17:38:31 <ehird> pikhq: One step at a time.
17:38:37 <kerlo> That's precisely why I said "hypothetically".
17:38:39 <ehird> Think "talk to a three year old".
17:39:59 -!- M0ny has joined.
17:40:24 <ehird> Also, use PAM for authentication of these accounts, not manually wrangling /etc/passwd like I just forsighted you doing. Yes, that means a dumbed-down web interface to the Unix tools you have will be a pain to implement due to PAM not being that easy to automate. That's a feature.
17:40:33 -!- Gracenotes has quit (Success).
17:42:27 <ais523> actually, you store the password hashes in a different file, so you can make most of the user database user-visible, but not the password hashes
17:42:51 <ehird> ais523: Unix. Accounts.
17:43:16 <ehird> Yes. I know. That's not kerlo's concern.
17:43:20 <ehird> There are tools that handle that.
17:44:04 <ehird> "Nanotechnology discoveries: Disks that can store 10000 times more data than a DVD and that can outlive stone inscriptions!"
17:44:08 <ehird> FUCK YEAH NANOTECHNOLOGY
17:44:39 <ehird> "They've further increased the storage density to 1.1 terabytes per cubic centimeter"
17:44:48 <ehird> "In a paper published online today in the journal Nature, Gu's group reports recording speeds of about a gigabit per second."
17:44:59 <ehird> 1Gbit/sec writing to media with 1.1TB/cm2.
17:45:12 <ais523> I wonder what its rewrite is like?
17:45:21 <ais523> although something like that would be amazing even in write-once form
17:45:25 <ehird> maybe it's write-once, yeah
17:46:02 <ehird> outliving stone inscriptions too
17:46:04 <ehird> is pretty damn awesome
17:46:11 <ehird> ais523: see, THIS is the ideal backup media
17:46:19 <pikhq> ehird: 1.1TB/cm³ is pretty spiffy. :)
17:46:26 <ehird> pikhq: er, ^3 yeah not 2.
17:46:53 <ehird> You can't mess up your backups, you can store all your backups for forever with only a few discs, and backups are instant (assuming sufficiently fast source media)
17:46:59 <ais523> hmm... that would be over a petabyte per litre
17:47:03 <ais523> and storing a litre isn't that hard
17:47:14 -!- oerjan has joined.
17:47:30 <pikhq> Imagine the bandwidth of a station wagon hurtling down the road with one of this things.
17:47:42 <ais523> oerjan: been logreading, or just an out-of-content reply?
17:47:52 <ehird> pikhq: I wonder how many library of congresses per disc it'd be?
17:48:28 <ehird> ais523: pikhq: 1Gbit/sec = 128MB/sec
17:48:32 <ehird> ok, that's rather less impressive
17:49:00 <ehird> ais523: well, it's what internet connections use
17:49:09 <ehird> maybe CDs/DVDs/Blu-Ray are measured in that too
17:50:16 <kerlo> So, suppose we let our users log in via the web. If I'm not mistaken, the chief non-weird way to do this is cookies. Would these cookies contain the users' passwords?
17:50:32 <ehird> kerlo: AAAAAAAAAAAAAAARFGH
17:50:37 <ehird> Stop coding stop coding stop coding
17:51:17 <ehird> Give or take a day
17:51:21 <oerjan> kerlo: no, and i can say that without even having read the answer. what about session ID?
17:51:32 <kerlo> "No" is the answer I expected.
17:51:40 <ehird> kerlo: not cookies.
17:51:47 <ehird> Sessions. If you don't know what they are, stop coding some more.
17:51:53 <ais523> kerlo: cookies generally contain autogenerated nonsense that's used to identify a session
17:51:58 <ais523> and the session is used for everything else
17:52:04 <ehird> ais523: just tell him to use PHP sessions
17:52:16 <ehird> we can't save him, but we can edge him in a sort of right direction to avoid unspeakable disaster
17:52:20 <ais523> then you don't have to worry about how they're implemented (even though they're normally implemented with cookies)
17:52:30 <ais523> OTOH, isn't PHP open to session fixation unless you take steps against it?
17:52:50 <ehird> My current strategy is attempting to make him outsource as much code as possible to Unix and his language. Less dangerous :)
17:53:14 * kerlo reads about sessions.
17:53:16 <ehird> ais523: argh, "One discovery can store way more than a DVD, and the other has a long lifetime. Furthermore, the first one is write-only, and I don't see indications of mass storage for the second."
17:53:19 <ehird> ais523: so two separate things
17:53:41 <ais523> and Slashdot lying as usual?
17:53:57 <ehird> The blog title lying to get sensationalist hits
17:54:06 <ehird> and the reddit submitter— bet it's the author of the blogspam— using the same title
17:54:23 <ehird> at least slashdot clarifies its sensationalism in the summaries
17:55:03 <oerjan> ehird: wait, write-only? you mean it _cannot_ be read again? :D
17:55:17 <ehird> oerjan: http://www.supersimplestorageservice.com/
17:55:21 <ehird> Write-only storage!
17:55:26 <ehird> Super-secure! Competitive pricing!
17:55:48 <ehird> It's provably secure.
17:55:50 <ais523> err, how can you prove they're storing at all?
17:55:57 <ehird> Would they LIE to you
17:56:15 <ais523> argh, that website has one of the most annoying JS advertising sidebars I've ever seen
17:56:17 <ehird> ais523: Here's their proof: http://www.supersimplestorageservice.com/secure/s4/pipeline/cache/throughput/proof.aspx
17:56:21 * ais523 opens in Firefox instead
17:56:32 <ehird> ((read carefully))
17:56:36 <ais523> now I know why I have JS off by default
17:57:18 <oerjan> ehird: oh it's a joke, must be
17:57:22 <ais523> it doesn't look /quite/ like a genuine IIS formatting page
17:57:23 <kerlo> Should I ask how to store little bits of information associated with each user another time?
17:57:30 <ais523> formatting-wise, I mean
17:57:46 <ehird> ais523: yes, they don't usually reference watching (a) the game, (b) cute kitten videos or (c) female mud wrestling
17:57:50 <ehird> or people starving in China
17:58:01 <ais523> ehird: it's not the text I'm talking about
17:58:04 <ais523> which is obviously different
17:58:59 <ehird> 17:57 oerjan: ehird: oh it's a joke, must be
17:59:04 <ehird> oerjan: very AnMaster of you
17:59:08 <oerjan> "All complaints and feature requests will be immediately stored using our S4-backed user request database." :D
17:59:43 <oerjan> ehird: i _was_ trying to stretch my mind for a moment thinking of any way it could not be a joke
17:59:44 <AnMaster> ehird, when is that quote from?
17:59:55 <ehird> AnMaster: almost every time you're mentioned, ever
17:59:58 <AnMaster> or where you impersonating again...
18:00:02 <ais523> also, Amazon allow PUT requests?
18:00:12 <ais523> that's one interesting piece of information to come out from that site, if it's true
18:00:14 <AnMaster> ehird, just see above for a counter example
18:00:16 <ais523> I thought PUT requests were dead
18:00:21 <ehird> yes, I was nefariously POSING AS YOU, AnMaster
18:00:30 <ehird> And making you FORGET WHAT "ALMOST" MEANS
18:00:39 <ehird> ais523: nah, REST is popular nowadays
18:00:40 <AnMaster> ehird, I think we need secure tripecodes! ~
18:00:44 <oerjan> ehird: i mean, _backup_ is sort of write-only in the short term. but it's not anything new so...
18:00:47 <ais523> ehird: but nothing supports PUT
18:00:47 <AnMaster> (or whatever the spelling was)
18:00:54 <ehird> ais523: sure it does
18:01:22 <ehird> [[Micheal Lynton, the guy who said 'I'm a guy who doesn't see anything good having come from the Internet. Period.' has posted an editorial at the Huffington Post titled Guardrails for the Internet, in which he defends his comment, and suggests that just as the interstate system needs guardrails, so too does the information superhighway.]]
18:01:30 <ais523> haha: https://www.paypal.com/us/cgi-bin/webscr?cmd=_flow&SESSION=g--DKTemd4TMMMq8aU4EeABuqWILa_aMmkQbyhwzrxbrHtx8tmKmlH7irWC&dispatch=5885d80a13c0db1f998ca054efbdf2c29878a435fe324eec2511727fbf3e9efc
18:01:32 <AnMaster> ehird, isn't PUT used in webdav iirc
18:01:38 <ais523> they have a genuine paypal link to buy the entire company
18:01:40 <ehird> AnMaster: ais523:, you mean
18:01:47 <ehird> ais523: that link doesn't work, it's user-specific
18:01:54 <ais523> ehird: but I don't have a paypal account
18:01:59 <ais523> so would it not work for me either?
18:02:13 <ehird> ais523: the author said that he'd probably refund any actual transactions :-P
18:02:19 <ais523> so why can't you steal my session by following the link? IP-based?
18:02:24 <ehird> thus demonstrating he could never own the lottery
18:02:26 <ais523> AnMaster: the S4 people
18:02:27 <ehird> ais523: cookie-based?
18:02:38 <ais523> hmm... cookies /and/ URL seems redundan
18:02:38 <AnMaster> ais523, I'm afraid I missed who S4 were
18:02:43 <ehird> AnMaster: SCROLLBACK!
18:02:46 <ehird> SCROLLBACKSCROLLBACKSCROLLBACKSCROLLBACKSCROLLBACKSCROLLBACKSCROLLBACKSCROLLBACKSCROLLBACKSCROLLBACKSCROLLBACKDSKJFHKSDFKSDJFS
18:03:00 -!- tombom has joined.
18:03:01 <ehird> so just waste OUR time instead?!
18:03:10 <ehird> because clearly only -your- time is valuable
18:03:18 <AnMaster> ehird, a short one line summary would be enough
18:03:31 <ais523> Apache=147.188.254.232.327301199300722205; KHcl0EuY7AKSMgfvHl7J5E7hPtK=K1qt0zJ42nOKrHXQ512V7BuFW5fs_zXxwdkA8IlYo00HhjwGiO42drcE6jsj5EHtVawXwNuuhXfQnm49; cookie_check=yes; jNS36F3v1LVZP8Qp0a2pJWhXNKy=mpG_l1Xk9WVVMJwKZVHh_9nI392FXiTNMVIAFRlmKad9qdQtn72dqh9D_3Ksf_3nREVnPjQCIKyE4epIJ6zrmi4ZEsW; LANG=en_US%3bUS; navlns=0; cwrClyrK4LoCV1fydGbAxiNL6iG=DctYoOYqVUNLQKpmwSnWT6mVvE5ZBi3RJ8UuNQw5FdKNngFlzJV1MaUVDLge9BcQPU-SGvT-
18:03:33 <ais523> dqAUo3AQzWWvZ-IHZGNfi0Z7Aw_GC2pam5X39xkdFHQI7ggSNE53QMYBSeyAJ0%7cbqkyg7YNVpHIk_tGrE-68fFnc0kX79Qa913-xSLSZPAMHlyIubz_4wrZrNk2u9vRhHOv7W%7cIMMKtMGhuMz8vkVo7mLi9UQCNkMi6CQhsTCulqkVImjKunHUWsHfINxwd7vLcmfnTR_FYm%7c1243357268; navcmd=_xclick; pNTcMTtQfrJuaJiwEnWXQ6yNxfq=5mv9RU4DctpsQ9re8jiTEh_JImwBUbicIr7FvspGKVQe8hjrwdl_RlC99UYGDLzzp3vd1gAmrIbXQqXt8oETB_fq6Ge1UBpjmfKeevJ8goC2hs5tc3KX_Ho1I6wIgIN0vjhyfY9ftf2FxLYIdIEu8bg7EBQXC-
18:03:34 <ais523> TVzUYFXxJETr2tshK8PcCE4y5z5tFohs4wYca_1iIRds5Rz1VAhhu84PwLZYw_JAKatkZtmxn84zZYjC-TSp4nhXOFctIUj50m0BTwK8IVThZXxMZ-DgPVnkrX36pzZ4ZxTWdks-7k7i5p8Boqy0JoN-kIeG1qgV4uTn4ajaZP9ZNCsmGOsBCvXvC9cc68-moa-DQxi_oE_0OpF7tE; 6Vt_kuBZl8VlyHAqyfqTECtzKXS=4FmsD86w5SZMRES5DmDJOUf8i2MUEbgVvNrxy-YD1bl1DQlvynIlaY86brrCYbUQzov9tNBAz90bxc47
18:03:40 <ais523> what brilliant cookies PayPal have!
18:03:47 <ehird> ais523: don't do that, you'll use up all of AnMaster's scrollback
18:03:48 <AnMaster> I only know of one S4... And that is an ACPI sleep state
18:03:51 <ehird> he only has three lines, after all
18:10:58 <ais523> oerjan: you're going to have trouble looking after them when they grow up!
18:11:22 * oerjan notes that wasn't a fact
18:11:25 <ehird> ais523: but I AM right back!
18:11:43 -!- bsmntbombdood has quit (Success).
18:11:59 <oerjan> actually lizard pets are illegal in norway
18:12:04 <oerjan> unless they changed it
18:12:10 <ais523> oerjan: do you know why?
18:12:19 <Slereah> They're like scaly ferrets
18:13:13 <oerjan> ais523: norway has (almost?) no native lizards, i guess is the rationalization
18:13:34 <ais523> oerjan: ah, to prevent them disrupting the ecosystem?
18:13:54 <oerjan> ais523: maybe. there is a movement to change the law though.
18:14:07 <ehird> what a fun law to challenge
18:14:31 <ehird> "I will prove that anything written in a higher-level language will not be as fast as my implementation of it in C. I leave this challenge out to anyone to take. (*)"
18:14:35 <ehird> Kill stupid. Kill stupid.
18:14:42 <ehird> Override mode activated. Kill target sequence initiated.
18:15:59 -!- bsmntbombdood has joined.
18:17:16 <oerjan> ais523: actually that's probably not a real reason, as the reason we only have 5 (googled) native reptile species in norway is that it's too cold for others to survive
18:18:16 <oerjan> many think the law is stupid. but then occasionally there is a news item about someone's (illegal) snake growing too big for them...
18:18:45 <ehird> int fact_table[] = { 1, 1, 2, 6, 24, 120, 720, 5040, 40320, 362880, 39916800, 479001600 };
18:18:46 <ehird> #define factorial(n) fact_table[n]
18:18:57 <ehird> ↑ most efficient factorial implementation over 32-bit integers :-)
18:19:01 <ehird> well, maybe a switch/case could be faster
18:19:25 <ehird> ofc, 32-bit is useless for anything involving numbers bigger than your average pony
18:20:07 <oerjan> "i'm not a number, i'm a free pony!"
18:20:21 <ehird> ais523: so you mentioned slashdot and I tried to read the comments of a post about lisp
18:20:35 <ehird> and people called it LISP, were ignorant about how it can be faster than C in many cases,
18:20:37 <ehird> so on and so forth
18:20:41 <ehird> then I gave up 'cuz it's worse than reddit.
18:20:44 <ais523> Slashdot is an interesting mix
18:20:45 <ehird> thanks for that :P
18:20:50 <ais523> you do get very interesting comments from time to time
18:20:53 <ais523> but there's a lot of junk too
18:20:58 <ehird> i didn't really see any gems
18:21:08 <ais523> they're normally good at checking whether the story is true or not
18:22:14 <ehird> [[The service is going to feature in a new unscripted series that will "harness Twitter to put players on the trail of celebrities in an interactive, competitive format". No, I am not sure what that means and there are no further details as yet. ]]
18:22:22 <ehird> Twitter... TV... shoh god.
18:22:51 <ais523> wow, that's even worse than the fictional deliberately bad reality TV program that's the centrepiece of BlogNomic at the moment
18:23:23 <ehird> ais523: you have to work http://picoup.com/ into it somehow, then
18:23:39 <ehird> ais523: Twitter except you only get 18 characters
18:23:46 <ehird> plus one @user for free
18:23:56 <ais523> as a joke, or as a serious attempt?
18:24:09 <ehird> ais523: if 18 is too verbose for you, http://femto.picoup.com/ lets you have a character and a user reference
18:24:25 <ehird> express your feelings like @user !, @user ? and @user …
18:24:31 <ehird> (that's one character)
18:25:06 <ais523> I like the random usernames feature
18:25:17 <ehird> apparently I'm ImpulseLeast
18:25:26 <ais523> how would you log on with the same name in future?
18:25:37 <ehird> ais523: if you click change it (with JS), you can give it a password
18:25:50 <ehird> but you can't use a nick that isn't either (a) automatically assigned, (b) registered
18:25:58 <ehird> and I assume it doesn't generate already registered nicks
18:26:13 <ais523> also, any way to follow anyone in particular there?
18:26:24 <ehird> if you click a nick there's a watch link
18:26:44 <ais523> hmm... well, it might catch on
18:37:25 <oerjan> <ais523> and storing a litre isn't that hard <-- literally trivial
18:41:15 <ehird> http://timesonline.typepad.com/technology/2009/05/new-iphone-2009-a-possible-checklist.html ← "And a pony"
18:42:38 * oerjan sees no pony and feels cheated.
18:46:50 -!- jix_ has joined.
18:47:54 <pikhq> ehird: Any idea how to to do an arrow with the compose key?
18:48:08 <ais523> pikhq: AltGr-I gives → for me
18:48:22 <pikhq> ais523: I don't have no stinking AltGr.
18:48:23 <ehird> pikhq: Nope, I just did it the macfag way and made a keyset file with the characters I want triggered on control-option-<key>
18:48:32 <ehird> ← = control-option-, for me (analogy with <)
18:48:55 * pikhq will have to futz with the compose configuration
18:49:09 <pikhq> Probably stick that on compose <-
18:55:16 <ehird> http://namakajiri.net/diary/things-they-asked-me-in-the-us-visa-interview/en/
18:58:46 -!- jix has quit (Read error: 113 (No route to host)).
19:02:46 -!- olsner has joined.
19:03:59 * ehird attempts to make a smaller typeface than http://en.wikipedia.org/wiki/3x3
19:06:39 * ehird concludes 'tis impossible
19:08:24 <ehird> pikhq: yeah, but somehow still readable
19:09:22 -!- FireFly has quit ("Later").
19:10:44 -!- FireFly has joined.
19:18:52 <ais523> because the list of programs is sorted in reverse order
19:18:59 <ais523> then you eliminate all past the tenth place
19:19:15 <GregorR-L> Oh, I thought they were from best to worst! D-8
19:20:36 <ehird> GregorR-L: Also, rewipe the hill.
19:20:40 <ehird> Since it's all the worst programs now.
19:21:43 <ehird> ais523: add impomatic_shortsword?
19:21:52 <ehird> GregorR-L: oh, it'd be nice if you made it automatically prepend nickname_ to the programs
19:22:17 <ais523> ehird: I don't have it on me
19:22:20 <ais523> but I'll add some of mine
19:22:24 <GregorR-L> ais523: It takes a REALLY long time to timeout if the programs are unhappy :(
19:22:26 <ehird> ais523: with ais523_ pls
19:22:45 <ais523> !bfjoust ais523_attack5 [>[-]-.-.-.-.-.-]
19:22:46 <EgoBot> Score for ais523_attack5: -1 (maximum 5)
19:22:59 <ais523> !bfjoust ais523_defend5 >+>+([{>[(.)*20-]+}]<..........-[++[[]<(-..-.)*300>[>[-]+]]]<(+..+.)*300>[>[-]+])%2000
19:23:00 <ehird> erm, isn't that in the default hill?
19:23:07 <ais523> ehird: no, that was attack/defend/fool 1
19:23:23 <EgoBot> Score for ais523_defend5: -1 (maximum 6)
19:23:24 <ehird> GregorR-L: do you think you could do that nick_ prepending thing & rename the default programs to have ais523_?
19:23:30 <ehird> that plus correct in-channel score reporting = heaven
19:23:52 <GregorR-L> I don't think the default ones should be renamed, they're just examples, but yeah, I can add that.
19:24:06 <ehird> GregorR-L: er, they're ais523's base warriors
19:24:17 <ehird> seems reasonable to prepend ais523_ to me. ais523?
19:24:24 <ais523> <Google> Did you mean: pimpomatic shortsword
19:24:30 <ais523> ehird: well, they are mine, but they're all pretty simple
19:24:41 <GregorR-L> They're like four characters long :P
19:24:44 <ehird> well, okay, it's just "attack1" vs "ais523_attack5"
19:25:18 <ais523> !bfjoust impomatic_shortsword (>++>--)*2(>)*6([-[+]]>)*20
19:25:20 <EgoBot> Score for ais523impomatic_shortsword: -1 (maximum 7)
19:25:38 <ehird> otherwise we'll get impomaticimpomatic
19:25:51 <GregorR-L> I had just added name, then saved, then went "Oh yeah, name_".
19:25:59 <GregorR-L> There was about a five second span in there where it had name and not name_
19:26:07 <ehird> GregorR-L: once you've done that, fix the in-channel reporter, wipe the hill and we can get on with playing
19:26:10 <ais523> GregorR-L: could you just rename the shortsword to the correct name on the hill?
19:26:25 <ehird> But also correct in-channel score reporting!
19:26:34 -!- impomatic has joined.
19:26:49 <ais523> impomatic: good timing
19:26:53 <ais523> the hill is now the right way up
19:26:57 <ais523> as of about 10 minutes ago
19:27:00 * impomatic grumbles about ( ) { and } in BF Joust!
19:27:07 <ais523> impomatic: in what way?
19:27:15 <impomatic> I'm writing an interpreter in asm!
19:27:22 <ehird> GregorR-L: oi, rename ais523impomatic_ to impomatic_ kay? :P
19:27:26 <ais523> impomatic: those are just abbreviations, you don't have to handle htem
19:27:35 <ais523> but they make programs a lot faster
19:27:43 <ais523> to execute, that is, realtime
19:27:54 <ehird> GregorR-L: oh, but the report didn't regenerate
19:27:56 * pikhq grumbles about (){}
19:28:07 * pikhq grumbles still more about (){}();
19:28:39 <ehird> GregorR-L: ok, remaining wishlist, ordered by importance first: (a) correct in-channel score reporting, (b) report doesn't disappear while regenerating, (c) timeouts don't take so long
19:28:46 <ehird> first two can be fixed by you, last needs ais523 probably
19:28:50 <impomatic> Are names automatically added to submissions now?
19:29:13 <GregorR-L> A) If you want the actual final score, this can't be done, or at least not quickly, it'd be slooow.
19:29:27 <GregorR-L> B) This is actually a big PITA >_>
19:29:30 <ais523> !bfjoust bigdecoy >(-)*9>(+)*9>>>>>>>[(+)*6[-]>+]
19:29:45 <EgoBot> Score for ais523_bigdecoy: -3 (maximum 7)
19:29:48 <ehird> (A) is very fixable
19:29:52 <ehird> It counts wins as losses.
19:29:53 <impomatic> !bfjoust stranger >>>>>>>>>([[-][(-)*127(+)*127]]>)*20
19:29:55 <ais523> and C) by my calculationsm, reducing the timeout to 20000 from 100000 should not affect too many programs
19:29:58 <ehird> Just fix that, dammit.
19:30:00 <EgoBot> Score for impomatic_stranger: -1 (maximum 8)
19:30:02 <GregorR-L> ehird: Oh, then it's ais523's fault :P
19:30:07 <ehird> GregorR-L: And how is (B) a big pitter?
19:30:14 <ehird> do >tmp; mv tmp foo
19:30:26 <GregorR-L> ehird: I kill the process generating a report when a new report starts.
19:30:41 <ehird> GregorR-L: "Don't do that"
19:30:45 <ehird> ais523: fix (A) and (C), anyway :P
19:31:02 <ehird> (B) probably isn't too important, but it's irritating having this trivial bug in-channel
19:31:08 <ais523> while ($steps++ < 100000) { and print "Timeout." if $steps >= 100000; are the only two lines that need changing to fix (C)
19:31:08 <ehird> and the timeouts are truly annoying
19:31:11 <ehird> making (B) more annoying
19:31:21 <ais523> for (A), the current code appears to run the program each way round and see if the answer was the same
19:31:25 <ais523> rather than looking at exit status
19:31:39 <GregorR-L> ais523: Right, because that's how FYB works :P
19:31:51 <ehird> ais523: if it can report score quickly
19:31:54 <ehird> how come the report takes so long
19:32:01 <ais523> ehird: because it runs all the programs against each other
19:32:02 <ehird> oh, because then it runs every other combination
19:32:03 <ais523> even the ones you didn't just add
19:32:08 <ais523> maybe that should be cached somehow
19:32:16 <EgoBot> Score for impomatic_simple: -1 (maximum 9)
19:32:46 <ais523> if we're in ridiculous wishlist mode, why not post a link to the report in-channel when it's finished generating?
19:33:21 <GregorR-L> ais523: I looked in to caching it and started rewriting report in Python to get non-argh SQL bindings to cache in an SQLite database.
19:33:40 <ehird> cache in a database?
19:33:48 <ehird> Just serialize some data structures.
19:33:50 <ehird> "import pickle" done
19:34:04 * ais523 wonders why bigdecoy did so badly
19:34:08 <pikhq> Why don't more people use interrobang‽
19:34:18 <ais523> because you're using it inappropriately when you are?
19:34:35 <impomatic> !bfjoust shield (>--)*3>((+)*12(>-)*4<<<<)*9999
19:34:43 <EgoBot> Program 2's flag fell.
19:34:49 <GregorR-L> Heh, whoops, forgot to get rid of that output :P
19:35:50 <EgoBot> Program 2's flag fell.
19:36:00 <ehird> !bfjoust bitchesdontknowboutwhichflagismine >+[-->+]
19:36:06 <EgoBot> Score for ehird_bitchesdontknowboutwhichflagismine: -9 (maximum 11)
19:36:11 <ais523> impomatic: how does your shield program work?
19:36:26 <ehird> GregorR-L: is it fixed?
19:36:29 <ehird> is -9 my actual score :D
19:37:43 <impomatic> ais523: not sure how it works, I've slept since I wrote it
19:37:57 <ais523> anyway, I just had a new idea for a defence program
19:38:03 <ais523> or rather, an old one, but I figured out how to make it work
19:38:50 <ais523> ehird: yes, your program really did do that badly
19:39:04 <ehird> ais523: run off the tape, I guess
19:39:16 <ehird> Uh, hey GregorR-L.
19:39:26 <ehird> 58.26-9ehird_bitchesdontknowboutwhichflagismine.bfjoust
19:39:26 <ehird> 70.00-3impomatic_shield.bfjoust
19:39:26 <ais523> and why the hey, now everything's working?
19:39:32 <ehird> GregorR-L: That's not good ordering, yo.
19:39:41 <ehird> ais523: actually, it's not
19:39:42 <ehird> 8 | + + + + + + + + + + + | 100.0| 11| impomatic_shortsword.bfjoust
19:39:43 <ais523> ehird: it's ordered by "Score"
19:39:51 <ais523> which seems to be something other than wins-losses
19:39:51 <ehird> his program is so good that it fucks up the tabs
19:39:57 <ehird> GregorR-L: add another tab after score
19:40:19 <ais523> you could avoid tab problems by always making the score 5 characters long
19:40:33 <GregorR-L> ehird: I don't understand, this output is exactly the same as FYB, and FYB is always aligned >_<
19:40:53 <ehird> GregorR-L: Because,.
19:40:58 <ehird> All others are one char less
19:41:08 <ehird> It's how tab characters work.
19:41:12 <ehird> You need another to pad out
19:41:28 <ais523> just bring all the scores up to 5 chars with padding
19:41:33 <GregorR-L> ehird: But that one is aligned to where the top is. If that one was the problem, it would be the maligned one.
19:41:35 <ais523> that way you could right-align them too at the same time, for free
19:41:46 <ehird> It is the misaligned one, GregorR-L.
19:41:53 <ehird> But what ais523 said.
19:41:54 <ais523> ehird: the title is also 5 chars
19:42:02 <ais523> so actually, it's correctly aligned, everything else is misaligned
19:42:39 <GregorR-L> I'm just saying I don't understand because the code is the same but always comes out perfect for FYB :P
19:44:06 -!- Sgeo has joined.
19:45:35 <ehird> http://codu.org/eso/fyb/report.txt
19:45:39 <ehird> There aren't any 100.0s there.
19:45:45 <ehird> So it's because no program is good enough.
19:46:03 <GregorR-L> ehird: But the 100 is the only one that's CORRECTLY aligned in bfjoust, LIKE I SAID
19:46:08 <impomatic> How about keeping an age for each program to keep track of how many successful challenges each one survives?
19:46:19 <impomatic> Then we could have a hall of fame ;-)
19:47:49 -!- M0ny has quit ("Read error: 182 (Connection reset by beer)").
19:50:11 -!- Sgeo has quit (Read error: 104 (Connection reset by peer)).
19:50:39 -!- Sgeo has joined.
19:50:55 <oerjan> !slashes http://oerjan.nvg.org/esoteric/slashes/counter3.sss
19:51:03 <impomatic> How come ehird's program scores more than shield? Ehird = 9 losses, 2 ties. Shield = 3 losses, 8 ties.
19:51:11 <EgoBot> Supported commands: addinterp bf_txtgen bfjoust daemon daemons delinterp fyb help info kill mush userinterps 1l 2l adjust asm axo bch bct befunge befunge98 bf bf16 bf32 bf8 bfbignum boolfuck c chiqrsx9p choo cintercal clcintercal cxx dimensifuck echo forth glass glypho google hello kipple lambda lazyk linguine malbolge ook pbrain perl qbf rail rhotor rot13 sadol sceql sh show slashes test trigger udage01 underload unlambda whirl yodawg
19:51:13 <ehird> 'cuz it's terrible
19:52:12 <impomatic> Surely a tie should be worth more than a loss?
19:52:15 <Sgeo> I hate collegeboard
19:52:16 <Sgeo> I email them saying I don't remember what I put for my security question's answer. They email me asking for personal information. Among the information they're asking for: "Web Account Security Answer:"
19:52:25 <ehird> Sgeo: I remember you saying that a few minutes ago in Sine too!
19:52:34 <ehird> Why not just crosspost everything you say to every IRC channel?
19:53:08 <GregorR-L> It's about WHAT you tie or lose against.
19:53:19 <oerjan> GregorR-L: !slashes is not working :(
19:53:20 <Sgeo> ehird, good idea!
19:53:29 <GregorR-L> impomatic: http://codu.org/eso/fyb/SCORES (same scoring used for this)
19:53:53 <EgoBot> Installed user interpreters: bct bfbignum chiqrsx9p choo echo google hello ook rot13 slashes yodawg
19:54:07 <oerjan> GregorR-L: previously i had a problem with it only giving the first line, it didn't use DCC
19:54:21 <oerjan> (although it worked with !show then)
19:54:51 <GregorR-L> However, I'll have to get to that after I fix this :P
19:55:13 <oerjan> GregorR-L: fix the backslashes then too pretty please
19:55:33 <GregorR-L> oerjan: I haven't fixed that because I don't know where the bug is.
19:56:06 <oerjan> !ook ++++++++[->++++++++<]>.
19:58:13 <pikhq> !addinterp echo_sh sh cat
19:58:13 <EgoBot> Interpreter echo_sh installed.
19:58:21 <pikhq> !echo_sh Hi there!
19:58:34 -!- KingOfKarlsruhe has joined.
19:59:24 -!- oerjan has quit ("Later").
20:00:40 <impomatic> Oh, undocumented instruction in BF Joust :-)
20:01:22 <EgoBot> Score for GregorR-L_attack1: -3 (maximum 10)
20:01:54 <impomatic> Ah, not undocmented I just didn't read the Wiki properly.
20:02:03 <ehird> which one did you mean?
20:03:26 <impomatic> I assumbed it was a * on the Wiki for some reason
20:03:56 <impomatic> 09/05/26 19:32:07 <impomatic> !bf joust simple [>[-]+]
20:04:12 <impomatic> 09/05/26 20:01:23 <GregorR-L> !bf joust attack1 [>[-]+]
20:05:15 <ehird> impomatic: "!bf joust"?
20:05:19 <ehird> what's up with your logs?
20:05:40 <ehird> !bfjoust attack1a [>[+]-]
20:05:42 <EgoBot> Score for ehird_attack1a: 2 (maximum 11)
20:05:52 <impomatic> I added the space in case egobot responded to what I pasted
20:05:59 <ehird> impomatic: only at the start of a line
20:06:32 <ais523> !bf joust ,[.,]!Hello, world!
20:06:36 <ais523> !bf joust,[.,]!Hello, world!
20:06:44 <ais523> hmm... input with ! no longer workd
20:09:35 <EgoBot> Score for impomatic_flux: 4 (maximum 12)
20:10:24 <ehird> !bfjoust flux_a_counteracting_monomorphism_cocktails (>)*8(>[+]-)*21
20:10:27 <EgoBot> Score for ehird_flux_a_counteracting_monomorphism_cocktails: 6 (maximum 13)
20:12:57 <ehird> ais523: how do you shorten nests, again?
20:13:41 <ais523> even if a and c contain the matching halves of square brackets
20:14:18 <impomatic> !bfjoust kicks_ehird >---->++++>-->++(>)*4(>++[-]+)*21
20:14:19 <EgoBot> Score for impomatic_kicks_ehird: 5 (maximum 9)
20:14:25 <GregorR-L> Gee, this is much faster with caching :P
20:14:30 <ehird> ais523: what if I want n=0->a, n=1->[ab],n=2->[[ab]b], etc?
20:14:34 <ehird> do I have to manually do that?
20:14:54 <ehird> GregorR-L: can you remove score? pts seems so much more useful
20:15:28 <GregorR-L> ... score is much better. Writing something that beats all the most useless programs but fails against any good ones shouldn't do well :P
20:15:41 <ehird> GregorR-L: then make the in-channel thing report score
20:16:16 <GregorR-L> I suppose now that report is caching, that's not wildly insane *sigh*
20:16:37 <ehird> !bfjoust modernist_decorum ([{[]}+])%30
20:16:44 <EgoBot> Score for ehird_modernist_decorum: -9 (maximum 10)
20:17:01 <ehird> It should constantly defend its flag.
20:17:49 <ehird> ais523: ([{[]}+])%blah is like [[[[[[]+]+]+]+]+], correct?
20:17:54 <ehird> !bfjoust modernist_decorum [[[[[[[[[]+]+]+]+]+]+]+]+]
20:18:01 <EgoBot> Score for ehird_modernist_decorum: -9 (maximum 10)
20:19:06 <ehird> ais523: so why does it fail?
20:20:20 <ehird> !bfjoust elena_lady_of_the_french_moving_picture_association (>+)*8[[-].>]
20:20:29 <EgoBot> Score for ehird_elena_lady_of_the_french_moving_picture_association: -8 (maximum 10)
20:20:41 <ehird> i should probably spend more time coming up with the programs than their name
20:23:31 <ehird> hey GregorR-L, what would I need to get a local version of the !bfjoust command?
20:23:42 <ehird> working the same way, updating a local scoreboard, etc
20:23:46 <ehird> I want to genetically evolve some proggies
20:23:57 <EgoBot> EgoBot is a bot for running programs in esoteric programming languages. If you'd like to add support for your language to EgoBot, check out the source via mercurial at https://codu.org/projects/egobot/hg/
20:24:09 <ehird> GregorR-L: That's... not helpful. At all.
20:24:39 <ehird> GregorR-L: I don't want a version that commits to an hg repository and is hooked up to IRC, is the point.
20:25:16 <GregorR-L> ehird: The report program in interps/bfjoust does exactly what you need, without any hgism or whatnot.
20:25:21 <ais523> ehird: all the programs needed are there
20:25:26 <ais523> there's glue to link them to IRC
20:25:27 <ehird> That's nice, then.
20:25:28 <ais523> but you can just not use it
20:28:12 <ehird> % perl5.10.0 bfjoust
20:28:12 <ehird> Both programs finished.
20:28:41 <ehird> ais523: #!/usr/bin/perl
20:28:44 <ehird> please fix that to #!/usr/bin/env perl
20:28:48 <ehird> my perl5.10 didn't run on it
20:28:52 <ehird> since it's in /opt/perl/bin
20:28:59 <ehird> not everyone has control over their vendor :)
20:29:15 <ais523> ehird: you could just run it with perl as a separate program
20:29:21 <ehird> but it's still a bug.
20:29:23 <ais523> not all computers have env
20:29:44 <ais523> (I have Perl and not env on my Windows computer at home running under DJGPP; it ignores the path in #! lines but parses the program)
20:29:45 <ehird> all UNIX-alikes and POSIX-alikes do, and Windows doesn't have /usr/bin
20:29:57 <ehird> ais523: then what's in the #! line doesn't matter to it
20:30:05 <ais523> because it'd try to run env, not perl
20:30:18 <ehird> ais523: then DJGPP is buggy
20:30:25 <ehird> 20:29 ais523: ehird: you could just run it with perl as a separate program
20:30:52 <GregorR-L> Incidentally, perl is particularly problematic since there's a flag you're always "supposed" to add, but #!/usr/bin/env perl can't add flags.
20:31:05 <ehird> GregorR-L: no, you're not meant to use perl -w
20:31:08 <ehird> you're meant to "use warnings;"
20:31:23 <ehird> -rwxr-xr-x 1 ehird staff 3500 26 May 20:27 bfjoust
20:31:23 <ehird> -rwxr-xr-x 1 ehird staff 3500 26 May 20:27 bfjoust.pl
20:31:27 <GregorR-L> Then I've been taught wrong oh noes
20:31:28 <ehird> ais523: GregorR-L: errrrrrr.
20:31:45 <ais523> ehird: "use warnings;" is scoped
20:31:58 <ehird> 20:31 ehird: -rwxr-xr-x 1 ehird staff 3500 26 May 20:27 bfjoust
20:31:59 <ehird> 20:31 ehird: -rwxr-xr-x 1 ehird staff 3500 26 May 20:27 bfjoust.pl
20:32:10 <ehird> Use: report <bfjoust program> <cache dir> <bfjoust files>
20:32:10 <ais523> ehird: so you can turn warnings on and off, obviously
20:32:13 <ehird> GregorR-L: That's a remarkably unhelpful help
20:32:19 <ehird> ais523: ARGH!! ARE YOU BLIND?!?!?!
20:32:21 <ehird> 20:31 ehird: 20:31 ehird: -rwxr-xr-x 1 ehird staff 3500 26 May 20:27 bfjoust
20:32:21 <ehird> 20:31 ehird: 20:31 ehird: -rwxr-xr-x 1 ehird staff 3500 26 May 20:27 bfjoust.pl
20:32:50 <ais523> ehird: I noticed that, but aren't replying to it
20:32:55 <GregorR-L> ehird: cd programs; ../report ../bfjoust ../cache *.bfjoust
20:33:00 <ehird> ais523: that was what my question was about
20:33:05 <ehird> so why did you answer a question I didn't ask?
20:33:33 <ehird> GregorR-L: Umm... and how do I get a certain program's score out of this?
20:33:46 <ehird> Additionally, how can I get the latest scoreboard of programs?
20:34:17 <GregorR-L> Hold your horses. I'm saving scores to files now, the scoreboard is spit out stdout.
20:35:32 <GregorR-L> Previously the scores were only in the stdout report, so don't complain :P
20:35:49 <ehird> GregorR-L: So what do I need to do?
20:42:07 <GregorR-L> That's the best you'll get due to my crippling apathy :P
20:42:57 <ehird> GregorR-L: So, cd programs; echo bitch>goddamn.bfjoust; ../bfjoust ../cache *.bfjoust; cat ../cache/goddamn.score?
20:43:27 <ehird> Now, GregorR-L, some questions: Does this trim the hill to 10 items? How do I get the current scoreboard programs in here?
20:45:09 <GregorR-L> This does trim the hill, yes. That command is ../report ../bfjoust ../cache *.bfjoust, and the scoreboard is the output of report.
20:45:43 <ehird> GregorR: No. I mean: how do I get the current egobot programs in there?
20:45:56 <GregorR-L> http://codu.org/eso/bfjoust/in_egobot/
20:46:00 <ehird> impomatic: have you got all your favorite warriors in EgoBot?
20:46:06 <ehird> GregorR: Ew. Fine.
20:46:34 <ehird> GregorR: Will removing the hill-trimming code make, say, the program think a warrior is great while EgoBot gives it a low score?
20:46:39 <ehird> i.e. will it bias my results
20:46:59 <GregorR-L> It could, but not by a lot, it gives more weight to beating programs that do well themselves.
20:47:20 <GregorR-L> What should I limit the hill to now? 10 is no longer necessary, and seems to small. 50 mebbe?
20:47:20 <ehird> GregorR-L: I'm just wondering whether to evolve with or without the limited hill
20:47:55 <ehird> ais523: i don't need to generate the ()/%/* stuff, right? I can just preprocess them in before sending the program to bfjoust
20:48:12 <GregorR-L> ehird: I would say fight against a static hill ...
20:48:20 <GregorR-L> ehird: That is, always delete your own.
20:48:28 <ehird> GregorR: err, but then I'll only beat the best one currently on the hill
20:48:39 <GregorR-L> OH, you want to evolve against yourself, right.
20:48:39 <ehird> if I add mine to the hill, I come up with a strategy that beats the current crop, then beat that, etc
20:48:46 <GregorR-L> In that case, definitely keep them all.
20:48:58 <ehird> GregorR: does that require intensive surgery?
20:49:00 <GregorR-L> Beating programs that don't do well doesn't earn you much score-wise.
20:49:08 <GregorR-L> Nope, just remove about five lines.
20:49:35 <GregorR-L> It should be painfully obvious what to remove.
20:49:47 <GregorR-L> And yes, I know, that's gross, but report.c is a hack :P
20:49:52 <ehird> GregorR: You said "without any hg stuff".
20:49:55 <ehird> system("hg commit -m -");
20:50:02 <ehird> I assume that, then, is not "hg stuff"...
20:50:07 <GregorR-L> That's the only time report.c uses hg. I forgot about that because it shouldn't :)
20:50:26 <ehird> GregorR: which hg repo has the programs?
20:50:31 <ehird> I don't want to download 'em manually.
20:50:47 <GregorR-L> wget -r -l inf -np http://codu.org/eso/bfjoust/in_egobot/
20:50:57 <ais523> I was just going to suggest wget, too
20:51:10 <ais523> GregorR-L seems to have more wget-fu than me, but mine is good enough for something like that
20:51:11 <ehird> GregorR: that ends up spidering the whole web
20:51:25 <ehird> ais523: it certainly downloads non-.bfjoust files
20:51:25 <ais523> there's an option not to spider above where you start
20:51:34 <ais523> ehird: yes, presumably it's the directory index in 20 different formats
20:51:41 <ehird> which kind of defeats the damn point...
20:51:53 <GregorR-L> ehird: Good freaking lord, stop complaining and mv *.bfjoust <elsewhere>
20:52:20 <ehird> GregorR: i'm irritated
20:53:25 <ehird> GregorR-L: report is idempotent, right?
20:53:34 <ehird> as in, running it without changing its args or the files will have no effect
20:53:40 <ehird> % mkdir ../cache;../report ../bfjoust ../cache *.bfjoust
20:53:42 <ehird> 00.000ais523_attack5.bfjoust
20:53:44 <ehird> 90.000impomatic_shortsword.bfjoust
20:53:46 <ehird> 80.000impomatic_kicks_ehird.bfjoust
20:53:48 <ehird> GregorR-L: cool story bro
20:54:14 <GregorR-L> So long as you removed that deletion, yes.
20:54:29 <ehird> GregorR-L: so everything's just lame huh
20:54:37 <ehird> No, because ais523 broke it.
20:54:55 <ehird> ais523: and "perl ../bfjoust" doesn't work with gregor's report program
20:54:58 <ehird> because it requires a filename
20:55:08 <ehird> ais523: so what did you say about "you can run it manually with 'perl bfjoust'?"
20:56:03 <ehird> % ../report ../bfjoust ../cache *.bfjoust
20:56:05 <ehird> 00.000ais523_attack5.bfjoust
20:56:07 <ehird> 90.000impomatic_shortsword.bfjoust
20:56:09 <ehird> GregorR-L: (yes, I fixed it)
20:56:19 <ehird> ../cache was broken.
20:57:23 <ehird> GregorR-L: d'you think removing the programs whose pts are below 0 is a wise decision?
20:57:29 <ehird> otherwise it'll fuss over them
20:58:18 <GregorR-L> The problem is that it may be a program that defeats one really "important" program, but doesn't do well otherwise.
20:58:23 <GregorR-L> And you'd want to keep that alive.
20:59:11 <GregorR-L> (Which is why the whole "score" vs "points" system exists :P )
20:59:43 <ehird> GregorR-L: ../cache never changes anything, right?
20:59:48 <ehird> As in, rm -rf ../cache only makes it slower
21:00:01 <GregorR-L> If you /update/ a program, you have to remove its relevant cache entries.
21:00:20 <ehird> Ahahahahahahahahaha GregorR-L and which are these
21:01:24 <ehird> GregorR-L: What about .score?
21:01:40 <ais523> $ ./bfjoust.pl attack1.bj defend6.bj | wc -l
21:01:43 <ais523> with 2 lines of output per step
21:01:50 <ais523> that means that 10000 is not enough
21:01:56 <ais523> but 20000 probably is, that was quite a long tape
21:01:59 <ehird> ais523: please, can you use .bfjoust instead of .bj? the connotations!
21:02:15 <ais523> but the connotations are only in your mind
21:02:27 <ehird> yes, well, the connotations of everything is only in your mind
21:04:30 <ais523> incidentally, defend6 beats shortsword too
21:04:36 <ais523> it's just /slightly/ too long for an IRC lien
21:05:01 <ais523> which is why I'm pasting it
21:05:08 <ehird> ais523: !bfjoust butt http://foo
21:05:59 <ais523> !bfjoust defend6 http://pastebin.ca/raw/1435349
21:06:14 <ehird> Wow, that's a fucked up codeulation, ais523.
21:06:17 <ehird> Care to expand it for us mortals?
21:06:21 <EgoBot> Score for ais523_defend6: 97.0
21:06:28 <ehird> ais523: just look at it!
21:06:36 <ais523> ehird: it consists of lots of not-quite-repeats
21:06:44 <ais523> if I expanded it, it would be massively long
21:06:49 <ais523> as there are no loops but the [] at the start
21:07:05 <ais523> GregorR-L: not surprising, they're both defence programs
21:07:11 <ais523> and wait for the opponent to arrive
21:07:34 <GregorR-L> Well, looks like you're on top now.
21:07:43 <ais523> maybe I should add in a defence-program-detector
21:07:45 <ehird> GregorR-L: So's your mom
21:07:54 <ais523> that switches to a counter-defence strategy if it notices one
21:08:04 <ehird> 295.8710ais523_defend6.bfjoust
21:08:05 <ais523> (a detector is easy, just wait for a few thousand turns and see if anything happens)
21:08:08 <ehird> GregorR-L: Why is it differen?
21:08:20 <ehird> I have the same programs.
21:08:26 <ais523> impomatic: you'll have to come up with something that can beat that, now I've knocked shortsword off the leaderboard
21:08:26 <ehird> Because I have 11 of them.
21:08:30 <ais523> ehird: tape length randomisation, too
21:08:38 <ehird> ais523: ugh, it's nondeterministic?
21:08:42 <ais523> it might make a difference, although probably not
21:08:56 <ais523> the randomisation's just to prevent people using degenerate strategies, and rarely changes the result
21:09:33 <ehird> there's probably an optimum strategy for a constant tape length N
21:09:56 <ais523> ehird: yes, just running straight to your opponent's flag
21:11:01 <ehird> [ehird:~/Code/scraps/2009-05/evolve_bfjoust/working/evolvist] % ruby evolvist.rb
21:11:03 <ehird> [ehird:~/Code/scraps/2009-05/evolve_bfjoust/working/evolvist] % ruby evolvist.rb
21:11:09 <ehird> GregorR-L: ais523: that's a pretty large variation...
21:11:14 <ehird> (testing impomatic_shortsword)
21:11:30 <ehird> maybe I should do it, say, 5 times, and take the average
21:11:36 <ehird> def remove_cache(name)
21:11:37 <ehird> `rm ../cache/*:#{name}.bfjoust:*`
21:11:41 <ehird> def test_program(name)
21:11:43 <ehird> remove_cache(name)
21:11:44 <ais523> ehird: if it's an evolutionary algorithm, a bit of randomness is fine anyway
21:11:45 <ehird> `../report ../bfjoust ../cache *.bfjoust`
21:11:47 <ehird> File.read("../cache/#{name}.bfjoust.score")
21:12:03 <GregorR-L> Any change there can only be from tape lengths ...
21:12:04 <ehird> i'm not goign to get anywhere with this
21:12:10 <ehird> it takes multiple seconds just to run once
21:12:25 <ehird> [ehird:~/Code/scraps/2009-05/evolve_bfjoust/working/evolvist] % time ruby evolvist.rb
21:12:26 <ehird> ruby evolvist.rb 2.16s user 0.13s system 97% cpu 2.342 total
21:12:28 <ehird> [ehird:~/Code/scraps/2009-05/evolve_bfjoust/working/evolvist] % time ruby evolvist.rb
21:12:29 <GregorR-L> ehird: Quay? With the cache in place it shouldn't?
21:12:32 <ehird> ruby evolvist.rb 1.78s user 0.13s system 98% cpu 1.936 total
21:12:41 <ehird> GregorR-L: it removes the cache for the program we're testing
21:12:45 <ehird> since it's to be used when you change it
21:12:54 <ehird> although, I suppose I won't ever reuse a prorgam
21:12:56 <GregorR-L> Oh, it's because your programs almost always time out, right? :P
21:13:16 <ehird> GregorR-L: anyway, you need <0.1s for evolving to be practiacl
21:13:21 <ehird> otherwise it'll take years
21:13:52 <ehird> GregorR-L: in short, make it faster :P
21:13:57 <ais523> ehird: just some programs are naturally very slow
21:14:03 <GregorR-L> Any slowness at this point is due to timeouts.
21:14:10 <ais523> defend6 wins really quite slowly, most of the time
21:14:10 <ehird> ais523: just get AnMaster to rewrite it
21:14:15 <ais523> juts turn on debug and look at the log
21:14:19 <ehird> it'll be really fast :p
21:14:22 <ais523> you could speed it up by not reparsing the program every step
21:14:25 <ais523> or in numerous other ways
21:14:52 <AnMaster> rewrite the bfjoust interpreter?
21:15:07 <AnMaster> No thanks. I'm not really interested in corewars or similar things
21:15:18 <ehird> Gee I sure was not joking
21:15:29 <ehird> "AnMaster" in string implies seriousville, true logic equation facst.
21:16:22 <ehird> GregorR-L: Hahahahaha
21:17:37 <ais523> ehird: as for that BF Joust program, I have a version with whitespace too which shows what it's doing
21:25:21 <ehird> ais523: that'd be nice
21:26:05 <ais523> ehird: http://pastebin.ca/1435376
21:29:09 <ais523> it attacks something in particular, which is the [-] or [+] loop
21:29:13 <ais523> nearly all programs have one
21:29:37 <ais523> and if they go into such a loop on its flag, then defend6 will keep them in it forever, whilst running off and sinking their flag
21:32:25 <EgoBot> Score for impomatic_tweaked: 37.5
21:32:49 <ais523> although it does beat defend6
21:33:28 <ais523> beating a program that beats all the others is quite good for its score...
21:34:06 <ais523> and kicks_ehird is amusing, it beats all the programs that don't start ais523_, and loses to all the ones that do
21:34:32 <Sgeo> Game of Werewolves (called Mafia) at irc.xkcd.com #mafia
21:34:49 <ais523> Sgeo: why did you just advertise that in #esoteric and ##nomic?
21:35:15 <Sgeo> I thought peoople there might be interested. Agora had a Werewolves thing
21:35:51 <ais523> impomatic: interesting strategy in stranger, by the way
21:36:00 <ais523> you seem to be thinking about counter-defence strategies
21:37:06 <impomatic> Hmmm... I've also slept since I wrote that one ;-)
21:37:18 <ais523> I can tell how it works, though
21:37:27 <ais523> detects defence strategies by looking to see if the flag has changed
21:37:45 <ais523> and trying up-runs and down-runs to see if either beats the defence strategy, before moving on
21:38:59 -!- bsmntbombdood has quit (Read error: 60 (Operation timed out)).
21:42:33 <ehird> "I don't know very English but I know very much Spanish and I suppose you don't know very Spanish but you know very English because you are American but I am not American."
21:42:52 <ehird> (Or should I say "Fidel Zzo38stro"? :D)
21:43:30 <ais523> wow, is that a genuine quote?
21:43:41 <ais523> except the grammar isn't quite as good
21:43:55 <ehird> ais523: it is genuine
21:43:56 <ehird> http://everything2.com/title/Young%2520Fidel%2520Castro%2527s%2520letter%2520to%2520President%2520Roosevelt
21:44:03 <ehird> beforehand he asks Roosevelt for a 10 dollar bill
21:44:14 <ehird> because he wants to see one, apparently
21:44:24 <ehird> he was, admittedly, 12 years old
21:45:20 <Slereah> http://www.underfoule.net/mika/src/12433705409.jpg
21:45:42 <ehird> Slereah: I see no /prog/snake; downvoted.
21:46:11 <Slereah> Where does the snake come from, anyway?
21:46:14 <ehird> http://images.cafepress.com/jitcrunch.aspx?bG9hZD1ibGFuayxibGFuazoyX0YuanBnfGxvYWQ9TDAsaHR0cDovL2ltYWdlczkuY2FmZXByZXNzLmNvbS9pbWFnZS8zNDEyMjU4OV80MDB4NDAwLmpwZ3x8c2NhbGU9TDAsMTcwLDE0NSxXaGl0ZXxjb21wb3NlPWJsYW5rLEwwLEFkZCwxNTUsMTI1fGNwPXJlc3VsdCxibGFua3xzY2FsZT1yZXN1bHQsMCw0ODAsV2hpdGV8Y29tcHJlc3Npb249OTV8
21:46:36 <ehird> Slereah: Because SICP.
21:46:39 <ehird> "force e = e" —/prog
21:49:44 -!- inurinternet has joined.
21:50:16 <Slereah> He is handsome and all, but
21:52:08 -!- bsmntbombdood has joined.
21:52:12 <ehird> Carts, cartographers,
21:54:50 <impomatic> !bfjoust dumb (>)*9([(-)*128.[-]]>)*20
21:54:52 <EgoBot> Score for impomatic_dumb: 50.0
21:55:11 <Slereah> http://twitter.com/progsnake
21:57:30 <Slereah> http://img87.imageshack.us/img87/1861/screenshotlm2.png
21:58:26 <Slereah> ehird, do you know, or do I have to ask /soc/?
21:59:31 <ehird> Slereah: It is just random.
21:59:47 <ehird> Also, your questioning is unscientific and ultimately destructive.
22:01:28 <Slereah> A lot of memes actually have some sort of origin
22:09:43 <ehird> Slereah: I'm waiting for the conclusion to your colon.
22:10:48 -!- bsmntbombdood has quit (Connection timed out).
22:11:48 -!- bsmntbombdood has joined.
22:12:19 <Slereah> The conclusion of my colon is my asshole
22:15:34 -!- MizardX has joined.
22:16:00 <ehird> [[Q: How do you vote in elections?
22:16:00 <ehird> A: I usually vote for the fattest candidate, on the basis that they'll take up more room on the House of Commons benches, thereby giving me more democracy for my valuable franchise.]]
22:18:09 <Slereah> I always vote for the tallest guy
22:20:08 <pikhq> I vote for the guy most likely to gum the works.
22:20:26 <pikhq> In this day and age, that means someone with reasonable opinions.
22:23:34 <ehird> "The Apple M9178 23-Inch Cinema HD Display is a liquid crystal display (LCD) monitor and has a default resolution of 1920x1200 dpi (dots per inch)."
22:23:39 <ehird> holy fucking shit!
22:23:52 <ehird> pikhq: You know we talked about that IBM display with hugh dpi?
22:23:56 <ehird> THIS IS THE REAL DEAL :P
22:24:06 <ehird> It's like a bajillion times more dense than paper!
22:24:37 <ais523> ehird: that's probably a bug
22:24:52 <ehird> it's a bug in the wetware of the human who authored it, rather.
22:26:20 <ehird> pikhq: Think how many VMs you could run.
22:26:28 <ais523> one in each square inch
22:26:31 <ais523> at a decent resolution
22:26:36 <ehird> A really decent resolution.
22:26:43 <ehird> It's pretty much the max anyone runs.
22:26:53 <ehird> Well, some people game at 2048x1536
22:26:59 <ehird> (Or was it 2560x1600?)
22:27:11 <pikhq> That's watching every single HD station at once.
22:27:18 <ehird> pikhq: The screen is 44160x27600
22:27:24 <ehird> And the total DPI is 3063
22:27:25 <pikhq> And a small handful of bluray discs.
22:27:32 <ehird> And 0.0083mm dot pitch
22:27:48 <ehird> pikhq: Haha. That's 8.3 microns separating each pixel.
22:29:05 <pikhq> I didn't know that there was an 8 micron silicon process out there.
22:29:31 <pikhq> Oh, the micron is larger than the nanometer.
22:29:49 <pikhq> That's still pretty impressive.
22:30:02 <ais523> wow, that's the first time I clicked on a link to Goatse
22:30:09 <pikhq> For comparison, a red blood cell is 7 microns.
22:30:12 <ais523> and I recognised it as Goatse before it finished loading, and managed to not look at it
22:30:31 <ehird> pikhq: 10 microns was the state of the art process in 1971-1972
22:30:37 <ehird> 3 microns was reached in 1975.
22:30:40 <ehird> But that's microchips.
22:30:45 <ais523> well, I recognised it wasn't what it claimed to be
22:30:52 <ais523> and guessed it was a shock image
22:30:55 <ehird> This would be the first monitor that requires you to spend inordinate amounts of money just to detect the pixel separation, pikhq :D
22:31:04 <pikhq> BTW, the proper terminology is the 'micrometer'. As in µm.
22:31:14 <ehird> pikhq: microns is also acceptable
22:31:23 <ehird> A micrometre or micron (American spelling: micrometer; symbol µm) is one millionth of a metre, or equivalently one thousandth of a millimetre. It can be written in scientific notation as 1×10−6 m, meaning 1/1 000 000 m.
22:31:35 <pikhq> Yes, but µm is something I can type that most can't. :p
22:32:10 * ais523 wipes the goatse from browser cache
22:32:41 * pikhq can't wait for picometer CPU processes
22:33:16 <ais523> and I didn't even copy-paste from you
22:33:26 <ehird> comex: this is not a nomic channel, before you say anything
22:33:34 <pikhq> ais523: Yes, but not for Windows users.
22:33:42 <comex> do you consent to ais523 joining bayes
22:33:44 <ehird> pikhq: Is picometer even possible? I mean, 11nm is nanotechnology.
22:33:45 <pikhq> Alt+0something or other.
22:33:48 <ehird> comex: I do not consent.
22:33:49 <ais523> comex: I'm not trying to
22:33:59 <comex> ais523: well, I asked you if you wanted to and you didn't respond
22:34:02 <ais523> and I suspect you need my consent too
22:34:05 <ehird> ais523: he wants to stop it being deregistered by announcement
22:34:13 <ais523> also, still wrong channel
22:34:14 <pikhq> A picometer process would have the circuit as a very complex molecule.
22:34:23 <ehird> comex: [[I deregister Bayes.]]
22:34:33 <ehird> pikhq: i'm not sure that would work too well
22:34:50 <pikhq> Lemme check to see how big a silicon atom is.
22:34:54 <ehird> helium atom = 31 picometers
22:35:09 <pikhq> 117.6 picometers for Si.
22:35:15 <ehird> pikhq: Heh. Silicon's a fatso.
22:35:43 <ehird> pikhq: So 0.1176nm.
22:35:45 -!- impomatic has left (?).
22:35:46 <pikhq> ... Arguably, we're already dealing with complex molecules.
22:35:59 <ehird> We're currently at 45/32nm
22:36:04 <ehird> 32nm can do RAM and shit, so let's say that
22:36:05 <pikhq> I mean, really. We're discussing traces that are about 100 atoms across.
22:36:29 <pikhq> (3 or 4 hundred across for common processes)
22:36:31 <ehird> pikhq: 32nm = 272 * a silicon atom
22:37:00 <ehird> pikhq: 11nm - nanoelectronics - will be 93
22:37:07 <ehird> so really, we're near that stage
22:37:26 <pikhq> IIRC, Intel was working on 20nm stuff.
22:37:56 <ehird> pikhq: On August 18, 2008, AMD, Freescale, IBM, STMicroelectronics, Toshiba and the College of Nanoscale Science and Engineering (CNSE) announced that they jointly developed and manufactured a 22 nm SRAM cell, built on a traditional six-transistor design on a 300 mm wafer, which had a memory cell size of just 0.1 square μm.[5] The cell was printed using immersion lithography.[6]
22:38:10 <ehird> pikhq: that's a bit too prototype for my tastes
22:38:36 <ehird> pikhq: But that's 187 * Si
22:39:06 <pikhq> I seem to recall Intel wanting to have their next generation on a 22nm process. Anyways.
22:39:20 <ehird> pikhq: the next tock will be a 32nm shrink of Nehalem
22:39:27 <ehird> and Intel are converting a fab to it
22:39:32 <ehird> so that's basically where we're at now
22:40:07 <ehird> Some predictions for the 22 nm node come from the ITRS. For example, it is predicted that silicon devices will no longer be planar, but will require ultrathin sections mostly surrounded on the sides by gates. The silicon body in each section is fully depleted, i.e., the free charge carrier concentration is deliberately suppressed. The sections basically protrude as fins from the surface (sometimes these are known as FinFETs). The creation of fins is a ne
22:40:11 <ehird> w challenge for the semiconductor industry, which has become accustomed to building transistors on a flat silicon surface. As of late 2008, several technical risks remain for implementation of non-planar 22nm transistors for logic applications.[2]
22:40:15 <ehird> According to the ITRS, the 22 nm node also marks the first time where the pre-metal dielectric, separating the transistor from the first metal layer, is a porous low-k material, replacing traditional, denser CVD silicon dioxide. The introduction of a porous material closer to the front end presents numerous integration challenges. In particular, the extent of plasma damage to low-k materials is typically 20 nm thick,[3] but can also go up to approximatel
22:40:22 <ehird> pikhq: tl;dr: "22nm requires some thinking before we can use it"
22:40:28 <ehird> whereas 32nm has no real issues vs 45nm
22:40:38 <ais523> they want to put the transistors vertically, sticking out from the chip
22:40:40 <pikhq> Still, very damned impressive.
22:40:41 <ehird> 16 nm resolution is difficult to achieve in a polymeric resist, even with electron beam lithography. In addition, the chemical effects of ionizing radiation also limit reliable resolution to about 50 nm, which is also achievable using current state-of-the-art immersion lithography. Hardmask materials and possibly iterated double patterning will be required.
22:40:43 <ais523> to help cooling and density, or something
22:40:45 <ehird> A more significant limitation comes from plasma damage to low-k materials. The extent of damage is typically 20 nm thick,[3] but can also go up to approximately 100 nm.[4] The damage sensitivity is expected to get worse as the low-k materials become more porous.
22:40:49 <ehird> and 11nm, well, that's nanotechnology
22:40:52 <ehird> quantum tunneling and shit
22:41:23 <pikhq> The point is, we're getting to that pretty quick.
22:43:46 <pikhq> Which is pretty damned spiffy.
22:44:29 <pikhq> Hmm. Intel is planning to start shipping 22 nm in 2011.
22:44:39 <pikhq> No guarantees of it actually happening, of course.
23:04:10 -!- tombom has quit ("Peace and Protection 4.22.2").
23:04:57 -!- FireFly has quit ("Later").
23:06:04 -!- BeholdMyGlory has quit (Remote closed the connection).
23:06:21 -!- coppro has joined.
23:13:42 -!- olsner has quit ("Leaving").
23:25:49 -!- ais523 has quit (Remote closed the connection).
23:29:08 <pikhq> BTW, way up the logs, s/Örjan/Ørjan/
23:46:45 <ehird> pikhq: intel say they'll get 11nm by 2015
23:47:04 <pikhq> ehird: Which is unlikely.
23:47:44 <ehird> nothing remotely close to the advanceness of what Intel wants to do at 11nm even exists in a "this could work" sketch for a prototype, as far as I know
23:48:14 <pikhq> Unless Intel is sitting on stuff. Which would be stunning to say the least.
23:48:36 <ehird> I find that incredibly unlikely.
23:48:53 <ehird> If they have 11nm stuff right now, they could make bajillions and woo everyone by demonstrating it
23:48:58 <ehird> Because it's literally sci-fi.
23:56:38 -!- AnMaster has quit (Connection reset by peer).
23:59:34 -!- AnMaster has joined.