00:00:10 <olsner> excel's "language" and VBA exist in localized versions
00:00:45 <maedhros777> It seems strange that all of them are in English
00:01:00 -!- pikhq has joined.
00:01:19 <soupdragon> I think a sanskrit lisp would be very nice
00:01:37 <soupdragon> and I've seen some people talking about ideas for a chinese APL
00:01:39 <pikhq> Welp, rebooting the router isn't helping HTTP any.
00:01:57 <pikhq> Chinese APL? That would work quite well.
00:02:07 <Sgeo> I think Python has a Spanish version
00:02:07 <olsner> it might be nice to take chinese characters and just use a random dictionary to map them to meanings in a programming language
00:02:25 <pikhq> Of course, I'm a guy who uses Chinese characters to write English. :P
00:02:58 <maedhros777> Look at this: http://blog.aegisub.org/2008/12/if-programming-languages-were-religions.html
00:03:30 <olsner> pikhq: nice! I was considering learning to do something like that, inventing my own english dialect using kanjis and translating all the linux software into it for sweet compactness and weirdness
00:04:26 <maedhros777> Another thought: how are accents represented in HTML?
00:04:27 <pikhq> Hmm. 漢字-書en programming 語... The 意 amuses 僕 大tly.
00:04:43 <pikhq> maedhros777: Depends on the encoding format in use.
00:05:22 <pikhq> Though the most general means of representing them is with the equivalent HTML entity. AKA a Unicode codepoint.
00:05:41 <olsner> pikhq: kanji-something-en programming language... the thought amuses me greatly?
00:06:05 <pikhq> olsner: Kanji-written programming language. The idea amuses me greatly.
00:06:48 <pikhq> I use 考 for thought, and 意 for idea. The mapping is a bit imperfect, I admit, but it kinda-sorta works.
00:07:19 -!- Tritonio_GR has quit (Read error: Connection reset by peer).
00:08:45 -!- maedhros777 has quit (Remote host closed the connection).
00:09:05 <olsner> the "idea" kanji looked similar to how I remembered japanese "thought", but after looking it up they're quite different
00:09:41 <olsner> they just share the "heart" radical(?) in the bottom
00:10:45 -!- pikhq_ has joined.
00:10:46 -!- pikhq_ has quit (Read error: Connection reset by peer).
00:12:11 -!- pikhq has quit (Ping timeout: 264 seconds).
00:15:48 -!- pikhq has joined.
00:16:21 <pikhq> GAH STOP DROPPING CONNECTIONS
00:16:51 <pikhq> How the hell is it that this damned thing manages to drop *specific connections*?
00:18:01 <pikhq> I cannot browse the web, but I can browse Gopher.
00:19:37 * Sgeo had that experience
00:19:43 <Sgeo> Can you get to https:// stuff?
00:20:19 <pikhq> What's a good https server to try?
00:20:54 <Sgeo> Used to use https://secondlife.com
00:21:31 * Sgeo looks for the login page for eBay
00:21:45 <Sgeo> https://signin.ebay.com/ws/eBayISAPI.dll?SignIn&ru=http%3A%2F%2Fwww.ebay.com%2F
00:25:19 <pikhq> Well, I almost lost my Freenode connection, and didn't get the page loaded.
00:27:16 <pikhq> Oh, and I can't get an IM connection going at all.
00:30:46 -!- pikhq has quit (Quit: leaving).
00:31:16 -!- biduzido has joined.
00:32:14 -!- maedhros777 has joined.
00:34:41 -!- pikhq has joined.
00:35:32 -!- Slereah has quit (Ping timeout: 248 seconds).
00:35:34 <pikhq> Nope, can't browse the web with Elinks. *While no other connections exist*.
00:36:15 <pikhq> I think I've got everything-but-what-normal-people-use Internet service.
00:36:41 <Sgeo> Tunnel HTTP over IRC!
00:36:54 <pikhq> Very, very tempting.
00:37:07 <pikhq> "Connection reset by peer". :(
00:37:39 <Sgeo> Can you connect to SSH?
00:37:57 <pikhq> Dunno; don't have a shell account handy.
00:38:17 <Sgeo> uorygl, emergency ping
00:38:32 <pikhq> Hmm. What was the normal HTTP request?
00:38:47 <Sgeo> GET / HTTP/1.1
00:38:52 <pikhq> I'd look it up, but I can't.
00:39:18 <Sgeo> And you need to include the Host: header
00:41:07 <maedhros777> Does anyone know if there are any Gravity compilers/interpreters out there?
00:41:13 <maedhros777> http://esoteric.voxelperfect.net/wiki/Gravity
00:41:39 <Sgeo> pikhq, I'm getting someone to contact uorygl to see if I would be allowed to grant you a temporary account
00:41:53 -!- Slereah has joined.
00:42:16 <Sgeo> maedhros777, I thought it was mathematically impossible to make a Gravity interp
00:42:26 <Sgeo> At least, one that will actually run
00:42:27 <soupdragon> maybe you could approximate an interpreter
00:43:00 -!- pikhq_ has joined.
00:43:13 <maedhros777> Isn't that, like, the definition of a programming language, that it is run?
00:43:40 -!- pikhq has quit (Ping timeout: 260 seconds).
00:43:53 <pikhq_> Well, I'm having difficulty getting a response *using netcat to google*.
00:45:00 <pikhq_> Hmm. Anyone know of an easy way to discover if it's MTU issues?
00:45:53 <maedhros777> Oh yeah -- Sgeo, how do you do red text in IRC?
00:46:08 <Sgeo> maedhros777, it only looks red because your name was mentioned
00:46:16 <Sgeo> Genuinely red text is different
00:46:31 <soupdragon> Content-Type: text/html; charset=UTF-8
00:46:32 <Sgeo> pikhq_, did you see my privmsg
00:46:43 <Sgeo> maedhros777, I see it as red
00:46:53 <Sgeo> pikhq_, any luck?
00:46:53 <pikhq_> soupdragon: I have managed to get responses out of Google. *Only to invalid requests*.
00:47:01 <maedhros777> I was always wondering why that happened :)
00:47:11 <Sgeo> pikhq_, bad connection, or did I mistype the password?
00:48:42 <Sgeo> pikhq_, so should I delete the account?
00:51:47 -!- pikhq_ has quit (Ping timeout: 264 seconds).
00:52:40 -!- pikhq has joined.
00:53:03 <pikhq> Well, that was succesful at making me lose every single connection.
00:53:13 <pikhq> Each and every one.
00:54:10 <Sgeo> And you weren't successfully able to actually use it, though, I guess
00:54:17 <pikhq> debug1: Remote protocol version 2.0, remote software version OpenSSH_4.7p1 Debian-8ubuntu1.2
00:54:20 <pikhq> debug1: match: OpenSSH_4.7p1 Debian-8ubuntu1.2 pat OpenSSH_4*
00:54:22 <pikhq> debug1: Enabling compatibility mode for protocol 2.0
00:54:25 <pikhq> debug1: Local version string SSH-2.0-OpenSSH_5.3
00:54:27 <pikhq> debug2: fd 3 setting O_NONBLOCK
00:54:30 <pikhq> debug1: SSH2_MSG_KEXINIT sent
00:54:30 <pikhq> And then... *Nothing*.
00:54:32 <pikhq> debug3: Wrote 792 bytes for a total of 813
00:54:43 * Sgeo has no clue what any of that means
00:54:44 <pikhq> Yes, it stopped working during the key exchange.
00:55:15 <pikhq> The connection was made, but everything stopped working as soon as key negotiation began.
00:55:21 <Sgeo> So can I delete the account?
00:55:33 <pikhq> Yeah, I can't use it at all.
00:55:55 <olsner> oh, you're using encryption? must be something subversive then!
00:56:07 <maedhros777> Does anyone know of a program to generate LaTeX code, or a good tutorial?
00:56:39 <soupdragon> you just write \sqrt{...} for square roots and \frac{...}{...} for division
00:56:44 <soupdragon> that's all you need to know realyl.....
00:56:59 <soupdragon> at lesat that's all I know and I get by
00:57:29 <Ilari> host header is mandatory in HTTP/1.1.
00:57:52 -!- pikhq has quit (Read error: Operation timed out).
01:00:16 -!- pikhq has joined.
01:00:38 <pikhq> I've got a theory that it's the router's fault.
01:01:12 <Sgeo> pikhq, if you sent something containing HTTP-like stuff to IRC, what would happen?
01:01:14 <Ilari> pikhq: So nc www.googele.com 80 -> 'GET http://www.google.com/<ENTER>Host:<ENTER><ENTER>' hangs?
01:01:27 <olsner> fiddle furiously with the router settings for content filtering, firewall functions and intrusion prevention/detection
01:01:30 <pikhq> ifconfig certainly isn't showing any errors.
01:01:52 * Sgeo notes that pikhq is probably glad that he's computer-savvy
01:02:01 <Sgeo> Or at least, savvy enough to use IRC
01:02:11 <olsner> maedhros777: yes, no other way to fiddle really
01:02:12 <maedhros777> * Maedhros777 is wondering Sgeo makes text blue
01:02:48 <pikhq> Okay, watching ifconfig...
01:02:56 <Sgeo> Is +c color filtering?
01:03:11 <pikhq> No errors. No dropped packets. Just an inexplicable lack of a response.
01:03:55 <Sgeo> GET http://google.com is meaningless
01:04:22 <Sgeo> Anyone want to -c the channel?
01:05:32 <olsner> woah, who put swedish in the topic?
01:07:17 -!- pikhq has quit (Read error: Connection reset by peer).
01:09:22 -!- pikhq has joined.
01:09:43 <pikhq> Okay, now with the only IRC connect being to Freenode.
01:10:19 <pikhq> And let's make the only room I'm in #esoteric.
01:10:58 <pikhq> AAGH WHY DONT I HAVE TCPDUMP
01:12:03 <Sgeo> pikhq, what happens if you disable all firewalls?
01:12:42 <pikhq> Sgeo: I am behind no firewalls ATM.
01:13:12 <pikhq> I am directly connected to the modem, and I have no iptables rules, and the policy for INPUT, FORWARD, and OUTPUT is ACCEPT.
01:13:46 <Sgeo> Is there a DMZ thing?
01:13:57 <Sgeo> Also, try a different OS?
01:14:02 <pikhq> I am DIRECTLY ATTACHED TO THE MODEM
01:14:29 <Sgeo> Um, different distro then, or at least LiveCD?
01:14:35 <Sgeo> See if it's some configuration issue
01:14:35 <pikhq> And lack the means to *install* a different OS *because I CANT MAKE ANY NON-IRC CONNECTIONS*
01:15:19 <pikhq> I'm strongly suspecting that there's an MTU issue.
01:16:01 <Sgeo> Uh, hown do I USE dcc chat on this
01:16:02 <Ilari> pikhq: After local updates finish, I could maybe help with testing...
01:16:09 <myndzi> dcc isn't an irc connection
01:16:18 <myndzi> it is Direct Client Connection
01:16:25 <myndzi> by definition you connect directly to the target :)
01:16:37 <myndzi> rather, they connect to you (if you are sending (unless you use "passive" dcc))
01:16:56 <pikhq> From the looks of things, any sufficiently large packets are completely and utterly ignored?
01:17:14 <myndzi> well at least you can change the packet size with /dcc packetsize
01:17:32 <pikhq> This is consistent with how the only HTTP response I have gotten is an error page.
01:17:33 <Sgeo> Ok, anyone want to send pikhq a LiveCD?
01:17:47 <olsner> hmm, I think there's a tool for discovering the smallest mtu between you and some other host
01:17:55 <Sgeo> Or invent HTTP-over-DCC?
01:17:56 <Ilari> pikhq: MTU problems would cause that. More specifically, MTU problems on return path.
01:17:57 <myndzi> i don't have one, and also my upload is slooow
01:18:26 <Ilari> pikhq: What happenss if you set interface MTU to 1280?
01:18:28 <pikhq> Ilari: Fucking wonderful. Fucking *wonderful*.
01:18:37 <pikhq> Ilari: Same thing.
01:18:43 <pikhq> I actually have been playing with my MTU.
01:18:49 <Sgeo> I don't have anything new, but I'll try
01:18:58 <pikhq> Because I have actually had this happen before when my MTU was set wrong.
01:19:35 <pikhq> If the other end has its MTU set wrong, I am going to hunt down my ISP and kill some people.
01:19:55 <Sgeo> Hm, maybe I shouldn't bother sending a livecd
01:20:48 <Sgeo> pikhq, do you at least see me trying to send?
01:20:48 <pikhq> How do you accept a DCC send with irssi?
01:21:52 <Ilari> Ugh. Locale generation is slow and I apparently have lot of them enabled.
01:22:54 <pikhq> Hmm. Anyone willing to set up an HTML server with an obscenely low MTU?
01:23:13 <Sgeo> Is there a way to change MTU in Apache?
01:23:22 <pikhq> Or maybe I could try and figure out what the MTU *is* for this link?
01:23:32 <pikhq> Sgeo: "ifconfig eth0 mtu number-goes-here"
01:23:48 <olsner> http://en.wikipedia.org/wiki/Path_MTU_discovery
01:23:50 <Sgeo> pikhq, um, is there a risk of losing my ssh connection?
01:24:13 <pikhq> olsner: Thanks. Thanks for the link that I cannot read.
01:24:42 <Sgeo> Ok, what MTU shall I set?
01:25:06 <pikhq> Sgeo: 1000 I guess?
01:25:09 <Sgeo> The original MTU of Normish is 1500
01:25:15 <Ilari> 1280 is high enough for IPv6 but low enough not to usually be a problem...
01:25:36 <Sgeo> But now I can't seem to type into ssh
01:25:37 <pikhq> Hmm. Could someone google the management IP for a "ViaSat Surfbeam satellite modem"?
01:25:45 <Sgeo> http://normish.nomictools.com
01:26:28 <Sgeo> I think I broke Normish
01:26:56 <Sgeo> Oh crap, I typoed
01:27:09 <pikhq> "Connection reset by peer"
01:27:11 <Sgeo> I forgot mtu, so I did sudo ifconfig eth0 1000
01:27:23 <pikhq> What are the *other* possible causes of that?
01:27:41 -!- biduzido has quit (Quit: Leaving).
01:28:51 * Sgeo blames pikhq for my typo
01:30:25 <pikhq> Anyone? "Connection reset by peer" causes?
01:31:15 <olsner> pikhq: tracepath says that the path-mtu between you and me is 1500
01:31:47 <pikhq> olsner: Okay, so it's not the MTU, then.
01:33:19 <Sgeo> So I killed Normish for nothing?!?
01:33:52 -!- pikhq has quit (Read error: Connection reset by peer).
01:37:07 -!- pikhq has joined.
01:38:08 <Ilari> pikhq: Can I send few pings there for testing?
01:38:53 <pikhq> I've been using ping -s to google.
01:38:57 <pikhq> And getting truncation.
01:39:36 <maedhros777> Does anyone know how to use a sigma in LaTeX?
01:40:03 <Ilari> 1024 byte IP packets seem to work OK. Latencies are slow.
01:41:02 <pikhq> Well, I do have a very, very high-latency link.
01:41:04 <Ilari> So do 1500, and return packets do not seem truncated.
01:41:36 <pikhq> What's your IPv4 address, Ilari?
01:41:46 <olsner> wow, I've never seen ping times this high before
01:42:00 <pikhq> olsner: Satellite-tastic link.
01:42:26 <olsner> you must really be in the middle of nowhere
01:42:32 <olsner> move somewhere they have internet!
01:42:34 <Ilari> pikhq: You know how to do 6to4 address to ipv4 conversion?
01:42:50 <pikhq> Ilari: No, I'd look it up but I can't.
01:43:25 <pikhq> 12 packets transmitted, 0 received, 100% packet loss, time 10999ms
01:43:34 <olsner> that's weird, I can ping you
01:43:48 <pikhq> Your host is probably set to drop ICMP?
01:43:59 <olsner> ah, right, I probably am
01:45:45 <olsner> so that means a TCP connection takes at least 4-6 seconds to set up?
01:46:04 <pikhq> More if DNS hasn't been cached.
01:47:27 -!- poiuy_qwert has quit (Read error: Connection reset by peer).
01:47:28 <Ilari> pikhq: 2002:5870:32ae::1 -> 587032ae -> 1483748014.
01:47:41 <pikhq> Ilari: Thank you for not being helpful.
01:48:04 <pikhq> Could you stick that in a more useful format, like 4 octets?
01:48:31 <pikhq> ... Wait. You can just use the full number, can't you?
01:48:44 <pikhq> 8 packets transmitted, 0 received, 100% packet loss, time 6999ms
01:48:50 -!- poiuy_qwert has joined.
01:49:46 <pikhq> Hmf. No truncation when pinging slashdot.
01:50:16 <pikhq> But nor is there any actual access to slashdot via HTTP.
01:50:36 <pikhq> I've got half a mind to take a sledgehammer to a computer.
01:50:53 -!- maedhros777 has quit (Remote host closed the connection).
01:51:09 <Ilari> pikhq: There is small HTTP server on port 7682.
01:54:51 <olsner> maybe they just use a timeout < 6s :)
01:54:56 -!- pikhq has quit (Ping timeout: 260 seconds).
01:55:17 -!- pikhq_ has joined.
01:55:58 <Ilari> pikhq_: <Ilari> pikhq: There is small HTTP server on port 7682.
01:56:47 <Ilari> Saw request (200 OK reply).
01:57:04 <pikhq_> I'm certainly not getting the reply.
01:57:22 <Sgeo> On Slashdot, or something else?
01:58:21 <pikhq_> Ilari: How large was the reply?
02:01:53 <Ilari> pikhq_: The page size is about 900 bytes.
02:02:26 -!- pikhq_ has quit (Read error: Connection reset by peer).
02:02:42 -!- pikhq has joined.
02:05:14 <Ilari> Gah. I can't use tcpdump (I have it installed but trying to use it sends load average through the roof).
02:05:30 <pikhq> And I cannot has tcpdump.
02:08:54 <Ilari> pikhq: Well, I set firewall to log all outbound IPv4 packets.
02:11:14 -!- pikhq_ has joined.
02:11:42 -!- pikhq has quit (Read error: Connection reset by peer).
02:13:37 <Ilari> pikhq_: Maybe try requesting it again (I should see the outbound packets in firewall logs now)...
02:17:52 <Ilari> I see what looks like repeated TX attempts of 52 byte ACK + 98 byte ACK/PSH.
02:19:17 <Ilari> ... DST=75.106.103.227 LEN=98 TOS=0x00 PREC=0x00 TTL=64 ID=56999 DF PROTO=TCP SPT=7682 DPT=46258 WINDOW=432 RES=0x00 ACK PSH URGP=0
02:19:46 <Ilari> ... DST=75.106.103.227 LEN=52 TOS=0x00 PREC=0x00 TTL=64 ID=61552 DF PROTO=TCP SPT=7682 DPT=46259 WINDOW=432 RES=0x00 ACK URGP=0
02:19:47 -!- pikhq has joined.
02:21:00 -!- pikhq_ has quit (Read error: Connection reset by peer).
02:21:22 <Ilari> I also saw this: DST=75.106.103.227 LEN=938 TOS=0x00 PREC=0x00 TTL=64 ID=27331 DF PROTO=TCP SPT=7682 DPT=46261 WINDOW=432 RES=0x00 ACK PSH FIN ... No re-TX seen.
02:21:50 <pikhq> What was there before "I also saw this"?
02:21:56 <pikhq> 20:13 < Ilari> pikhq_: Maybe try requesting it again (I should see the outbound packets in firewall logs now)...
02:22:56 <Ilari> I see what looks like repeated TX attempts of 52 byte ACK + 98 byte ACK/PSH.
02:23:30 <pikhq> So this implies... What?
02:23:57 <Ilari> Also hmm... What was that 938 byte ACK/PSH/FIN packet there about? I didn't see any retransmits...
02:23:58 <pikhq> My ISP is about as reliable as smoke signals from on top of an active volcano?
02:25:06 -!- sshc has quit (Remote host closed the connection).
02:26:22 <Ilari> Obiviously not MTU problem when packets that small have problems and packets of 1.5kB make through...
02:27:08 -!- pikhq has quit (Read error: Connection reset by peer).
02:27:18 -!- pikhq has joined.
02:29:41 <Ilari> pikhq: Also, why IRC is pretty much only thing that works (it is unstable, but...)?
02:29:54 <pikhq> IRC is presumably low-bandwidth.
02:30:02 <pikhq> I can't connect to irc.foonetic.net.
02:30:15 <pikhq> But still, mostly I can do IRC.
02:34:46 -!- pikhq has quit (Read error: Connection reset by peer).
02:34:58 -!- Gregor-L has quit (Remote host closed the connection).
02:35:02 -!- pikhq_ has joined.
02:35:09 <AnMaster> Ilari, try wireshark as root, generally it works, probably the tcpdump issue is just a command line flag issue or such
02:35:20 <AnMaster> of course, wireshark will complain loudly
02:35:39 <AnMaster> it is up to you to decide if you want to do that
02:35:52 <Sgeo> Is there a way to make my .avis take up much less space?
02:36:23 <AnMaster> Sgeo, ogg theora, mpeg, or whatever
02:36:34 <AnMaster> Sgeo, iirc avi is uncompressed
02:36:39 <Ilari> AnMaster: I can't run graphical apps as root.
02:36:55 <AnMaster> Ilari, or ksu or whatever kde has
02:37:36 <AnMaster> Ilari, for me graphical apps for root work with "sudo programname" but not "sudo -s" or su or "sudo su -"
02:38:06 <Sgeo> How do I get Archive Manager to compress to maximum possible extent?
02:38:34 <Sgeo> the one with GNOME
02:38:42 <AnMaster> uh no idea, I use cmd line tools
02:38:55 <Ilari> AnMaster: None of those work (I am not authorized for those).
02:39:10 <AnMaster> Ilari, eh? so you don't have root access locally?
02:39:21 <AnMaster> Ilari, then you are pretty much fucked
02:39:59 * Ilari deletes the firewall log rule...
02:41:03 <AnMaster> Sgeo, something like tar -jcf foo.tar.bz2 foo for bzip2, not sure if it is max, but it would be trivial to do something like: tar -cf - foo | xz -z9e - > foo.tar.xz
02:41:10 <AnMaster> that may not work exactly like that
02:41:15 <AnMaster> haven't test run the latter command
02:41:26 <AnMaster> but it should work, or something pretty close to it
02:41:48 * Sgeo somehow doubts it will get 12GB of video down to less than 1GB
02:41:53 <AnMaster> oh and if you want zip I have absolutely no clue about anything except unzip
02:42:00 -!- pikhq_ has quit (Read error: Connection reset by peer).
02:42:08 <AnMaster> Sgeo, not without lossy compression
02:42:22 <Sgeo> Ok, how do I do lossy compression on the command line?
02:42:25 <AnMaster> so you want mpeg or ogg theora or such
02:43:02 <Ilari> Well, IIRC gzip could compress ~59GiB of raw video data (video dump from game) to about 2GB...
02:43:58 <AnMaster> Sgeo, I don't know, I would read the manual for the the relevant programs. In this case that is probably ffmpeg or such
02:44:11 <Ilari> Well, it has large single-color surfaces, which really tends to compress well...
02:44:25 <AnMaster> Sgeo, oh and -9e to xz might not be a good idea, I wouldn't be surprised if it took days
02:44:33 <AnMaster> something like just -9 might be better
02:45:53 <Sgeo> tgz -9cf . > futurama.tgz
02:45:58 <Sgeo> Why is that not working like I want?
02:46:34 <AnMaster> Sgeo, I have no clue what tgz is
02:46:42 <AnMaster> it is an unknown command to me
02:46:56 <coppro> tgz doesn't accept options
02:47:27 <AnMaster> coppro, someone made an alias for tar -z?
02:47:42 <coppro> AnMaster: no, for tar | gzip
02:47:48 <coppro> which is strictly worse
02:47:57 <AnMaster> coppro, well passing -f and redirecting stdout makes no sense at all
02:49:29 -!- maedhros777 has joined.
02:49:46 <Sgeo> With just tgz futurama.tgz, will it be much larger than I want/
02:50:17 -!- maedhros777 has quit (Remote host closed the connection).
02:51:51 -!- pikhq has joined.
02:52:22 <pikhq> Hey, look. Freenode's back.
02:52:29 <pikhq> I doubt any other Internet is.
02:54:11 <pikhq> Nope. Still don't have *the fuckinsadfgnklasdjgh;blohzsadvjhnasb;hg web*
02:54:34 <AnMaster> Sgeo, tar -zcf futurama.tgz dirwithfuturama
02:54:45 <AnMaster> Sgeo, because then it might try to include itself
02:54:47 <pikhq> Sgeo: .avis are compressed.
02:54:52 <AnMaster> and yes it would be much bigger
02:55:13 <pikhq> A .avi file is just a container for some video compression format.
02:55:19 <Sgeo> AnMaster, should I include 9?
02:55:23 <AnMaster> pikhq, sure? I'm pretty sure they were uncompressed, or perhaps it can be uncompressed as well?
02:55:33 <AnMaster> Sgeo, tar doesn't take that option afaik, check man page
02:55:36 <pikhq> AnMaster: It's a container format.
02:55:44 <AnMaster> pikhq, and can it contain uncompressed?
02:55:54 <pikhq> Yes, that is one such codec.
02:56:01 <pikhq> Almost *nobody* uses it for raw video though.
02:56:04 <Ilari> pikhq: Maybe phone ISP tech support? Or is the average IQ below CMB temperature there?
02:56:44 -!- lament has joined.
02:59:10 <Sgeo> AnMaster, I'm not getting status on the command line the way I did with tgz
02:59:43 <Sgeo> AnMaster, I want to see the progress as it's being made.. wait, that's -v, isn't it?
03:00:13 <AnMaster> but if as pikhq said it is compressed then I doubt this will help at all
03:00:26 -!- pikhq has quit (Ping timeout: 264 seconds).
03:02:08 <Sgeo> Indeed, it seems to not be helping
03:02:09 -!- poiuy_qwert has quit (Quit: Leaving).
03:02:26 <Sgeo> So, how do I reencode the videos?
03:02:34 <Sgeo> Basically, I just want to play these things on my phone
03:07:14 -!- pikhq has joined.
03:15:26 -!- lament has quit (Ping timeout: 264 seconds).
03:21:31 -!- lament has joined.
03:23:00 -!- Gregor-L has joined.
03:29:24 -!- lament has quit (Ping timeout: 248 seconds).
03:30:58 <pikhq> Well, I guess my IRC connection has stabled out.
03:31:07 <pikhq> I'm still not getting any not-IRC.
03:39:18 <Sgeo> pikhq, can you telnet?
03:39:33 <Sgeo> try telnet alt.nethack.org
03:39:48 -!- sshc has joined.
03:40:08 <Sgeo> telnet nethack.alt.org
03:41:14 <pikhq> Lagtastic, but yes.
03:44:13 <Sgeo> pikhq, irc.xkcd.com#xkcd
03:44:26 <pikhq> And I cannot connect to Foonetic.
03:45:07 <coppro> You can't just switch ISPs?
03:45:13 <Sgeo> Maybe I can bridge a channel on FreeNode with Foonetic
03:45:27 <pikhq> coppro: If I could I would.
03:45:36 <pikhq> Actually, I can *connect* to Foonetic.
03:45:52 <pikhq> Joining #xkcd causes the connection to drop.
03:46:47 <Sgeo> pikhq, can you connect to any other channels on Foonetic?
03:47:39 <Sgeo> So, join #pikhq and I'll direct anyone who wants to help to there?
03:48:55 <pikhq> God dammit it autojoined.
03:49:04 <pikhq> But I'm still connected.
03:49:41 <Sgeo> pikhq, are you receiving chat?
03:50:02 <pikhq> Nothing from #xkcd after "<@res0> puddle dried up"
03:50:11 <pikhq> And... Lagging out.
03:50:23 <Sgeo> Ok, I'll set up a bridge
03:50:40 <Sgeo> #pikhq here will connect to #pikhq there
03:53:03 <pikhq> Gah. At the rate this is going I'm starting to think some asswad at the ISP is actually monitoring for connections that use "too much bandwidth" to drop them or something.
04:03:55 <Sgeo> pikhq, ping, you're still in Foonetic, right?
04:06:50 <Sgeo> pikhq pikhq pikhq
04:07:20 -!- soupdragon has quit (Quit: soupdragon).
04:08:37 <Sgeo> pikhq, why are you not responding in Foonetic???
04:19:03 -!- pikhq_ has joined.
04:20:27 -!- pikhq has quit (Read error: Connection reset by peer).
04:24:16 -!- FireFly has joined.
04:27:43 -!- pikhq has joined.
04:28:42 -!- pikhq_ has quit (Read error: Connection reset by peer).
04:36:16 -!- pikhq_ has joined.
04:37:01 -!- pikhq has quit (Read error: Connection reset by peer).
04:42:47 -!- pikhq has joined.
04:43:17 -!- pikhq_ has quit (Read error: Connection reset by peer).
04:45:58 -!- lament has joined.
04:49:11 -!- pikhq has quit (Read error: Operation timed out).
04:50:21 -!- pikhq has joined.
04:51:01 <pikhq> So, apparently I have 88% packet loss.
04:51:18 <pikhq> I'm amazed I can do IP.
04:51:40 <uorygl> How are you connected to your ISP?
04:51:52 <Gregor-L> You can "do IP" with 99% packet loss :P
04:52:02 <pikhq> uorygl: I send my packets TO SPACE
04:52:12 <pikhq> And a satellite sends them back FROM SPACE
04:52:18 <uorygl> Satellite? Sounds laggy.
04:52:30 <pikhq> Yes, I normally have 2-second pings.
04:55:54 <uorygl> At least it isn't Freenet.
04:56:09 <pikhq> Yeah, not as laggy as Freenet.
04:56:25 * uorygl ponders what the mail system would be like if it were like Freenet.
04:56:36 <uorygl> <Gregor-L> It would be like Freenet.
04:57:04 <uorygl> I guess people wouldn't receive messages until they look in their mailboxes.
04:57:11 <uorygl> Which is already how it works, so I guess it would be exactly the same.
04:57:29 <uorygl> Satellite Internet: Not As Laggy As The United States Postal Service.
04:58:57 -!- pikhq_ has joined.
04:59:21 -!- pikhq has quit (Read error: Connection reset by peer).
05:08:22 -!- pikhq_ has quit (Quit: Plugging in the router).
05:10:45 -!- pikhq has joined.
05:11:50 -!- oerjan has joined.
05:16:06 <Sgeo> pikhq, Internet success?
05:16:20 <pikhq> I appear to have issues with #xkcd still. XD
05:16:33 <pikhq> But I can browse t3h webs!
05:17:22 <Sgeo> pikhq, you are in #xkcd
05:19:01 <oerjan> <maedhros777> I've just been wondering -- are there any esoteric or real programming languages in another language (besides English)?
05:19:12 <oerjan> i recall an icelandic one being discussed
05:19:22 <pikhq> There's Perligata.
05:19:48 <oerjan> skrifastreng ("writestring") was one of the commands, iirc
05:20:17 <oerjan> which was enough to google it, it's called Grunnur
05:20:41 <oerjan> that was just another command
05:20:58 <oerjan> http://en.wikipedia.org/wiki/Fj%C3%B6lnir_(programming_language)
05:21:46 <oerjan> which also leads us to http://en.wikipedia.org/wiki/Category:Non-English-based_programming_languages
05:22:24 <oerjan> although a lot of those are esolangs not based on a real natlang at all :D
05:23:01 -!- adu has joined.
05:23:55 <oerjan> i guess Zuse's Plankalkül is also pretty famous
05:24:32 <pikhq> Yup, /join #xkcd kills my Freenode link.
05:24:33 <oerjan> (and presumably german, although the wikipedia sample doesn't really show it)
05:29:27 * coppro is reading about GOTO++. If he translates the French correctly, each program is compiled to an intermediate representation and a custom VM to interpret it
05:41:40 -!- gm|lap has joined.
05:53:24 <gm|lap> downloading ghc because hugs interactive mode is ass
05:53:39 <gm|lap> at least from what i can establish
05:56:28 <gm|lap> oh dear, i just realised what i typed
05:58:41 <pikhq> ICANN is doing non-ASCII TLDs.
05:59:00 <oerjan> it's all greek to me. or was that arabic.
05:59:54 <gm|lap> actually if unicode gets a cheeseburger glyph
06:00:03 <pikhq> http://xn----rmckbbajlc6dj7bxne2c.xn--wgbh1c/ This is the hostname of the Egyptian Ministry of Communications and Information Technology.
06:00:04 <gm|lap> then ICANN has (cheeseburger)
06:01:12 -!- Oranjer has left (?).
06:14:06 <Gregor-L> gm|lap: I wurve the man command.
06:25:44 -!- Gregor-L has quit (Remote host closed the connection).
06:28:18 -!- oerjan has quit (Quit: leaving).
06:29:06 -!- kar8nga has joined.
06:35:50 -!- HackEgo has quit (Remote host closed the connection).
06:35:50 -!- EgoBot has quit (Remote host closed the connection).
06:35:53 -!- HackEgo has joined.
06:35:53 -!- EgoBot has joined.
06:38:12 -!- kar8nga has quit (Remote host closed the connection).
06:46:49 -!- kar8nga has joined.
06:50:00 -!- Gregor-L has joined.
07:03:50 -!- tombom has joined.
07:25:40 -!- kar8nga has quit (Remote host closed the connection).
07:30:13 -!- asiekierka has joined.
07:37:30 <asiekierka> "code esoteric language do make real brain make"
07:37:42 <asiekierka> "Esoteric language code is actually generated by the brain"
07:40:46 <asiekierka> now, "talk make esoteric malbolge do do and create code generation"
07:41:06 <coppro> doesn't always work: http://translationparty.com/#7352244
07:41:44 <lament> i got a 2-cycle http://translationparty.com/#7352245
07:41:55 <coppro> http://translationparty.com/#7352249
07:42:08 <lament> and the site doesnt test for cycles :(
07:42:29 <asiekierka> lament, do you want to lament about it? :P
07:43:02 <lament> http://translationparty.com/#7352252
07:44:05 <coppro> http://translationparty.com/#7352257
07:44:05 <asiekierka> "talk make esoteric malbolge do do and create code generation"
07:44:23 <asiekierka> gives "Malbolge Thus, the code generator can create a powerful statement"
07:44:57 <coppro> http://translationparty.com/#7352261
07:47:24 <asiekierka> http://translationparty.com/#7352274 - a certain volcano you all know
07:47:37 <asiekierka> gives The Castle District in the North Fiji fatal Raja
07:48:12 <gm|lap> http://translationparty.com/#7352277
07:50:45 -!- tombom has quit (Quit: Leaving).
07:50:52 <gm|lap> http://translationparty.com/#7352287
07:51:12 <coppro> this one is deep: http://translationparty.com/#7352286
07:51:58 <gm|lap> http://translationparty.com/#7352295 vs http://translationparty.com/#7352293
07:59:35 -!- asiekierka has quit (Ping timeout: 264 seconds).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:00:34 <coppro> I have managed to get it to produce a word that it cannot translate in either direction
08:02:03 <coppro> I've also yet to produce an input including both 'quark' and 'gluon' which doesn't cause it to fail to translate at least one word
08:02:12 <lament> http://translationparty.com/#7352347
08:02:32 <lament> kanji makes it into english
08:02:52 -!- asiekierka has joined.
08:03:34 * Sgeo should sleep now
08:04:18 <pikhq> My goodness. Actual kanji.
08:04:53 <pikhq> "Stars hidden in a crowd of his [face-haven]"?
08:07:45 -!- Sgeo has quit (Read error: Connection reset by peer).
08:14:04 -!- asiekierka has quit (Ping timeout: 276 seconds).
08:16:57 -!- asiekierka has joined.
08:21:08 -!- asiekierka has quit (Ping timeout: 240 seconds).
08:30:54 -!- MigoMipo has joined.
08:37:37 -!- MizardX has quit (Ping timeout: 276 seconds).
08:54:09 -!- BeholdMyGlory has joined.
08:57:52 <gm|lap> ugh... is translationparty broken?
09:35:28 -!- lament has quit (Ping timeout: 276 seconds).
09:59:28 -!- Tritonio_GR has joined.
10:22:26 -!- pikhq has quit (Read error: Connection reset by peer).
10:32:42 -!- pikhq has joined.
10:51:41 -!- pikhq_ has joined.
10:52:24 -!- pikhq has quit (Ping timeout: 246 seconds).
11:31:38 -!- cheater3 has joined.
11:34:05 -!- cheater2 has quit (Ping timeout: 246 seconds).
11:37:03 -!- Tritonio_GR has quit (Quit: Leaving.).
11:45:17 -!- gm|lap has quit (Quit: ilua).
11:46:38 -!- Tritonio_GR has joined.
12:01:00 -!- Tritonio_GR has quit (Ping timeout: 246 seconds).
12:09:32 -!- coppro has quit (Quit: sleeeeeep).
12:30:29 -!- pikhq_ has quit (Read error: Connection reset by peer).
12:31:39 -!- pikhq has joined.
12:31:55 -!- pikhq has quit (Read error: Connection reset by peer).
12:42:45 -!- pikhq has joined.
12:43:16 -!- pikhq has quit (Read error: Connection reset by peer).
12:54:08 -!- pikhq has joined.
12:54:50 -!- pikhq has quit (Read error: Connection reset by peer).
12:58:26 -!- pikhq has joined.
12:59:01 -!- pikhq has quit (Read error: Connection reset by peer).
13:04:16 -!- pikhq has joined.
13:05:09 -!- pikhq has quit (Read error: Connection reset by peer).
13:13:44 -!- pikhq has joined.
13:29:51 -!- BeholdMyGlory has quit (Remote host closed the connection).
13:47:08 -!- MigoMipo has quit (Remote host closed the connection).
14:10:59 -!- oerjan has joined.
14:37:59 -!- pikhq has quit (Read error: Connection reset by peer).
14:41:27 -!- pikhq has joined.
14:41:30 -!- pikhq has quit (Read error: Connection reset by peer).
14:46:30 -!- pikhq has joined.
14:53:09 -!- pikhq_ has joined.
14:54:01 -!- pikhq has quit (Ping timeout: 258 seconds).
14:56:29 -!- pikhq_ has quit (Read error: Connection reset by peer).
14:59:44 -!- pikhq has joined.
15:06:26 -!- pikhq_ has joined.
15:07:15 -!- pikhq has quit (Ping timeout: 240 seconds).
15:12:49 -!- pikhq_ has quit (Read error: Connection reset by peer).
15:13:15 -!- pikhq has joined.
15:14:32 -!- augur has quit (Remote host closed the connection).
15:19:36 <AnMaster> I have an idea for how to make an AI. It is probably crazy. And someone probably thought of it already...
15:20:36 <AnMaster> (oh and it probably won't work)
15:21:56 -!- pikhq_ has joined.
15:22:21 <oerjan> the latter being a pretty safe assumption
15:22:23 <AnMaster> basically simulate evolution with genetic algorithms, don't aim for AI that can pass the turing test right away, rather aim for something that can respond to it's simulated environment, say, along the lines of a very very simple bacteria, when you have something reasonable there, extend the fitness test to require something more
15:22:27 -!- pikhq has quit (Ping timeout: 240 seconds).
15:22:34 -!- adu has quit (Quit: adu).
15:23:13 <oerjan> well i'm sure it's been thought of before yeah...
15:23:14 <AnMaster> presumably you could work yourself up to something that could seem like a simple AI in some hundred years. Unlikely to take as long as real evolution since computers are rather fast.
15:23:48 <oerjan> i didn't say i knew who did it
15:23:56 <oerjan> just that it's too obvious
15:24:25 <AnMaster> oerjan, heard that story of Columbus and the egg?
15:24:27 <oerjan> i suspect the problem is you are enormously underestimating the complexity of environment needed to evolve intelligence
15:25:00 <AnMaster> (there is even a separate wikipedia page on it heh)
15:25:39 <oerjan> i've heard one such story, which fits the subject so i assume that is it
15:26:14 <AnMaster> oerjan, hm I did consider complexity but I wasn't sure how bad it would be, bad yes but no idea how bad.
15:26:15 <oerjan> ("anyone could have thought of that" "but they didn't!" essentially)
15:26:22 <AnMaster> but probably I underestimated it yes
15:27:02 <oerjan> also i think there is a theory that much of the final intelligence evolution was due to social interaction between humans
15:27:32 <AnMaster> oerjan, considering you can get interesting things with breading FPGAs (famous example is that circuit able to detect two different tones, without using a reference clock)
15:27:45 <oerjan> presumably we tried to evolve to outsmart each other :)
15:28:10 <oerjan> (seeing as the other highly intelligent animals are also highly social)
15:28:12 <AnMaster> oerjan, okay so you could introduce that in your environment after you get something that is able to be aware of other AIs in that environment
15:28:33 -!- pikhq has joined.
15:28:40 <AnMaster> for speed reasons you could probably skip that during the initial stages, but it should probably be introduced not too late either
15:29:11 -!- pikhq_ has quit (Ping timeout: 240 seconds).
15:29:46 <AnMaster> oerjan, but yes, it probably would take too much resources to be feasible :/
15:30:34 <AnMaster> oerjan, on the other hand I suspect designing an AI with "traditional engineering" may be unfeasible for humans.
15:31:22 <oerjan> as long as we don't know how _humans_ manage to be intelligent, sure
15:31:42 <AnMaster> oerjan, heck if you wanted to evolve AIs like that you would probably have to make the environment (the fitness function) dynamic too. I mean, consider on earth, life changed the environment a lot
15:32:13 <AnMaster> stuff like production of oxygen I mean
15:32:14 <oerjan> yes. although maybe artificial evolution could skip some of the long static periods.
15:32:46 <AnMaster> oerjan, what do you mean here?
15:33:08 <AnMaster> (I can think of several interpretations of that statement)
15:33:49 <oerjan> life didn't evolve fast _all_ the time, because the environment didn't always change much. punctuated equilibrium, it's called afair
15:35:42 <AnMaster> oerjan, but it still changed, surely you can't just keep the same generation and extrapolate the environment 10 million years? It seems quite possible that then everything will be unfit. I mean, small gradual changes in environment are not too hard to adapt to, compared to sharp changes.
15:35:45 -!- pikhq has quit (Read error: Connection reset by peer).
15:36:17 -!- pikhq has joined.
15:36:49 <oerjan> well species _do_ occasionally die out when they are unfit for a new environment, i presume
15:37:18 <AnMaster> oerjan, even if nothing much happened it seems unlikely to have been completely static, and perhaps some random mutation in the skipped generations would have caused far reaching consequences.
15:37:27 <oerjan> although i don't know how much is environment and how much is new competition
15:37:42 <AnMaster> oerjan, just think of production of oxygen, iirc things were pretty static before that for quite some time.
15:38:01 <oerjan> well naturally not completely static there are _all_ kinds of cycles and events
15:38:48 <AnMaster> btw, I doubt actually simulating the chemical earth is a good idea for such AI evolution.
15:40:02 <AnMaster> oerjan, ooh new idea that ensures a high probability of friendly AI! However it requires some technology we don't yet have. And I have no idea how much storage space it needs...
15:40:24 <AnMaster> oerjan, basically, model a human. Every molecule
15:40:41 <oerjan> primates presumably evolved some of that brain to be able to distinguish a lot of different foods (fruit) visually, they have an extremely varied (and time-varying) diet
15:41:35 <oerjan> because not all forest plants bear fruit at the same time
15:41:45 <oerjan> especially in the tropics
15:42:15 <AnMaster> actually, just simulating a cell down to molecular level should be interesting enough and quite hard.
15:42:42 <oerjan> i'm not entirely convinced emulating humans is the best recipe for friendliness. especially once you try to go _beyond_ human intelligence
15:42:45 <AnMaster> and might help find out how feasible doing anything more advanced would be (I guess it wouldn't at all)
15:43:05 <AnMaster> oerjan, hm true. but at least you would get an AI passing the turing test
15:43:08 <oerjan> we're not known for being friendly with inferior species
15:44:52 <AnMaster> bbl, strange sound from computer
15:45:29 -!- poiuy_qwert has joined.
15:45:42 -!- pikhq has quit (Ping timeout: 246 seconds).
15:48:12 -!- Gregor-L has quit (Ping timeout: 252 seconds).
15:55:57 -!- pikhq has joined.
16:00:53 -!- Gregor-L has joined.
16:07:10 -!- pikhq has quit (Ping timeout: 240 seconds).
16:12:28 -!- pikhq has joined.
16:14:17 -!- lament has joined.
16:15:07 -!- pikhq has quit (Read error: Connection reset by peer).
16:17:36 -!- pikhq has joined.
16:17:37 -!- pikhq has quit (Read error: Connection reset by peer).
16:22:41 -!- pikhq has joined.
16:30:35 -!- benuphoenix has joined.
16:31:20 <benuphoenix> I just spent three-fourths of my savings on dvds and printer ink last night. Now, I finally have more in my checking account than in my savings account. This sucks.
16:33:18 <benuphoenix> oh, and main(){cout << "I'm an idiot that still uses iosttream." << endl;}
16:33:43 -!- pikhq_ has joined.
16:34:21 -!- pikhq has quit (Ping timeout: 246 seconds).
16:38:24 <benuphoenix> I have a question about brainfuck. How does one output text to standard output using it?
16:38:55 -!- Phantom_Hoover has joined.
16:43:07 -!- pikhq has joined.
16:43:10 -!- pikhq_ has quit (Ping timeout: 240 seconds).
16:43:21 <benuphoenix> I gotta log off. I'm using my blackberry...
16:46:21 <benuphoenix> sorry about telling the world how broke I am. I'm just pissed at myself for not realizing that having $70 in savings is a bad idea.
16:46:47 -!- benuphoenix has left (?).
16:47:15 <Phantom_Hoover> I love the way that xkcd's irc bot isn't much more sophisticated than our very own fungot.
16:47:15 <fungot> Phantom_Hoover: it's time you jumped off this mortal coil... are you all right? of course! you were so young! you ran around saying " daddy! the children are going?! the king hast been injured? what's the big deal? so what if we won a war out there! can't it see i love my daddy! the children are going!
16:48:30 <fungot> Available: agora alice c64 ct* darwin discworld europarl ff7 fisher ic irc jargon lovecraft nethack pa speeches ss wp youtube
16:48:38 -!- kar8nga has joined.
16:48:54 <fungot> ^<lang> <code>; ^def <command> <lang> <code>; ^show [command]; lang=bf/ul, code=text/str:N; ^str 0-9 get/set/add [text]; ^style [style]; ^bool
16:49:04 <fungot> ^<lang> <code>; ^def <command> <lang> <code>; ^show [command]; lang=bf/ul, code=text/str:N; ^str 0-9 get/set/add [text]; ^style [style]; ^bool
16:50:06 -!- benuphoenix has joined.
16:50:23 -!- benuphoenix has quit (Remote host closed the connection).
16:50:40 <fungot> Available: agora alice c64 ct* darwin discworld europarl ff7 fisher ic irc jargon lovecraft nethack pa speeches ss wp youtube
16:50:47 <fungot> Selected style: youtube (Some YouTube comments)
16:50:52 <fungot> Phantom_Hoover: holy crap, that was just making that up in time.
16:52:06 -!- pikhq has quit (Ping timeout: 265 seconds).
16:52:07 <fungot> Selected style: lovecraft (H. P. Lovecraft's writings)
16:52:12 <poiuy_qwert> lol, in some programming channel: "<***> its like brainfuck to me"
16:52:33 -!- benuphoenix has joined.
16:52:47 <fungot> oerjan: " you, mr. de marigny and phillips, across the putrid moat and under the prevailing stimulus my son wingate would often go up to different heights and scan the fnord waste for signs of dim, fumbling terror about the way he knew it would be
16:55:29 * Phantom_Hoover is going to do horrible things to his X server so as to play Portal
16:56:25 -!- pikhq has joined.
16:57:24 <poiuy_qwert> yay im on my way to making my IRC bot in an esoteric language!
16:58:26 -!- Phantom_Hoover has quit (Remote host closed the connection).
16:58:28 -!- pikhq has quit (Read error: Connection reset by peer).
16:58:40 -!- lament has quit (Ping timeout: 240 seconds).
17:01:53 -!- pikhq has joined.
17:12:32 -!- pikhq has quit (Read error: Connection reset by peer).
17:12:58 -!- pikhq has joined.
17:25:09 -!- tombom has joined.
17:33:04 -!- benuphoenix has left (?).
17:36:40 -!- poiuy_qwert has quit (Quit: Leaving).
17:42:44 -!- AnMaster has quit (Ping timeout: 248 seconds).
17:51:15 -!- jcp has quit (Ping timeout: 240 seconds).
17:56:01 -!- jcp has joined.
18:00:29 -!- AnMaster has joined.
18:15:56 -!- MizardX has joined.
18:17:24 -!- AnMaster has quit (Ping timeout: 248 seconds).
18:21:09 -!- BeholdMyGlory has joined.
18:21:25 -!- tombom_ has joined.
18:22:00 -!- AnMaster has joined.
18:22:40 -!- tombom__ has joined.
18:23:13 -!- sebbu2 has joined.
18:24:15 -!- MigoMipo has joined.
18:24:20 -!- tombom has quit (Ping timeout: 240 seconds).
18:26:00 -!- tombom_ has quit (Ping timeout: 246 seconds).
18:26:20 -!- sebbu has quit (Ping timeout: 240 seconds).
18:31:22 -!- ais523 has joined.
18:39:23 -!- AnMaster has quit (Read error: Operation timed out).
18:41:47 <pikhq> Calling all alise. Calling all alise.
18:41:53 <pikhq> Is there an alise in the room?
18:43:50 -!- oerjan has changed nick to MIB.
18:44:03 <MIB> There is no such thing as an alise
18:44:07 -!- MIB has changed nick to oerjan.
18:45:03 <fizzie> Swamp gas, weather balloons, so on.
18:45:40 -!- augur has joined.
18:48:15 <pikhq> http://www.ioccc.org/1984/mullender.c I love obfuscated C.
18:49:35 -!- kar8nga has quit (Remote host closed the connection).
19:00:12 <ais523> hmm, anything really interesting happen while I was gone?
19:00:21 <ais523> preferably eso-wise, but I don't mind too much if it's offtopic
19:01:07 <pikhq> Well, I did have odd network behavior wherein I could access IRC and Gopher and nothing else.
19:10:08 -!- poiuy_qwert has joined.
19:13:39 -!- Gregor-L_ has joined.
19:14:34 -!- AnMaster has joined.
19:17:12 -!- Gregor-L has quit (Ping timeout: 252 seconds).
19:18:18 -!- Gregor-L_ has quit (Ping timeout: 252 seconds).
19:18:45 -!- Gregor-L_ has joined.
19:19:37 <ais523> pikhq: that sounds amusing
19:20:09 <pikhq> Good God it was annoying.
19:22:10 -!- poiuy_qwert has quit (Ping timeout: 240 seconds).
19:26:52 -!- poiuy_qwert has joined.
19:54:38 <AnMaster> I think my adsl modem is having hardware issues
19:56:00 <AnMaster> after rebooting it (due to it becoming unresponsive to dns, then a few minutes later connections dying and finally the webui of the modem itself going down), it had reset about half of the settings
19:56:06 <AnMaster> some with garbage, some with factory defaults
19:56:28 <AnMaster> now I did a clean install of it, and found out I can't let it reboot after it, nor can I restore config
19:56:38 <AnMaster> so there are some settings I can't change
19:56:41 -!- Sgeo has joined.
19:56:59 <AnMaster> ais523, pikhq: ever heard about anything like that?
19:57:20 <ais523> but then, I tend not to mess with modems/routers, I just leave them be
19:57:31 <AnMaster> ais523, well I was leaving it alone when this happened
19:57:52 <AnMaster> ais523, plus I can't use defaults, no password for admin and wep for wlan... :P
19:57:57 <AnMaster> no way I will leave a modem like that
19:58:17 <AnMaster> plus I need to enter PPPoE settings anyway
20:01:11 <AnMaster> ais523, plus it worked okay before...
20:07:20 -!- Alex3012 has quit (Read error: Connection reset by peer).
20:07:29 -!- soupdragon has joined.
20:07:57 -!- Alex3012 has joined.
20:22:38 -!- Phantom_Hoover has joined.
20:23:01 -!- soupdragon has quit (Quit: soupdragon).
20:23:17 -!- soupdragon has joined.
20:23:28 <Phantom_Hoover> If I make a filesystem on a file, then mount it, then try to move the file into the mounted filesystem, what happens?
20:24:54 <Phantom_Hoover> The file would have to be larger than the free space in the filesystem.
20:25:07 <pikhq> Phantom_Hoover: Either the system call fails, or a copy of the filesystem is stuck in the file system, and the reference count of the filesystem's file is decremented.
20:25:21 <pikhq> Meaning that once you unmount the filesystem, it completely disappears.
20:26:39 <pikhq> Same place any non-referenced inodes and blocks go: the free list.
20:26:47 <AnMaster> Phantom_Hoover, hm squashfs is read-only isn't it?
20:27:53 <AnMaster> pikhq, also which system call(s) fails? open/read/write/close or unlink?
20:28:09 <AnMaster> well the size thing makes it probably the former group
20:28:16 -!- Guest_854 has joined.
20:28:32 <pikhq> AnMaster: ... rename?
20:28:47 <AnMaster> pikhq, it works between filesystems?
20:29:01 <AnMaster> oh maybe the copy and remove is done in kernel then
20:29:42 <pikhq> rename will return EXDEV, which means the move command does open, read, write, close, and unlink.
20:29:49 -!- hiato has joined.
20:29:53 <pikhq> So, it's the second behavior I described, not the first.
20:30:20 <AnMaster> pikhq, but it would fail since unless the fs is compressed it will be larger than the space available on that disk
20:30:29 <AnMaster> and compressed fs have a tendency of being read-only
20:30:40 <pikhq> AnMaster: He never said anything about it being compressed.
20:30:49 <pikhq> And many, many filesystems support sparse files.
20:31:09 <AnMaster> pikhq, hm good point, but doesn't those need some special command to create it?
20:32:12 <AnMaster> Phantom_Hoover, you have to remember mv basically copies the file then removes the original when moving between different mount points
20:32:24 -!- Guest_854 has quit (Quit: He who laughs last, thinks slowest).
20:32:43 <pikhq> AnMaster: GNU mv attempts to keep files sparse.
20:32:48 <AnMaster> within a mount point it will just be moved in file hierarchy by the kernel in some way (probably varies between different FS)
20:32:53 <pikhq> And no, there's no "special command" to create it.
20:33:19 <pikhq> Instead, you just seek and write.
20:33:20 <AnMaster> pikhq, so how does it try to keep them sparse?
20:33:52 <AnMaster> well that explains why the copy worked
20:34:07 <pikhq> Phantom_Hoover: mv copies the file to the new filesystem. It then unlinks the file in the old filesystem.
20:34:17 <AnMaster> Phantom_Hoover, the image on there is likely corrupt since probably it was written to while mv was reading it
20:34:20 <pikhq> All that an unlink does, though, is remove the name for the inode.
20:34:49 <pikhq> Any program that had the file open is still reading from that inode.
20:35:02 <pikhq> When no programs access it, then the inode is marked free.
20:35:14 <AnMaster> pikhq, correction: if it was the last hardlink the file will be marked as free
20:35:36 -!- benuphoenix has joined.
20:35:37 <pikhq> AnMaster: Right, right. You can have multiple names attached.
20:35:40 <AnMaster> but the bit about hardlinks is important, otherwise it would cause havoc.
20:35:53 <AnMaster> pikhq, there is no way to find all attached names is there though?
20:35:57 <Phantom_Hoover> I can read and write files in the mounted fs without corruption.
20:36:09 <pikhq> AnMaster: Sure there is. Grep.
20:36:15 <AnMaster> Phantom_Hoover, expected, what I said was that the image stored on that fs is now corrupted
20:36:28 <AnMaster> pikhq, hm or find /mountpoint -inode or such I bet
20:36:45 <pikhq> Phantom_Hoover: Yes, that's because your mounted filesystem is still on the original filesystem.
20:36:45 <AnMaster> iirc you can get the inode with either some ls option or with stat(1)
20:36:52 <pikhq> It just no longer has a *name*.
20:37:20 -!- hiato has changed nick to sheep.
20:37:22 <pikhq> Yes, that's what *we've been fucking telling you*.
20:37:28 -!- sheep has changed nick to hiato.
20:37:37 <pikhq> A UNIX filesystem is reference counted.
20:37:45 <AnMaster> pikhq, hm can you have more than one name for a symlink? as in ln -s /bin foo; ln foo bar
20:37:46 <fizzie> "ls -i"'s the list-inode option; it's the easiest to remember.
20:37:57 <pikhq> AnMaster: Filesystem-dependent.
20:38:28 <AnMaster> I only have ext2/4 handy, and I'm not about to mess around on /boot to try ext2 behaviour
20:38:43 <AnMaster> fizzie, how is it easier to remember than ls -l ?
20:38:48 <pikhq> AnMaster: I'm pretty sure that ext2 filesystems shove the symlink data in the filename structure.
20:38:55 <pikhq> Well. The file structure.
20:38:56 <AnMaster> at least ls -l you use sometimes
20:39:15 <pikhq> (along with the filename, permissions, extended attributes, etc.)
20:39:27 <AnMaster> pikhq, so that means it wouldn't or would work?
20:39:39 <pikhq> AnMaster: Wouldn't, as a symlink does not have an inode.
20:39:47 <fizzie> AnMaster: The superlative may have been superfluous. Let's settle for "easy to remember".
20:41:18 <AnMaster> doh anyway this /boot is ext3, I wonder why
20:41:30 <AnMaster> pikhq, ext4 shows inode numbers for ls -i
20:41:54 <fizzie> "Ext2fs implements fast symbolic links. A fast symbolic link does not use any data block on the filesystem. The target name is not stored in a data block but in the inode itself."
20:41:59 <fizzie> So it has an inode; it doesn't have data.
20:42:00 <AnMaster> and yeah ext2 I can't find any of atm, seems I have all my /boots as ext3
20:42:10 <pikhq> AnMaster: Ah, I'm having a thinko.
20:42:23 <pikhq> The inode *is* where the file stuff is stored.
20:42:26 <fizzie> (Though they say the fast symlinks have a maximum length limit of 60 chars. Might be dependant on inode size though.)
20:42:44 <pikhq> And a hardlink is a new inode that points to the same blocks as another one.
20:43:08 <AnMaster> pikhq, where is the filename then, is it not stored with the inode?
20:43:17 <AnMaster> hm probably in directory structure
20:43:24 <pikhq> Directory structure.
20:43:29 <AnMaster> also doesn't ext4 store very small files inline in the inode or something?
20:43:46 <AnMaster> so hardlinks wouldn't work for those?
20:43:48 <pikhq> Ext4 also doesn't use blocks for very *large* files when it can.
20:44:04 <pikhq> No, I think hardlinking will force ext4 to create a block.
20:44:23 <pikhq> Hardlinking *is* a system call. It can do that. ;)
20:44:40 <AnMaster> symlinking must be some system call too I guess
20:45:20 <fizzie> I'm not sure "hardlink is a new inode that points to the same blocks as another one" is true, because I believe it's the inode that stores the link count.
20:45:48 <AnMaster> and they have the same inode number
20:46:05 <AnMaster> yep test shows they get same inode number (regular file)
20:46:43 <AnMaster> "POSIX.1-2001 says that link() should dereference oldpath if it is a symbolic link. However, since kernel 2.0, Linux does not do so: if oldpath is a symbolic link, then newpath is created as a (hard) link to the same symbolic link file (i.e., newpath becomes a symbolic link to the same file that oldpath refers to)."
20:46:52 <fizzie> Also ls here seems to color "link count >1" files differently. If I have noticed this before, I've forgotten.
20:47:14 <AnMaster> fizzie, not here, but then ls on my laptop and desktop follow slightly different rules sometimes
20:47:42 <AnMaster> "POSIX.1-2008 changes the specification of link(), making it implementation-dependent whether or not oldpath is dereferenced if it is a symbolic link."
20:50:25 <fizzie> dircolors --print-database says "HARDLINK 44;37 # regular file with more than one link", so it at least provides a possibility to colorize it differently.
20:51:10 <fizzie> Also sockets and doors have the same color here. Not that I'm very likely to run across a door.
20:51:52 <AnMaster> btw wonderful word "a nondirectory"
20:52:06 <pikhq> One thing I absolutely *despise* about Wikipedia's new design.
20:52:26 -!- poiuy_qwert has quit (Read error: Connection reset by peer).
20:52:31 <AnMaster> pikhq, oh don't get me started on google's new design
20:52:34 <pikhq> They have an uncacheable Javascript file in the head of the file.
20:52:45 <pikhq> That takes *forever* to load.
20:53:06 <AnMaster> don't think I allowed scripts on wp
20:53:35 <pikhq> No, wait, those are cachable. What *else* is it that's taking forever?
20:53:43 <Phantom_Hoover> pikhq: If you have an account you can change it back to the old Monobook skin.
20:54:12 <pikhq> Phantom_Hoover: Oh, so I have an excuse to log in.
20:54:13 -!- poiuy_qwert has joined.
20:54:22 * pikhq shall log in and purge that
20:54:36 * Sgeo wikistalks pikhq
20:55:20 <pikhq> It's like Wikipedia never considered that some people have high-latency Internet.
20:56:36 <Sgeo> pikhq, why do you use Satellite?
20:56:56 <pikhq> Sgeo: The alternatives are worse.
20:57:00 -!- kar8nga has joined.
20:57:15 * Sgeo would die with high-latency Internet
20:57:33 <pikhq> Would you prefer a modem?
20:57:35 <uorygl> 1translate Kesyyntymisen alkuaikoina koirasta on ollut hyötyä varoittavana vahtina sekä jätteensyöjänä.
20:57:41 <pikhq> Because that's the alternative.
20:57:47 <uorygl> `translate Kesyyntymisen alkuaikoina koirasta on ollut hyötyä varoittavana vahtina sekä jätteensyöjänä.
20:57:53 <HackEgo> The early days of domestication, dogs have been useful as well as a warning vahtina jätteensyöjänä.
20:58:16 -!- Oranjer has joined.
20:58:17 <Deewiant> jätteensyöjänä == as a garbage eater
20:58:40 <pikhq> AnMaster: 30k, normally.
20:58:48 <pikhq> Phantom_Hoover: Web designers are assholes.
20:58:50 <uorygl> pikhq: do you live in a rural place, then?
20:58:53 <pikhq> AnMaster: Too much line noise.
20:59:04 <AnMaster> pikhq, anyway, way faster than what I had in 1996 or so. which was ~28k
20:59:17 <pikhq> Yes. In '96 people didnt'
20:59:24 <pikhq> didn't make megabyte-sized pages.
20:59:46 <AnMaster> pikhq, megabyte? that is small ;P
20:59:57 <AnMaster> (though more would be annoying over internet)
21:00:06 <uorygl> Feel free to run an Ethernet cable to my house and leech.
21:00:31 <pikhq> Man, logging into Wikipedia *really* makes Wikipedia go faster.
21:00:59 <AnMaster> what does cf. stand for in English exactly? I know how it is used but not what the letters stand for
21:01:03 <uorygl> Hmm, now I want to go to a rural area and figure out a way to get good Internet.
21:01:04 <pikhq> Even Google's AJAX stuff isn't as bad. At least their Javascript gets *cached*.
21:01:23 <ais523> it means "compare", but I don't know what it stands for either
21:01:32 <AnMaster> Deewiant, is that a real word?
21:01:41 <uorygl> Looks like "confer" is a Latin word meaning "compare".
21:01:45 <uorygl> And also an English word meaning something else.
21:01:46 <Deewiant> In that meaning, it's archaic, though
21:01:51 <pikhq> uorygl: Murder the phone company execs who *pocketed* the money that was given to them in order to let them offer good Internet in rural areas.
21:01:56 <ais523> looking it up, it does stand for "confer", which is Latin for "compare"
21:02:03 <AnMaster> "Definitions of confer on the Web:
21:02:03 <AnMaster> * have a conference in order to talk something over; "We conferred about a plan of action"
21:02:03 <AnMaster> * present; "The university conferred a degree on its most famous former student, who never graduated"; "bestow an honor on someone"
21:02:03 <AnMaster> wordnetweb.princeton.edu/perl/webwn"
21:02:03 <ais523> rather than the English "confer" which means something quite different
21:02:07 <Deewiant> It means compare in English as well, just hasn't been used with that meaning for some hundreds of years
21:02:36 <uorygl> pikhq: well, how will that help? :P
21:02:41 <pikhq> WE GAVE THEM 200 FUCKING BILLION DOLLARS IN ORDER TO GET FIBER ALMOST EVERYWHERE IN THE 90S. AND DAMMIT I WANT MY 100 MEGABIT INTERNET.
21:02:42 -!- Phantom_Hoover has quit (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100423140709]).
21:02:59 <ais523> the English word "confer" tends to be mostly used in quiz shows, talking about whether contestants are allowed to talk to team-mates before answering
21:03:10 <pikhq> Deewiant: No, I'm the "few hundred feet too far to get DSL".
21:03:35 <Deewiant> As in, "almost everybody can get DSL".
21:03:47 <pikhq> Yeah, that's not what we paid for.
21:04:01 <pikhq> We paid for Japan's infrastructure and got this piece of crap.
21:04:14 <uorygl> Is there another person within a few hundred feet of you who has DSL who you could get some Internet from?
21:04:34 <pikhq> AnMaster: Equivalent to Japan's current infrastructure.
21:04:56 -!- benuphoenix has quit (Quit: leaving).
21:05:21 <pikhq> If you'll note, $200 billion is a ton of money.
21:05:43 <oerjan> `calc 200 billion USD / 1 ton
21:05:45 <HackEgo> (200 billion U.S. dollars) / (1 short ton) = 220 462 262 U.S. Dollars / kg
21:06:02 <uorygl> So, how do we know that the phone company guys simply pocketed the money?
21:06:15 <oerjan> `calc 1 ton / 200 billion USD
21:06:17 <HackEgo> (1 short ton) / (200 billion U.S. dollars) = 4.5359237 10^-9 kg / U.S. Dollar
21:06:35 <oerjan> pikhq: i _say_ you're overestimating a bit
21:07:08 <pikhq> uorygl: Because they say as much.
21:07:50 <pikhq> They have nearly zero competition here. And nearly zero incentive to improve anything.
21:08:03 <uorygl> pikhq: so where do they say that?
21:09:58 <oerjan> hm wait that's underestimating
21:10:21 <uorygl> AnMaster: there is no command called convert.
21:10:31 <AnMaster> uorygl, it would give an error
21:10:39 <HackEgo> /bin/bash: line 1: type: convert: not found
21:10:58 <uorygl> Yep, that's what it does.
21:11:48 <uorygl> That's a rather excellent pun.
21:12:04 <oerjan> alas it's also mixing metaphors
21:12:19 <uorygl> Not only is it a cross between "bit the dust" and "kicked the bucket", but it also refers to the "bit bucket", a place where information goes to die.
21:12:59 -!- poiuy_qwert has quit (Read error: Connection reset by peer).
21:14:25 <AnMaster> uorygl, but it didn't bucket the bits :(
21:14:34 -!- poiuy_qwert has joined.
21:19:10 <uorygl> So, we can all agree that C is practically Turing-complete. Is it still practically Turing-complete without malloc?
21:19:26 <pikhq> No, C is not Turing-complete.
21:19:56 <pikhq> No. It still has finite memory and is thus not Turing-complete.
21:20:14 <oerjan> well then just use a large global array :)
21:20:22 <pikhq> It's about as stupid as saying "the reals are practically computable".
21:20:31 <uorygl> oerjan: yeah, I guess that works.
21:20:35 <pikhq> But anyways. malloc can be written in pure C.
21:20:39 <pikhq> Just need a very large array.
21:21:04 <pikhq> You have a very large array.
21:21:12 <pikhq> You use this as the heap for your malloc implementation.
21:21:42 <uorygl> What if you didn't have arrays?
21:21:53 <uorygl> Or if they were limited in size.
21:22:04 <pikhq> *Then* you're reduced to using the stack for allocation.
21:22:56 <pikhq> (this is necessarily finite, but only about as much as the rest of memory in C. Most implementations limit it further.)
21:25:55 -!- coppro has joined.
21:30:39 <uorygl> Does having to use the stack for allocation really interfere with flow control?
21:31:11 <oerjan> well to allocate more memory you need to do a call that never returns...
21:31:39 <oerjan> which means essentially using continuation passing style afaict
21:31:50 <uorygl> Yeah. And how difficult is CPS in C?
21:32:02 <pikhq> Actually, not that hard.
21:32:26 <pikhq> Tedious, but not hard.
21:32:45 <pikhq> As you have function pointers, and can manually pass the environment around.
21:35:11 <uorygl> How do you use a function pointer?
21:36:22 <pikhq> Let's say you have a void (*foo)(void). To call foo, you just do: foo();
21:43:15 <AnMaster> from man pts(4): "Pseudo-terminals can also be used to send input to programs that normally refuse to read input from pipes (such as su(1), and passwd(1))."
21:43:24 <AnMaster> well, why, if it is so easy to work around
21:43:32 <AnMaster> I mean, what is the point of them refusing pipes
21:45:53 -!- augur has quit (Ping timeout: 246 seconds).
21:45:59 <AnMaster> another wtf: "The behavior of grantpt() is unspecified if a signal handler is installed to catch SIGCHLD signals."
21:46:54 <uorygl> Well, for su(1), it's clear enough.
21:47:19 <pikhq> Probably to make it impossible for the password to be echoed?
21:48:06 -!- augur has joined.
21:48:12 <uorygl> If I do foo | su -c bar | baz, I want foo's output to go to bar, not be interpreted as a password.
21:48:23 <Sgeo> AnMaster, that sounds useful for Normish.. if Normish weren't dying
21:48:34 <ais523> presumably it's for avoiding stupid mistakes, rather than deliberate attempts to manipulate it
21:48:47 <Sgeo> The pseudo-terminal to send input to passwd
21:48:55 <Sgeo> So we could have a working addplayer
21:49:13 <ais523> doesn't adduser have an option to set the password at random and send it somewhere?
21:49:14 <Sgeo> uorygl, please save Normish kthx
21:49:19 <ais523> or am I thinking of the GUI version?
21:49:38 <AnMaster> ais523, some adduser/useradd have that
21:50:02 <AnMaster> it is funny how I have both adduser and useradd on my system
21:50:53 <uorygl> adduser is the friendly one, useradd is the unfriendly one.
21:51:00 <uorygl> adduser is the P one, useradd is the V one. :P
21:51:32 <AnMaster> well it is a bash script on arch and a perl script on ubuntu
21:53:03 <uorygl> P refers to things that are designed to be easy to use. V refers to things that are designed to be easy to implement.
21:53:22 <uorygl> Examples of P programming languages are Python, Perl, and PHP. Examples of V programming languages are Verilog and VHDL.
21:53:47 <AnMaster> uorygl, what about C++? It is neither
21:53:54 -!- augur has quit (Ping timeout: 246 seconds).
21:54:23 <AnMaster> uorygl, also I doubt implementing VHDL or Verilog synthesis is easy at all
21:54:36 <uorygl> I know very little about C++.
21:54:57 <AnMaster> uorygl, well I know only a tiny bit and it is more than I want to know
21:55:17 <uorygl> Verilog is like VHDL, except that the language itself is designed well, rather than simply tolerating good design. :P
21:55:28 <pikhq> uorygl: Take C++. Add poorly-implemented features.
21:55:41 <ais523> pikhq: no, that's SystemC
21:55:45 <pikhq> You then go into a loop with that body.
21:55:59 <ais523> which is orders of magnitude stupider than either Verilog or VHDL
21:56:25 <pikhq> uorygl: C++ has a Turing-complete type system and the C preprocessor for metaprogramming options.
21:57:15 <ais523> one thing that annoys me about VHDL is it doesn't allow recursion in the preprocessor
21:57:19 <AnMaster> pikhq, hm? do you mean the templates?
21:57:25 <AnMaster> I guess they are part of the type system
21:57:26 <ais523> at least, the syntax allows it, but every simulator I know crashes when you try
21:57:27 <pikhq> Because of this type system, C++ parsing is undecidable.
21:57:54 <uorygl> It's easy to hate C++ if you want to!
21:57:56 <ais523> however, it /is/ expressive enough that you can simulate bounded recursion by using a bunch of polynomials in the preprocessor, which is arguably worse
21:58:00 <ais523> uorygl: it's easy to hate it anyway
21:58:05 <AnMaster> pikhq, but the templates act all preprocessorish, as in, things you would expect worked if it was in compiler instead of preprocessor doesn't work
21:58:10 <ais523> I don't hate C++ because of an undecidable type system
21:58:14 <ais523> but there are other reasons to dislike it
21:58:16 <AnMaster> pikhq, like implementation outside header
21:58:38 <Sgeo> Wait, C++ types are Turing-complete?
21:58:41 <Sgeo> That sounds like a feature
21:58:46 <AnMaster> ais523, does VHDL have a preprocessor?
21:58:50 <pikhq> Sgeo: *Unintentionally so*.
21:59:00 <ais523> AnMaster: well, depending on your opinion, pretty much the entire language is a preprocessor
21:59:01 <Sgeo> Example, please/
21:59:04 <ais523> but I'm thinking of for-generate loops
21:59:14 <ais523> which I consider a preprocessor, based on what they actually do
21:59:17 <pikhq> Sgeo: template <typename T> struct foo { ... };
21:59:32 <pikhq> This is equivalent to the following Haskell statement: foo t = ...
21:59:33 <ais523> basically, for loops, except they're semantically unrolled rather than semantically being repetition
22:00:04 <AnMaster> ais523, what would you use recursion for there? I haven't used it but I imagine it would be useful to make stuff like n-bit adders from full adders or such
22:00:15 <AnMaster> combined with generics or such
22:00:26 <Sgeo> pikhq, how do you do Factorials in C++ type system?
22:00:36 <ais523> O(log n) time, O(n) space
22:01:17 <pikhq> template <int N> struct Factorial {enum { value = N * Factorial<N - 1>::value };};template <>struct Factorial<0>{enum { value = 0};};template<>struct Factorial<1>{enum {value = 1};};
22:01:18 <AnMaster> ais523, couldn't you make a perfect sorting network? I mean, just write a huge Karnaugh diagram for all possible inputs or something XD
22:01:33 <ais523> oh yes, you can get O(1) time for sorting in hardware if you really really want to
22:01:39 <pikhq> Yes, the pattern matching works the opposite of what you'd expect.
22:01:40 <ais523> but the space term is so horrible I don't want to imagine it
22:01:56 <ais523> (and of course, it only works for an input of given size)
22:02:00 <AnMaster> ais523, for something like 4 inputs it couldn't be too bad?
22:02:20 <uorygl> There, that should grow fast enough.
22:02:36 <ais523> AnMaster: depends on how wide the inputs are
22:02:38 <AnMaster> uorygl, size of that O(1) sort?
22:02:42 <pikhq> C++ metaprogramming should, IMO, be commented with the equivalent Haskell.
22:02:47 <ais523> the other thing to mention is that there's no such thing as a hardware comparison sort
22:02:54 <ais523> because you have to know the bounds of the data type to be able to implement them
22:02:56 <AnMaster> ais523, well I meant 4 input bits split into equal sizes
22:02:59 <uorygl> AnMaster: an upper bound thereon.
22:03:05 <ais523> AnMaster: yes, that sorts rather trivially
22:03:16 <pikhq> In this case: factorial 0 = 0;factorial 1 = 1;factorial n = n * factorial (n-1)
22:03:21 <AnMaster> uorygl, that notation doesn't make sense to me
22:03:23 <uorygl> (!)^n means "factorial n times".
22:03:44 <uorygl> So for 1, it's 1!; for 2, it's 2!! (to slightly abuse notation); for 3, it's 3!!!, and so on.
22:04:02 <AnMaster> ais523, what are the gates you use in a sorting network btw, you need something clocked I assume?
22:04:15 <ais523> yep, it was done synchronously
22:04:26 <AnMaster> well not need as such, presumably you could do it async
22:04:30 <ais523> and built out of serial 2-input sorters
22:04:52 <AnMaster> ais523, hm? I don't think I'm familiar with that
22:04:58 <Sgeo> Is Verilog FOSS?
22:05:01 <ais523> well, they aren't a standard component
22:05:05 <ais523> Sgeo: it's a programming language
22:05:10 <ais523> you might as well say "is C FOSS"?
22:05:25 <ais523> as for whether there are FOSS implementations, I'm not sure
22:05:38 <AnMaster> verilog have more FOSS implementations than VHDL
22:05:40 <ais523> but it seems likely that there are no completely-FOSS Verilog/VHDL synthesizers
22:05:46 -!- oerjan has quit (Quit: Good night).
22:06:15 <uorygl> Is there something that compiles Verilog into a circuit diagram? :P
22:06:29 <uorygl> A circuit diagram consisting only of transistors!
22:06:37 <AnMaster> electric (EDA for ICs) supports some limited VHDL for generating substrate layout (or whatever it was called)
22:06:44 <ais523> uorygl: the expensive synthesizers can synthesize for FPGA, then translate back into a circuit diagram
22:06:51 <ais523> but I'm not entirely sure why you'd want to do that
22:06:52 <AnMaster> not for generating schematics iirc
22:07:14 <uorygl> Hmm, we should compile Verilog into BBM instead.
22:07:20 <AnMaster> ais523, for making a pretty pic to show those who pay you?
22:07:42 <uorygl> The problem, of course, is that BBM is reversible.
22:07:51 <uorygl> The Billiard Ball Machine.
22:08:09 -!- ws has joined.
22:08:11 <uorygl> Computing via steel balls hitting blocks and each other.
22:08:29 <AnMaster> I wonder if you could design reversible circuits in VHDL
22:08:53 <uorygl> Sure you can. Just make your circuit reversible, and it'll be reversible.
22:09:02 <AnMaster> btw did I mention what I planned to do this summer?
22:09:13 <AnMaster> write a befunge-93 interpreter in vhdl (or verilog)
22:09:24 <AnMaster> (but since I took a VHDL course at uni it will probably be that)
22:09:31 -!- poiuy_qwert has quit (Ping timeout: 240 seconds).
22:09:39 <AnMaster> since I don't have any FPGA it will likely be ghdl-tested only
22:10:04 <uorygl> Build your circuit out of Toffoli gates.
22:10:32 <AnMaster> uorygl, exactly, can you do that in VHDL, after all it is rather high level often. And I doubt many FPGAs have such gates
22:10:52 <uorygl> Sure, a Toffoli gate can easily be described in VHDL.
22:11:31 -!- mre has joined.
22:11:50 <AnMaster> uorygl, well okay, but will it be synthesised into one. Rather than a combination of other gates
22:12:30 <uorygl> I think FPGAs never synthesise any gates at all into themselves.
22:12:34 -!- maedhros777 has joined.
22:13:09 <uorygl> Instead, they use lookup tables.
22:13:17 <maedhros777> hiato: Sorry to bother you, but I was wondering how your method of increasing primes would work with complex zeroes.
22:14:45 -!- poiuy_qwert has joined.
22:14:49 <uorygl> Yeah. The Wikipedia article on FPGAs gives a pretend FPGA architecture; it's like a typical FPGA but simpler.
22:14:59 <hiato> maedhros777: Why would there be complex zeroes?
22:15:39 <AnMaster> uorygl, also what about latches and such
22:15:40 <uorygl> Each cell has four inputs, which go into a lookup table, producing one output, which may go through a flip-flop.
22:15:53 <uorygl> So, you get one lookup table and one flip-flip.
22:15:57 <AnMaster> uorygl, what about a latch then? rather than a flip-flop
22:16:16 <uorygl> Depends on what the difference between a latch and a flip-flop is.
22:16:26 <AnMaster> uorygl, latch would be async afaik
22:16:53 <uorygl> I guess you could construct that easily. Feed its output back into it as an input, and also have the two other inputs.
22:17:28 <uorygl> Not necessarily, is it?
22:17:29 <AnMaster> you can make a flip-flop out of a latch (iirc that is the normal way of doing it) but not the other way around
22:17:40 <AnMaster> you could built it out of gates though
22:17:52 <uorygl> Right, that's what I'm saying to do.
22:17:56 <AnMaster> two nor with suitable feedback...
22:18:02 <maedhros777> hiato, any idea how to encode with complex zeroes?
22:18:43 <uorygl> You can do that with a single LUT.
22:19:31 <uorygl> You can make that SR-latch with a single LUT.
22:19:41 <uorygl> If it has one bit of state which is also its one bit of output.
22:19:50 <AnMaster> uorygl, well probably if you only have one of the outputs
22:20:07 <AnMaster> you need 3 inputs, set/reset/feedback
22:20:25 <AnMaster> the classical SR-latch have both "normal" and "inverted" outputs though
22:21:13 <uorygl> Then you'll have to use a second LUT to provide that.
22:21:19 <AnMaster> uorygl, well, because it is symmetric which one is which is arbitrary as long as you swap the inputs too
22:21:32 <maedhros777> Ilari, would you have any ideas for complex zeroes?
22:21:52 <uorygl> An FPGA doesn't let you use "just an inverter"; all you get is LUTs and flip-flops, more or less.
22:22:09 <uorygl> Anyway, how's this for a Toffoli gate? http://pastebin.com/HFZNVdc1
22:22:55 <AnMaster> I don't know verilog, only VHDL (and only basic VHDL so far)
22:23:27 <uorygl> Well, it can't be very much different from VHDL.
22:23:43 <AnMaster> well true, somewhat different syntax
22:23:52 <ais523> they're basically the same lang with different syntax
22:24:04 <ais523> they were originally rather different, but they each stole features from the other until they became basically identical
22:24:21 <ais523> and they're still doing it; if something becomes standard in Verilog, it tends to go into the next version of the VHDL standard, and vice versa
22:24:30 <maedhros777> Does anyone here know how complex zeroes can be encoded in a function so that their order is distinct?
22:24:52 <uorygl> maedhros777: what are we trying to do, here?
22:24:58 <maedhros777> http://esoteric.voxelperfect.net/wiki/Polynomial
22:25:33 <maedhros777> Reals are encoded by increasing primes, but i'm not sure what to do about complex zeroes
22:25:39 -!- Phantom_Hoover has joined.
22:26:06 <soupdragon> A Polynomial program is of the form f(x) = 3x^2 + x + 7, where the polynomial begins with "f(x) = "
22:26:25 <ais523> reminds me a bit of http://esolangs.org/wiki/Formula
22:26:50 <ais523> different, but related
22:26:52 -!- Oranjer has left (?).
22:26:59 <ais523> I was wondering about just editing it into the see also of your article
22:27:13 <AnMaster> uorygl, I think this would work for vhdl http://sprunge.us/YMcO
22:27:36 <uorygl> AnMaster: looks right.
22:27:51 <uorygl> maedhros777: I'm looking.
22:27:57 <AnMaster> uorygl, that verilog looks like a process what with the "always" thingy?
22:28:00 <ais523> hmm, that article needs categories while it's at it
22:28:08 <ais523> maedhros777: what year did the language first become public? this one?
22:28:20 <AnMaster> uorygl, oh and can't you have multiple architectures in verilog?
22:28:20 <uorygl> AnMaster: I guess Verilog doesn't let you constantly execute something; "always" means "whenever", and "*" means "something changes".
22:28:46 <AnMaster> uorygl, well my VHDL was async I think?
22:29:07 <uorygl> AnMaster: I guess an "architecture" is what you call what Verilog calls modules.
22:29:23 <Ilari> maedhros777: Maybe use newton to try to find one (start from complex number), then when found, divide it out and do the same again until you get constant function...
22:29:24 <AnMaster> uorygl, you can have multiple implementations of the same entity
22:29:34 <AnMaster> uorygl, that is what the architecture means
22:29:42 <AnMaster> uorygl, the entity is more like a module.
22:29:50 <uorygl> AnMaster: oh, how does that work?
22:29:58 <Phantom_Hoover> maedhros: There's a crapload of stuff on ordering multiple numbers in the logs from a few weeks ago.
22:30:15 <AnMaster> uorygl, the entity defines the "outside" appearance. The architecture what is done
22:30:34 <ais523> maedhros777: cats added, tweak at will
22:30:38 <AnMaster> uorygl, iirc you can select which architecture to use, it defaults to one of them (forgot which, ask ais)
22:30:43 <maedhros777> Ilari: What would I be trying to find from the complex number?
22:30:44 <Ilari> maedhros777: If all coefficients are real, you can divide complex roots in pairs and keep the function as real.
22:31:10 <AnMaster> uorygl, iirc part of the reason may be one is only possible to simulate and not to synthesise or such.
22:31:18 <AnMaster> at least that is one reason I heard
22:31:27 <AnMaster> ais mentioned some other ones iirc
22:31:46 <ais523> hmm, first time I've ever had an edit conflict on Esolang
22:31:49 <ais523> cats actually added this time
22:32:02 <AnMaster> uorygl, well, if you use console IO it probably makes a good test bench but not so good FPGA ;P
22:32:30 <uorygl> maedhros777: I find the "we might be tempted to deduce" slightly offensive. :P
22:32:47 <AnMaster> uorygl, but yeah, ask ais523 about why you want multiple architectures
22:33:09 <ais523> uorygl: basically, because of the hugely closed nature of most FPGA ecosystems
22:33:22 <uorygl> "I have a cat in this box. You might be tempted to deduce that this box contains a cat. But actually, by 'in', I meant 'on top of'!"
22:33:25 <ais523> "better" architectures are more expensive than worse ones
22:33:43 <ais523> uorygl: heh, "in" and "on top of" are the same word in Latin
22:34:09 <uorygl> Hmm, and the Spanish word "en" means both as well.
22:34:18 <Ilari> maedhros777: Like x^3 - x^2 + x - 1 = 0. x = i is solution, so you can divide by (x - i)(x + i) = x^2 + 1 to get x - 1.
22:34:26 <AnMaster> ais523, it isn't hard to imagine situations were it would be unclear which is meant and doing it wrong would be rather irritating
22:34:52 <maedhros777> Ilari: Why would you divide, though? Because some language statements involve complex zeroes
22:35:07 <ais523> AnMaster: there's a reason why the ancient Romans normally used Greek for discussions about science
22:35:26 <uorygl> maedhros777: you can still divide complex zeros out.
22:35:33 <Ilari> maedhros777: Ensures that root finding won't find that root again.
22:35:44 <uorygl> Unless, of course, it was a double root.
22:35:55 <Ilari> maedhros777: And you also see if there are higher-order roots.
22:36:06 <maedhros777> But how will you know which order the complex zeroes are in?
22:36:13 <AnMaster> <uorygl> Hmm, and the Spanish word "en" means both as well. <-- insert obvious joke with regards to Spanish economy
22:36:16 <Ilari> maedhros777: You can't.
22:36:28 <AnMaster> (at least i hope it is obvious)
22:36:44 <uorygl> I like the Newton's method idea. Start at 0, see which root you find; that's the first one you execute.
22:36:51 <AnMaster> (the placed the money on top of the bank instead in it, see not strange they have financial problems!)
22:37:29 <maedhros777> But that limits programs from executing in different order
22:37:44 <uorygl> maedhros777: I notice that all the language does is start with a number and perform a series of operations on it.
22:37:50 <AnMaster> maedhros777, do the roots have some well defined order at all?
22:38:06 <uorygl> Well, yeah. Esoteric programming languages limit you. :)
22:38:09 <uorygl> AnMaster: no, they don't.
22:38:54 <AnMaster> and you could order any roots anywhere, just do something like order numerically by real, then imaginary. Or write it down and use the ascii value
22:39:54 <AnMaster> maedhros777, "3+4.2i" == 51,43,52,46,50,105
22:40:07 <AnMaster> at least that is what converting it in erlang told me
22:40:36 -!- kar8nga has quit (Remote host closed the connection).
22:40:40 <AnMaster> if you limit yourself to ASCII you can get a number from it that is 6 bytes
22:40:48 <AnMaster> maedhros777, well you can sort by that
22:40:56 <AnMaster> to get a well defined ordering of the numbers
22:41:14 <uorygl> AnMaster: it's not really well-defined.
22:41:21 <uorygl> Suppose the real part is nonterminating.
22:41:39 <uorygl> Well, I guess it's obvious enough how to fix that.
22:41:41 <AnMaster> uorygl, bah, double isn't infinite precision ;P
22:41:58 <maedhros777> All complex numbers are Gaussian in Polynomial
22:42:21 <AnMaster> but sure, sorting numerically by real then imaginary is probably saner
22:42:32 <uorygl> What if you have a number with multiple decimal expansions? :P
22:42:38 <maedhros777> Limiting...but I may have no other choice.
22:43:00 <AnMaster> (it is fun combining math with engineering)
22:43:04 <uorygl> Well-defined and undefined are mutually exclusive concepts. :)
22:43:07 <AnMaster> (neither profession will like it)
22:43:24 <AnMaster> maedhros777, what is it you are trying to accomplish though
22:43:56 <AnMaster> maedhros777, so what is the language in question
22:43:57 <maedhros777> Something similar to the ascending primes idea
22:44:05 <maedhros777> http://esoteric.voxelperfect.net/wiki/Polynomial
22:44:08 <AnMaster> maedhros777, also just add a line number then. Like BASIC
22:44:08 <ais523> maybe you could sort anti-clockwise?
22:44:16 <ais523> then by distance from the origin?
22:44:28 <soupdragon> maedhros777: why not just write in a list of numbers? why go via polynomials
22:44:49 <AnMaster> soupdragon, this is #esoteric, that's why
22:44:54 <maedhros777> soupdragon: Because polynomials are cool :)
22:44:55 -!- MigoMipo has quit (Read error: Connection reset by peer).
22:45:32 <maedhros777> ais523: interesting idea, maybe I'll do that
22:46:04 <maedhros777> Because if I did it by order of a in a + bi, everything might be screwed up
22:46:10 <AnMaster> maedhros777, why not in the order written in the program?
22:46:24 <uorygl> Here's my esoteric programming language: http://pastebin.com/cmLyEHJM
22:47:22 <AnMaster> uorygl, Newtons method is numerical?
22:47:31 <AnMaster> I think there may be more than one Newton's method
22:47:37 <uorygl> Newton's method is numerical, yes, but too bad.
22:47:46 <uorygl> I recall only one Newton's method.
22:47:46 <AnMaster> uorygl, how many iterations then?
22:48:03 <AnMaster> uorygl, the one I remember was related to differentiation?
22:48:16 <uorygl> Enough that you're absolutely certain of which root it's approaching.
22:48:24 <uorygl> Newton's method does involve differentiation, yes.
22:54:15 <maedhros777> Maybe I could use the ascending primes method with b in a + bi
22:54:53 <maedhros777> b can't be too large because there are certain set possible values of b, so that might work
22:55:15 <maedhros777> Only problem would be with i and 2i (output and input)
22:56:55 <maedhros777> Any ideas for i and 2i (the special cases)?
22:57:15 <AnMaster> maedhros777, can't b be any integer?
22:57:20 <uorygl> maedhros777: sorry, I don't know what you're trying to get across.
22:57:32 <uorygl> How does the ascending primes method even work with a + bi?
22:58:46 <uorygl> So, i becomes 2^1 * i and 2i becomes 2^2 * i?
22:58:52 <uorygl> I don't see a problem there.
22:59:32 <maedhros777> I'm off to change the wiki page for Polynomial now
22:59:35 -!- maedhros777 has quit (Remote host closed the connection).
22:59:38 <uorygl> I must admit I prefer my language, though. :P
23:00:09 <soupdragon> the input is two numbers, the output is a binary sequence
23:00:20 <soupdragon> based on which paths are taking in the GCD algorithm
23:00:50 <soupdragon> the binary sequence is used to train (good/bad) an AI what to do
23:01:28 <soupdragon> but the AI is only allowed to communicate to the outside world using pictures downloaded from the internet
23:01:39 <soupdragon> (so that it can't trick you to let it out)
23:01:54 <ais523> soupdragon: I don't see why that restriction would matter much
23:02:08 <uorygl> Couldn't it just download pictures of letters?
23:02:26 <uorygl> How do you disallow it?
23:02:46 <uorygl> Yes, but how do you make it heed the difference?
23:03:08 <uorygl> If you can make it heed the difference between a letter and a non-letter, you might as well make it heed the difference between right and wrong.
23:03:24 <AnMaster> uorygl, you use a monitor AI to do it XD
23:03:43 <uorygl> That could get exciting. :)
23:03:54 <AnMaster> uorygl, an infinite series of AIs
23:05:08 <Sgeo> Well, the monitor AI wouldn't really be able to communicate with the monitored.. no, it could
23:05:17 <Sgeo> If the monitored AI receives "accepted/rejected"
23:05:27 <uorygl> The monitored AI wouldn't have to receive those.
23:05:54 <uorygl> It could just spit out lots of letters, blissfully unaware that they're being rejected. :)
23:05:59 <soupdragon> the binary sequence is LEFT/RIGHT commands for a turtle
23:06:20 <AnMaster> soupdragon, oh so now it can drive around in letter shapes
23:06:45 <AnMaster> and also being stabbed by a turtle pen
23:07:01 <AnMaster> look those robots can be vicious.
23:07:02 <uorygl> Programming languages should not have output, because communication can be used to take over the world. :)
23:09:59 <pikhq> uorygl: There should be a single output for all computers.
23:10:06 <pikhq> A single flashing LED, in NORAD.
23:10:37 * Sgeo doesn't know if he should be excited or upset about tomorrow
23:10:59 <uorygl> You should probably be excited, since it's probably not a school-required thing.
23:11:04 <Sgeo> Exciting: Major marriage in a webcomic (although there's now a cold feet issue). Upsetting: SG-1 will no longer be available
23:12:21 <pikhq> AnMaster: Air Force base. Underneath a mountain.
23:12:28 <Sgeo> Yeah, this is what my life is like. The most interesting.. dammit, Internet connection
23:12:28 -!- Sgeo has quit (Quit: Ex-Chat).
23:12:30 <pikhq> Designed to survive a nuclear blast.
23:12:33 -!- Sgeo has joined.
23:12:38 <Sgeo> Yeah, this is what my life is like. The most interesting.. dammit, Internet connection
23:12:41 <AnMaster> pikhq, a bit hard to take off what with the mountain on top?
23:13:12 <AnMaster> <Sgeo> Yeah, this is what my life is like. The most interesting.. dammit, Internet connection
23:13:28 <Sgeo> I was lagging out
23:13:33 <AnMaster> Sgeo, "* Sgeo has quit (Quit: Ex-Chat)"
23:13:36 <Sgeo> Didn't think it would be received
23:13:48 <uorygl> The thing about computers is that it's difficult to make them not output via heat and stuff.
23:14:01 <AnMaster> Sgeo, could be temporary lag spike
23:14:05 <Sgeo> Um, I disconnected and reconnected my router
23:14:08 <AnMaster> Sgeo, you have to wait a bit to be sure
23:14:09 <pikhq> uorygl: We'll magine.
23:14:10 <Sgeo> And did a /server
23:14:41 * Sgeo does not fake connection issues, really!
23:14:51 <Sgeo> AnMaster, hulu
23:15:13 <pikhq> Probably something like destroy them if their temperature fluctuates greater than some epsilon.
23:15:31 -!- augur has joined.
23:15:34 <uorygl> So they'll be able to determine whether they're destroyed or not.
23:15:45 <uorygl> I once devised a scheme for ensuring that information doesn't escape something.
23:15:58 <Sgeo> AnMaster, you mean, illegally?
23:16:00 <pikhq> uorygl: Yes, so they can send at most one bit. Ever.
23:16:11 <uorygl> On the contrary, they can do it at a symbolic time.
23:16:16 <AnMaster> Sgeo, well that is up to you, just tcpdump everything, I guess that isn't legal either then
23:16:36 <Sgeo> I don't think I have time
23:16:36 <pikhq> Yes, but it's still exactly one bit of output.
23:16:40 <Sgeo> Nor space on this computer
23:16:48 <AnMaster> pikhq, how would you ensure that?
23:16:50 <uorygl> If you can choose when it's output, that's not exactly one bit of output.
23:17:10 <AnMaster> pikhq, couldn't it change the state later to construct a serial stream?
23:17:16 <AnMaster> or would it have a one-off manual reset?
23:17:35 <pikhq> AnMaster: *Destroyed*.
23:17:42 <uorygl> Anyway. First, put the thing inside a case it cannot escape from. Then, submerge it in a tank filled with hot saltwater, with metal filings to keep it from floating to the top or sinking to the bottom.
23:18:26 <pikhq> uorygl: Okay, fine. We use a time machine. We erase the computer from existence if its temperature fluctuates.
23:18:42 <Phantom_Hoover> Is it practical to get effectively free data storage by sticking lots of cheap, low-ish capacity (1-2G) flash drives together in a RAID?
23:18:52 <AnMaster> pikhq, but then you will probably kill yourself before you could go back and do so!
23:19:07 <uorygl> Have a pipe at the bottom that sucks out the filings, making them into an inverted cone shape, so that the thing doesn't float to the sides. Have the filings detonated alongside a nuclear bomb.
23:19:16 <uorygl> Phantom_Hoover: how would that be effectively free?
23:19:35 <AnMaster> yeah what uorygl said, plus you need card read, lots of them
23:19:35 <uorygl> Flash drives cost money.
23:19:41 <Phantom_Hoover> I have *tonnes* of flash drives lying around my house.
23:19:53 <AnMaster> uorygl, they are handed out for free at conventions and such iirc
23:20:13 <AnMaster> I have some old I got from people who got them at conferences and such and didn't want them
23:20:32 <Phantom_Hoover> USB ports may be an issue, but they are very cheap and I have a few already.
23:20:43 <AnMaster> uorygl, well 256 MB and 1 GB are the ones I can find atm
23:21:07 <Phantom_Hoover> (I have them because my parents are doctors and drug companies are allowed to bribe them)
23:21:36 <AnMaster> well, here they aren't allowed to bribe I think
23:21:41 <uorygl> I should become a doctor so that I get free flash drives. :P
23:22:35 <AnMaster> heck some of the ones I have are also medical related (guy who I got them from works in neuroscience or such... academic, not doctor)
23:22:45 <AnMaster> one says something about ADHD research on it
23:24:27 <AnMaster> ais523, how do you create local variables in VHDL?
23:24:39 <ais523> I can't remember, but it's possible
23:24:41 <uorygl> Let's see, you can get a terabyte SSD for about $3,000...
23:24:43 <ais523> that's not really very helpful...
23:24:50 <ais523> (and both local variables and local signals are possible)
23:24:52 <AnMaster> ais523, what? internal signals?
23:24:58 <AnMaster> ais523, I know about local signals
23:25:14 <AnMaster> ais523, how exactly do vars and signals differ
23:25:18 <Sgeo> Since when is 1GB small? I remember when I was younger, learning about Gigabytes and thinking that that's a LOT
23:25:37 <uorygl> You can also get a plain old regular terabyte hard drive for $100. :P
23:26:03 -!- augur has quit (Ping timeout: 240 seconds).
23:27:13 <uorygl> That's stupid. You can buy an encrypted 4 TB hard drive for about $1,500.
23:27:24 <Sgeo> How difficult would it to put a 1TB drive in a laptop?
23:27:39 <uorygl> Whereas doing the same with software would be just as effective and way cheaper.
23:28:56 <uorygl> http://www.newegg.com/Product/Product.aspx?Item=N82E16822158093
23:29:09 -!- augur has joined.
23:29:23 <AnMaster> uorygl, but where do you enter passphrase?
23:29:34 <uorygl> What's the leading symmetric key cipher? Twofish?
23:30:20 <AnMaster> uorygl, don't know. AES isn't too bad
23:30:37 <uorygl> I think symmetric key ciphers are pretty easy to implement in hardware. I don't know, though.
23:31:10 <uorygl> Of course, using an ordinary processor instead of a cryptoprocessor is not necessarily a good idea.
23:32:13 <AnMaster> uorygl, a crypto cpu does add some interesting aspects to the design...
23:32:32 <AnMaster> you couldn't make it async since that would mean widely varying power usage
23:32:45 <AnMaster> still, widely varying for sync probably
23:34:35 <Phantom_Hoover> If someone can analyse your CPU's power usage, you're doing something wrong in the first place.
23:37:16 <ais523> Sgeo: side-channel attacks are where you attack the implementation of an algo rather than the algo itself
23:37:21 <AnMaster> Phantom_Hoover, on the other hand you don't want to use more power than required at any point, especially in a laptop
23:37:29 <ais523> like shining a really bright light into a quantum crypto link to try to make it leak extra photons
23:38:02 <ais523> on some versions of it, yes
23:38:12 <ais523> the point is it has nothing to do with the algo at all, it's something they didn't think of
23:38:23 <AnMaster> ais523, yes but how would it work?
23:38:40 <AnMaster> shouldn't it just drown out the single photon?
23:38:49 <ais523> well, the things use filters to produce photons in the first place, you're trying to guess their settings
23:38:55 <ais523> so instead of one photon, you get, say, 20 all the same
23:39:02 <ais523> then you can measure 19 of them and let the other one go through the link
23:39:09 <ais523> and have a pretty good idea what polarisation that 19 had
23:39:21 <AnMaster> ais523, you need to modify the sender equipment then
23:39:33 <ais523> to stop that happening?
23:39:39 -!- augur has quit (Ping timeout: 240 seconds).
23:39:53 <AnMaster> ais523, to be able to send more than one such photon?
23:39:56 <ais523> except you don't, you cut the cable and aim a blast of light down it back at the sender
23:40:04 <ais523> and some of the photons bounce off the sender and come back
23:40:11 <ais523> going through the filter in the process
23:40:18 <AnMaster> ais523, how do you prevent it?
23:40:25 -!- Phantom_Hoover has quit (Remote host closed the connection).
23:40:45 <ais523> I have no idea, I'm not a quantum cryptographer
23:40:54 <ais523> one obvious method is to detect the blast and guess you've been tampered with
23:41:39 <AnMaster> ais523, but would that be viable? Quantum + detecting/measuring tends to non-weird results ;P
23:41:41 <ais523> that's not the only side-channel attack, of course
23:41:53 <ais523> AnMaster: you aren't detecting anything you send at all
23:41:55 <ais523> just things you receive
23:42:04 <ais523> you can therefore use conventional means to help secure the quantum system
23:42:42 <AnMaster> ais523, what are the other side channel attacks apart from compromising the sending/receiving computer(s)
23:43:38 <ais523> there's also the possibility of MitMing the entire communication
23:43:56 <ais523> as in, setting up a separate quantum link to each of the participating people
23:44:08 <ais523> but there are known ways to solve that sort of problem (again, I don't know what they are)
23:44:31 -!- tombom__ has quit (Quit: Leaving).
23:45:28 <uorygl> Calculate the amount of time taken for light to go from point A to point B. If it takes longer, there's tampering. :P
23:45:43 <uorygl> Of course, it requires your connection to be a straight line!
23:45:57 <ais523> strangely, at the conference I was at recently, there was an entire paper about just that method of securing communication
23:46:01 <AnMaster> uorygl, or not, you can calculate for a bent path, no?
23:46:44 <uorygl> Only if the attacker can't unbend the path.
23:47:01 <AnMaster> uorygl, well okay, but I was thinking "transatlantic cable" kind of thing
23:47:13 <AnMaster> he can shorten the path as needed ;P
23:47:19 <uorygl> I don't think wormholes are something we need to worry about for the time being. :P
23:49:10 <AnMaster> uorygl, also time machine to send the bits back in time