←2008-02-26 2008-02-27 2008-02-28→ ↑2008 ↑all
00:03:34 -!- CakeProphet has joined.
00:05:02 <slereah__> Are Schönhage's machines easier to understand than Kolmogorov's?
00:05:21 <slereah__> Even a clearer paper might do!
00:15:16 <slereah__> Problem is, I can't find anywhere in his article an example that doesn't involve the graph itself!
00:15:40 <slereah__> The only example is "How to double the graph".
00:16:06 <slereah__> Not a whole lot helpful for a more useful algorithm.
00:48:51 -!- RainbowTrout has joined.
00:49:27 -!- sebbu has quit ("@+").
00:50:40 <RainbowTrout> Cake Prophet
00:51:08 <RainbowTrout> CAKE PROPHET
00:51:19 -!- RainbowTrout has quit (Client Quit).
00:51:35 <CakeProphet> ...
00:52:07 <slereah__> Impatient lad!
01:22:14 -!- CakeProphet has quit ("haaaaaaaaaa").
01:27:14 -!- jix has quit ("CommandQ").
02:09:39 -!- atsampson has quit (Read error: 113 (No route to host)).
02:14:45 -!- atsampson has joined.
03:04:45 -!- sekhmet has quit (Read error: 110 (Connection timed out)).
03:19:30 -!- cmeme has quit ("Client terminated by server").
03:19:41 -!- cmeme has joined.
03:32:45 -!- immibis has joined.
03:51:15 -!- sekhmet has joined.
03:52:47 -!- uvanta has quit ("If you're seeing this message, it's probably XChat's fault.").
04:03:58 -!- immibis has quit (Client Quit).
07:20:10 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:14:55 -!- jix has joined.
08:17:40 -!- FurFur has joined.
08:18:45 -!- FurFur has left (?).
09:47:34 -!- olsner has quit ("Leaving").
10:58:22 -!- Corun has joined.
11:56:31 -!- Corun has quit ("This computer has gone to sleep").
13:48:22 -!- timotiis has joined.
13:52:31 -!- slereah__ has quit (Read error: 104 (Connection reset by peer)).
13:53:03 -!- slereah__ has joined.
14:03:59 -!- RedDak has joined.
14:04:38 -!- slereah_ has joined.
14:33:34 -!- slereah__ has quit (Read error: 110 (Connection timed out)).
14:47:40 -!- RedDak has quit (Remote closed the connection).
15:58:56 -!- oerjan has joined.
16:55:15 -!- n0nsense has joined.
16:55:24 -!- n0nsense has quit (Client Quit).
17:15:26 -!- ais523 has joined.
17:56:56 -!- Sgeo has joined.
18:18:26 -!- sebbu has joined.
18:20:39 -!- Corun has joined.
18:36:21 -!- RedDak has joined.
18:43:24 -!- Corun has quit ("This computer has gone to sleep").
19:19:03 <ais523> why isn't one's complement used more often?
19:19:20 <ais523> it's easy to create a two's complement system by just using unsigned adders and disregarding the carry
19:19:40 <ais523> but looping the carry from the msb round to the carry input of the lsb gives you ones' complement
19:20:01 <ais523> (albeit with -0 being the standard representation of 0 and +0 being impossible to get unless you start a number at that value)
19:20:32 <ais523> one big advantage is that if you use the right number of bits, you can make the number of available integers a prime
19:20:40 <ais523> so that multiplication is completely reversible
19:21:01 <slereah_> Hello ais523.
19:21:07 <ais523> hello slereah_
19:22:50 <ais523> hmm... my rant about 1's complement doesn't seem to have sparked off a discussion yet
19:23:50 <slereah_> Well, I don't know if the others are here, and I don't know what's 1's complement.
19:24:13 <ais523> most computers represent negative numbers using 2's complement
19:24:38 <ais523> given that you're storing numbers in binary (unless you're using TriINTERCAL or Malbolge, this is likely), and that you have a fixed number of bits (i.e. not a bignum)
19:24:56 <ais523> then you form the 2's complement of a number by changing all 0s to 1s in the number and 1s to 0s, and then adding 1
19:25:03 <ais523> 1's complement is the same except that you don't add the 1
19:25:08 <slereah_> Oh.
19:25:40 <ais523> the advantage of 2's complement, which is the reason it's so commonly used, is that signed and unsigned arithmetic work the same way; for signed arithmetic, you just disregard the carry
19:25:43 <slereah_> Wouldn't you have problems differentiation between, say, -1 and 255?
19:26:01 <ais523> with an 8-bit system, you'd generally range from -127 to +127
19:26:08 <slereah_> Oh.
19:26:22 <ais523> 2's complement would go from -128 to +127
19:26:42 <ais523> (1's complement has two sorts of zero, which act identically, but can be confusing to someone used to how arithmetic normally works)
19:27:07 <ais523> They're normally referred to as +0 and -0, but both act like 0 with respect to all the usual arithmetic operations
19:28:42 <ais523> the real interest comes when the number of bits you use is not a power of 2
19:28:49 <ais523> (to be precise, when 2^n-1 is a prime)
19:29:05 <ais523> because then multiplication by any integer except 0 maps all integers onto unique integers
19:29:38 <ais523> and it's almost as easy to implement in hardware as 2's complement (at least in terms of wiring)
19:29:50 <ais523> it's slower on average, but the worst-case response time is the same
19:30:39 <slereah_> So where's the gain?
19:31:29 <ais523> more elegant maths IMO
19:31:50 <ais523> oh, and you can pick one of the zeros to use as 0 and the other one to detect uninitialised data
19:33:59 <slereah_> "More difficult to understand quite unlike a "computer" -- abstract or otherwise."
19:34:04 <slereah_> Boy they've got that right.
19:34:51 <ais523> what are you referring to?
19:35:40 <slereah_> It's from the wiki article on pointers machines.
19:38:51 <slereah_> It seems to be one of the most obscure computing method.
19:39:46 <ais523> yes
19:39:54 <ais523> I had a vague memory that such things existed
19:39:59 <ais523> but don't really understand them
19:40:40 <slereah_> Is there anything based on it?
19:40:58 <slereah_> It would be nice to have a concrete example.
19:44:29 <ais523> after reading a few Google links I think I understand what's happening to some extent now
19:45:53 <ais523> basically, at any given point the memory is a graph, with nodes connected to each other by directed coloured edges
19:46:13 <slereah_> Yes.
19:46:24 <slereah_> Though they don't have to be directed apparently.
19:46:38 <ais523> and at each step, there are replacement rules, such as 'if a node is connected to another node via a red edge and then a blue edge, add a green edge pointing to that node from the original node'
19:46:55 <ais523> hmm... maybe not quite like that, I still don't understand them properly
19:47:12 <slereah_> I understand roughly how it works.
19:47:17 <slereah_> There's plenty of examples
19:47:26 <slereah_> Proble is, I don't get much how it computes.
19:47:33 <ais523> that's usual with esolangs
19:47:50 <slereah_> None of the examples are using problems outside of the graph itself.
19:47:51 <ais523> one of the articles said that you could make a cellular automaton by connecting the original graph up as a grid
19:48:07 <slereah_> Swell.
19:49:31 <slereah_> There's an example that's sort of useful, though not that clear.
19:49:47 <slereah_> It shows how to do a list of integers.
19:50:13 <ais523> where?
19:50:38 <slereah_> "On the definition of an algorithm", by Andrei Kolmogorov and Uspenskii.
19:50:51 -!- wildhalcyon has joined.
19:51:38 <slereah_> Though the Kolmogorov definition adds some strange conditions, like every nodes must have a type and a node can't be connected to two identical types.
19:52:28 <slereah_> So there's a lot of nodes type 2-3-4-2-3-4-2-3-4... because of it.
19:52:42 <slereah_> (0 is the starting node type and 1 ending type)
20:11:37 -!- olsner has joined.
20:15:00 <SimonRC> hm
20:15:35 <SimonRC> dunnet FTW
20:16:01 <slereah_> Whut?
20:16:16 <SimonRC> specifically, M-x dunnet
20:16:40 <oerjan> whodunnet?
20:16:48 <SimonRC> no
20:16:51 <SimonRC> a game
20:16:58 <SimonRC> a text adventure
20:17:11 <SimonRC> though best run with emacs -batch -l dunnet
20:19:01 <Sgeo> There's minigolf in movie 4
20:19:20 <SimonRC> huh, what, "movie"?
20:20:00 <Sgeo> Futurama movie
20:20:15 <SimonRC> ok
20:20:20 <SimonRC> relevent to anything?
20:20:28 <Sgeo> no
20:20:31 <Sgeo> wrong chan basically
20:20:35 <slereah_> Well, it's Futurama, so probably!
20:29:26 <ais523> SimonRC: ever tried Lost Kingdoms?
20:29:31 <ais523> I'm stuck on the wildcat
20:29:52 <SimonRC> ais523: no
20:30:30 <ais523> I mentioned this to ehird, and wrote a serialising BF interp so that I could send a save file, but ehird couldn't solve that bit either
20:43:47 <SimonRC> wasn't it originally written in bfbasic?
20:44:23 <SimonRC> hmm
20:44:30 <SimonRC> maybe that page is wrong?
20:45:19 <ais523> there are two versions
20:45:22 <ais523> the original was in BASIC
20:45:33 <ais523> an improved version was written in BFBASIC and compiled to BF
20:45:50 <SimonRC> is there any bfbasic source?
20:46:20 <ais523> not that I know of
20:46:27 <ais523> even the original BASIC source is in binary
20:46:40 <ais523> I suppose you could try decompiling, because BFBASIC itself is open-source
20:46:55 <ais523> but someone would have to write a decompiler first
20:48:03 <SimonRC> the actual page only says that the original LK was written in bbc basic and translated to brainfuck...
20:48:11 <SimonRC> it doesn't say anything about bfbasic
20:48:47 <ais523> it did use BFBASIC, though, I'm pretty sure of it
20:49:04 <ais523> especially because the patterns at the end of the source are characteristic of it
20:50:59 <ais523> FWIW, Esolang's entry on Brainfuck states that BFBASIC was used
20:51:09 <SimonRC> but that's the only source I can find
20:51:20 <SimonRC> apart from various translations of it
20:52:28 <ais523> there aren't many mentions of that game on the Internet anyway, for some reason
20:53:09 <ais523> also, note the mention of Jeffry Johnston in the credits at the bottom of its page
20:53:32 <ais523> Jeffry Johnston created BFBASIC, and it also has contributions by Jon Ripley, which would seem to make the situation even more likely
20:53:52 <SimonRC> hmm
20:54:03 <SimonRC> though the bf world is quite small isn't it?
20:54:29 <oerjan> only 5 feet across
20:54:31 <ais523> not as small as the rest of the esolang world
20:54:35 <slereah_> It's the biggest communauty in esolangs !
20:54:54 <slereah_> Though I wonder if Intercal isn't used more.
20:55:01 <slereah_> It has seniority and all.
20:55:08 <SimonRC> comp.lang.intercal has com activity
20:55:12 <ais523> INTERCAL is probably used a lot
20:55:19 <ais523> and don't you mean alt.lang.intercal
20:55:23 <ais523> or are there really two newsgroups?
20:55:49 <slereah_> Well, I wouldn't go as far as "lot".
20:56:01 <ais523> (apparently ~50 Debian users who send popularity contest stats installed C-INTERCAL, ~30 installed CLC-INTERCAL, single-figures > 1 actually use them)
20:57:06 <ais523> latest stats from http://qa.debian.org/popcon.php?package=intercal are that 4 people used C-INTERCAL recently
20:57:31 <ais523> (bear in mind that this is restricted to Debian Linux users with popularity stats turned on, so the true number is likely much higher)
20:57:49 <ais523> and 3 used CLC-INTERCAL
20:58:03 <ais523> either the total installed numbers have gone down, though, or my memory of them was wrong
20:58:51 <ais523> oh, add one to the C-INTERCAL stats for me, because although I use Ubuntu, I don't have C-INTERCAL installed from packages as I'm the maintainer and have my own newer test versions
21:01:11 <ais523> (I do show up on the CLC-INTERCAL stats, I think)
21:01:59 <RodgerTheGreat> I really think lightning-fast mental math is neat: http://www.ted.com/index.php/talks/view/id/199
21:02:57 -!- GregorR has quit (Success).
21:03:15 -!- wildhalcyon has quit ("ChatZilla 0.9.81 [Firefox 2.0.0.11/2007112718]").
21:03:23 -!- EgoBot has quit (Read error: 110 (Connection timed out)).
21:03:24 -!- GregorR has joined.
21:05:06 <ais523> <-- GregorR has left this server (Success).
21:05:21 <ais523> no double quotes, so that was a server-generated quit message
21:05:29 <ais523> (presumably a case of errno=0 but an error happened anyway)
21:05:32 <SimonRC> odd
21:05:41 <ais523> I came across a program recently which was exiting with "success" as the error message
21:05:49 <ais523> because it was using perror but nothing had set errno
21:06:17 <ais523> (apparently on that system mmap doesn't always set errno on failure)
21:06:20 <SimonRC> that video seems to have no image
21:06:56 <SimonRC> is there supposed to be one?
21:07:25 <ais523> SimonRC: I don't understand what you're trying to say, can I have some context or did you type into the wrong channel?
21:07:38 <SimonRC> the video that RodgerTheGreat posted
21:07:42 <ais523> ah, you're referring to RodgerTheGreat's comment
21:08:05 <RodgerTheGreat> hunh. Tried refreshing?
21:08:12 <SimonRC> yes
21:08:18 <SimonRC> am I missing much
21:08:31 <RodgerTheGreat> not really
21:08:43 <RodgerTheGreat> the audio will honestly give you most of it
21:12:15 -!- oerjan has quit ("Good night").
21:13:01 <SimonRC> ah, the "missing digit" one is via casting out nines, I suspect
21:14:02 <SimonRC> and the previous ones are simplified by knowing your times tables up to 100
21:14:56 <SimonRC> thoguh eh might not
21:15:32 * SimonRC tries to recall the Lewis Caroll technique for finding the day of the week of a date.
21:18:33 <SimonRC> ah, yes, he is using his tables up to 100
21:19:16 <SimonRC> ooh, nice memorisation hack
21:19:20 <SimonRC> (though it is standard)
21:20:18 <RodgerTheGreat> pretty simple techniques, he simply applies them with tremendous speed and accuracy
21:21:57 <RodgerTheGreat> this is the same kind of thing: http://www.youtube.com/watch?v=wIiDomlEjJw
21:23:24 <SimonRC> otoh, see the tale of richard feynmann and the abacus-seller
21:23:52 <RodgerTheGreat> link?
21:26:22 <SimonRC> erm, in the book "Surely You're Joking Mr Feynmann"
21:26:27 <SimonRC> a great book for any geek
21:26:40 <RodgerTheGreat> I have heard it's very entertaining.
21:26:48 <RodgerTheGreat> hm... lemme see how much it is on Amazon...
21:27:47 <RodgerTheGreat> about $8 including shipping.
21:28:27 <ais523> I've read it too
21:28:37 <ais523> it's a good book, but the sequel is much worse
21:28:48 <SimonRC> sequel?
21:28:50 <ais523> (I can't remember what the sequel is called)
21:28:58 <ais523> (probably because it wasn't very good)
21:29:10 <SimonRC> amazon should know
21:29:18 <SimonRC> or WP for that matter
21:52:44 <SimonRC> wow: http://www.pdp8.net/sound/sound.shtml
21:53:08 -!- calamari has joined.
21:58:16 -!- jix has quit ("CommandQ").
22:05:43 -!- ais523 has quit ("my quit message will never be as good as GregorR's :(").
23:06:43 -!- RedDak has quit (Remote closed the connection).
23:11:16 * SimonRC goes.
23:12:03 <slereah_> bais
23:34:13 -!- timotiis has quit ("leaving").
23:54:09 -!- Corun has joined.
←2008-02-26 2008-02-27 2008-02-28→ ↑2008 ↑all