00:00:44 <RodgerTheGreat> almost every program consists of an initialization chunk, a big loop, some smaller loops called from the big loop and an optional cleanup when the program ends
00:00:55 <oerjan> Writing a kernel without understanding hardware is probably like trying to swim without entering the water. :)
00:01:23 <RodgerTheGreat> "possible to learn, but ultimately not very applicable"
00:12:33 <oerjan> In C I would expect the worst part to be the memory management.
00:17:17 <oerjan> what, the hm... part? :)
00:17:34 <RodgerTheGreat> I guess that as well as the memory management in C sucking things
00:17:44 <lament> scorpions and apples? Who added scorpions?
00:21:45 <SevenInchBread> ...with the ability to shift back into manual transmission (lol car analogy) when you want.
00:23:37 <oerjan> the simplest way to handle overflow would probably be to have more bits in the temporary variables than in each array cell
00:33:42 -!- nazgjunk has quit ("sleep").
00:36:35 <SevenInchBread> I think we should go with a far far unmonolithic kernel than say.... Linux. :) :) :)
00:37:41 <SevenInchBread> ....with many many many filesystems... that you can switch between.
00:41:32 <oerjan> TRUE esotericness would probably require a WYRD filesystem.
00:41:47 <SevenInchBread> file systems that look like MUDs - super-compressed file systems - stenographic/cryptographic filesystems - a filesystem based on worms.... somehow
00:46:44 <RodgerTheGreat> a filesystem consisting of rainbow tables in which you reverse-engineer MD5s instead of editing files
00:48:03 <RodgerTheGreat> SevenInchBread: by worms, do you mean rogue processes carrying data around in the computer, or physical earthworms? Or both/either?
00:48:36 <lament> gigantic sandworms roaming below the surface producing spice.
00:55:25 <SevenInchBread> ...well, I'm kind of convinced that... to store data... you need some method of organizing the data - regardless of how you store it.
00:55:26 -!- ShadowHntr has joined.
00:56:10 <oerjan> What an antiquated conviction.
00:57:14 -!- sebbu has quit (Client Quit).
00:57:44 <SevenInchBread> whether you call it a file system or not is regardless.
01:00:07 <SevenInchBread> couldn't you just use longs and forget about overflow?
01:01:40 <oerjan> on the contrary, base 3^42 is best.
01:12:31 <SevenInchBread> interprocess communication could use filesystems to communicate instead of the usual "stream".
01:29:18 <SevenInchBread> hmmm... representing files as linked lists allows you to create infinite files - and tree structures.
01:31:55 <oerjan> representing files as lazy thunks allows infinite files without loops.
01:32:27 <oerjan> (This message was presented to you by the Haskellers for World Domination)
01:32:33 <bsmntbombdood> represent files as lambd terms allows a turing complete filesystem
01:33:53 <SevenInchBread> so.... basically... a simple change in the OS's representation of a file gives you quite a number of possibilities for structure... universally accessible from any language.
01:36:07 <oerjan> Wicked, dysfunctional people might point out that a next step could be using persistent objects.
01:38:53 <oerjan> and at this point the relational database people will probably want a word in edgewise
01:39:45 <oerjan> that is when we throw them out of #esoteric and decide each file should be a running brainfuck program.
01:40:10 <SevenInchBread> ....brainfuck, coincidentally, is a very handy representation of the typical file. :)
01:40:54 <bsmntbombdood> pause execution while there's no IO with that file
01:41:32 <SevenInchBread> or just use... running brainfuck programs... file writes would just be like an input.
01:42:05 <SevenInchBread> so no waiting... just pause when we want input... and then all the changes are persistent...
01:44:08 * oerjan thinks that Unlambda could hot-swap pretty easily.
01:44:33 <oerjan> of course its I/O sucks.
01:45:14 <oerjan> you must test every input against every possible value
02:03:29 <lament> SevenInchBread: how's that different from brainfuck?
02:06:42 <lament> by the way, i certainly think an esoteric OS shoudn't have files
02:09:22 <oerjan> very small gnomes carrying papyrus rolls
02:11:38 <lament> bsmntbombdood: some sort of objects
02:12:05 <lament> bsmntbombdood: a "file" is raw data that anything can read. It's a fairly bizarre concept.
02:12:31 <lament> for example, i have an mp3 file
02:12:42 <lament> why should i be able to open it with a text editor?
02:12:49 <lament> it makes no sense. It breaks abstraction.
02:13:19 <SevenInchBread> all things shall be the POLYNUMEROUS PAM... which is left implemtnation specific, to confuse everyone as to what it is.
02:13:22 <lament> on the other hand, look at PalmOS for a working example
02:14:07 <SevenInchBread> and all forms of persistence, memory management, back scratching, and ultimate deliberation... shall be of the POLYNUMEROUS PAM .
02:16:03 <lament> SevenInchBread: polyamorous
02:16:38 <lament> bsmntbombdood: only in the same sense that "everything is a number"
02:17:00 <bsmntbombdood> files are described by numerical file descriptors ;)
02:17:12 <lament> bsmntbombdood: it's not the kind of an abstraction a decent OS should provide. It's too low-level.
02:17:25 <SevenInchBread> all those "everything is a <insert favorite representation>" people scare me.... it's like religious adherance to some weird all-powerful principle.
02:18:02 <SevenInchBread> ....instead, those people should believe that EVERYTHING SHALL BE OF POLYNUMEROUS PAM.
02:19:29 <SevenInchBread> ....everything in esoos should be represented in overly verboose XML....
02:26:32 -!- ShadowHntr has quit ("End of line.").
02:46:39 -!- crathman has joined.
03:17:13 -!- oerjan has quit (zelazny.freenode.net irc.freenode.net).
03:17:28 -!- oerjan has joined.
03:42:20 -!- SevenInchBread has quit (Remote closed the connection).
04:38:30 -!- oerjan has quit ("leaving").
04:38:34 -!- crathman has quit ("Chatzilla 0.9.77 [Firefox 2.0.0.3/2007030919]").
05:03:37 -!- oerjan has joined.
05:22:33 -!- anonfunc has joined.
05:34:12 -!- ShadowHntr has joined.
05:36:31 -!- digital_me has quit (Read error: 110 (Connection timed out)).
05:47:13 -!- Arrogant has joined.
06:02:05 -!- Sgeo has quit ("Leaving").
06:25:38 -!- mule| has quit (Read error: 110 (Connection timed out)).
07:09:11 -!- sekhmet has quit (Remote closed the connection).
07:09:14 -!- sekhmet has joined.
07:28:01 -!- ShadowHntr has quit (Client Quit).
07:30:49 -!- anonfunc has quit (Read error: 104 (Connection reset by peer)).
07:30:52 -!- anonfunc has joined.
07:44:58 <oklopol> haha... i scroll up and say something that'd be weird even there :P
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:10:40 -!- RodgerTheGreat_ has joined.
08:10:40 -!- RodgerTheGreat has quit (Read error: 104 (Connection reset by peer)).
09:14:16 -!- Arrogant has quit ("Leaving").
10:20:03 -!- mule| has joined.
10:28:18 -!- jix has joined.
10:58:39 -!- helios_ has joined.
10:59:08 -!- helios_ has quit (Client Quit).
11:00:36 -!- helios24 has joined.
11:01:48 -!- helios24 has quit (Client Quit).
11:03:18 -!- helios24 has joined.
11:04:15 -!- anonfunc has quit.
11:04:20 -!- helios24 has quit (Client Quit).
11:04:42 -!- helios24 has joined.
11:12:33 -!- ais523 has joined.
11:19:17 -!- jix has quit ("Bitte waehlen Sie eine Beerdigungnachricht").
11:31:18 <SimonRC> idea: every file is a datastructure equivalent to XML...
11:31:38 <SimonRC> ... but you have to all IO on the file as raw characters of XML
11:31:51 <SimonRC> actually that's just perverse, rather than esoteric
11:33:50 <ais523> why does everyone like XML so much these days. Why not a readable format that saves on characters (IMO JSON is more readable and more concise but not perfect)
11:34:40 <SimonRC> We talk about XML a lot here because it is easy to take the piss out of it.
11:35:33 <SimonRC> classic example: http://www.codu.org/rxml.php
11:35:43 <ais523> <IRCpost type="comment">because everything in XML has to be formatted more redundantly than needed </IRCpost>
11:36:13 <ais523> Why does nobody try to format XML in XML?
11:36:55 <ais523> <tag type="IRCpost"><attribute name="type" value="comment"><contents>because everything in XML has to be formatted more redundantly than needed</contents></tag>
11:37:27 <ais523> (and of course this can be continued ad infinitum if this is believed to be beneficial)
11:37:42 <SimonRC> BTW, codu is GregorR's site: http://www.codu.org/hats.php
11:40:08 <ais523> one idea I had for a nonstandard file structure that might actually be useful is to change the one-many relationship between directories and files into a many-many relationship
11:40:34 <ais523> and then loop it round so that any directory could be a subdirectory of any other directory (with the system maintaining lists of backlinks)
11:41:01 <ais523> in fact, you could go further and use just a combined file/directory concept with some contents and a list of files it linked to
11:41:01 <SimonRC> so you have a netwok rather than a tree
11:41:15 <SimonRC> the problem with that is that stuff can get lost
11:41:30 <ais523> it's a directed network where you're allowed to go backwards along the lines if you want, as long as you say what you're doing
11:41:34 <SimonRC> you can have files that are reachable but you have no idea how to reach them
11:41:48 <SimonRC> and things lack a canonical name
11:41:58 <ais523> well, it would be like an ordinary directory structure with more links
11:42:10 <ais523> and as for canonical names, modern filesystems don't really have them anyway
11:42:23 <ais523> irc://irc.freenode.org/esoteric/../esoteric
11:42:31 <ais523> (I have no idea if that would actually work!)
11:43:08 <ais523> (BTW, when I clicked on that link, nothing happened at all, but I'm not sure why)
11:43:09 -!- helios24 has quit ("leaving").
11:43:09 <SimonRC> would there be a root node?
11:43:17 -!- helios24 has joined.
11:43:42 <ais523> at least one. Probably one for the system and one for each user on a multi-user system, like c:\ and My Documents\ on Windows, or / and ~ on Unix
11:43:55 <SimonRC> you know about the plan 9 FS, right/
11:44:07 <ais523> I'm not too aware of that,
11:44:41 <SimonRC> it implements much of its security by not even giving people a *name* for things they aren't supposed to know about.
11:44:48 -!- helios_ has joined.
11:44:48 -!- helios_ has quit (Client Quit).
11:45:08 <ais523> The two systems could be combined easily enough
11:45:10 <SimonRC> there is lots of mounting and union directories
11:46:02 <SimonRC> silly idea: you have union directories in Plan9FS, so EsoFS should have intersection directories, set difference directories, and complement directories.
11:46:36 <SimonRC> actually, intersection directories don't sound like a bad idea
11:46:45 <ais523> That could actually be useful. I'd like to see the symmetric difference between C-INTERCAL versions 0.24 and 0.25
11:46:46 <SimonRC> but complement directories sure do!
11:47:30 <SimonRC> you could have a directory containing all the picture files you have, and another containing all the files from 2003...
11:47:37 <ais523> I've heard it suggested that a big long list of 'every file on the computer' would actually be easier to navigate than other systems
11:47:44 <SimonRC> the intersection of these would be all your pictures from 2003
11:48:02 <ais523> Your directory system would be a fully-featured categorisation system
11:48:08 <SimonRC> I would like that featuer a lot actuall
11:48:46 <SimonRC> a sort of cross between a big directory tree and a collection of DB tables
11:48:51 <ais523> In fact, wiki software has been doing this sort of thing for ages; in MediaWiki, the links and backlinks do what I suggest, the categories would do what you suggest if the developers got round to it, and subpages give you canonical names
11:49:39 <ais523> Another thing I'd like is unlimited amounts of metadata in files
11:49:48 <ais523> We have file-size and permissions at the moment
11:50:14 <oerjan> Macs have that don't they?
11:50:37 <ais523> I'd like a list of metadata so that we can find the right program to open a file, even if we've never come across the format before
11:50:45 <SimonRC> I personally don't like any metadata system between the small system unix uses and the fully-fleged resource fork of older macs
11:51:14 <ais523> Ideally, we want to load our BF programs in a BF IDE, but Emacs will do if we don't have one (as they're text files) and in the worst case a hex editor
11:51:33 <ais523> (or if really needed, you could edit the file using cat)
11:52:11 <SimonRC> BF programs are basically trees of strings of "<>+-,.", right?
11:52:28 * ais523 is the sort of person who uses cat /dev/random to get their terminal out of the wrong shift state
11:53:12 <ais523> in fact, the actual characters used don't really matter IMO; <>+-., are just conventional so that people can communicate BF programs and have them mean the same thing at their destination
11:53:19 <oerjan> there should also be a way to get a list of options for opening, including options in the metadata as well as the programs you have installed (in separate lists)
11:53:55 <ais523> you could have a 'relevance' result like you do on search results, according to how suited a program would be for a file
11:54:05 <ais523> and each program could have its own algorithm
11:54:30 <ais523> so a plaintext editor is going to be moderately-high-relevance for a plain-text file containing text
11:54:37 <oerjan> heh, so that what google's "I'm feeling lucky" really is - the default open
11:55:06 <ais523> and moderate relevance for a program in most non-esoteric languages, and low but still positive relevance for a uuencoded binary
11:55:59 <ais523> you know, esoteric programming languages are reaching the state where they are actually useful in some cases
11:56:27 <ais523> (For instance, ALPACA gives one of the simplest Life implementations around, and BF (not C89) is probably now the world's most portable language)
11:56:40 <SimonRC> <cough>Linear Lisp</cough>
11:57:44 * ais523 is being chased all over the Internet by SimonRC at the moment
11:59:29 <ais523> the principle of building enforced garbage-collection into all your commands is really quite esoteric
11:59:56 <ais523> in fact, it's logic is like a functional version of BF (you have to decompose something whilst creating two copies if you want to copy it)
12:01:06 * ais523 thinks that clog should respond to sed commands so that you can correct your typos in the logs too
12:03:10 <SimonRC> ISTR there is a plug-in for some IRC clients that does that for the client.
12:03:39 <ais523> I wouldn't be surprised; surely someone had the same idea before me
12:06:34 -!- helios24 has quit ("leaving").
12:06:42 -!- helios24 has joined.
12:47:11 -!- jix has joined.
12:54:11 <ais523> ~dof ,[.,]!Most simple BF programs also work in DoFuck
12:54:36 <ais523> (that's strange, I got the output in #bsmnt_bot_errors rather than here)
12:54:43 <ais523> ~exec exec self.bf7+self.bf8
12:54:47 <ais523> ~dof ,[.,]!Most simple BF programs also work in DoFuck
12:54:47 <bsmnt_bot> Most simple BF programs also work in DoFuck
12:54:47 <bsmnt_bot> Most simple BF programs also work in DoFuck
12:55:19 <ais523> ~exec self.raw_regex_queue.pop()
12:55:21 <ais523> ~dof ,[.,]!Most simple BF programs also work in DoFuck
12:55:22 <bsmnt_bot> Most simple BF programs also work in DoFuck
12:57:17 <oerjan> ~exec sys.stdout(os.listdir("/bot/scripts/"))
12:57:51 <ais523> I've taken to messing with bsmnt_bot in #bsmnt_bot_errors to avoid spamming #esoteric
12:58:23 <oerjan> oh. well i just wanted to remind you that you can save files in /bot/scripts/ now.
12:58:51 <ais523> how? We could put the BF and DoFuck interpreters in there
12:59:06 <oerjan> just with the usual python file commands
12:59:39 <ais523> (I don't actually know much Python at all; I've picked it up mostly from the #esoteric logs)
13:00:10 <ais523> besides, what does it do with the files there? Does it run them on startup, or do we have to run them manually, or what?
13:03:38 <ais523> ~exec with open("/bot/scripts/bf.py") as f: write bf3+bf4
13:03:55 <ais523> ~exec exec "with open("/bot/scripts/bf.py") as f:\n write bf3+bf4"
13:04:10 <ais523> ~exec exec "with open('/bot/scripts/bf.py') as f:\n write bf3+bf4"
13:04:27 <oerjan> the bot is still running an old version of python. i tried the same as you...
13:05:06 <ais523> ~exec exec "f=open('/bot/scripts/bf.py')\nf.write(bf3+bf4)\nf.close()"
13:05:06 <bsmnt_bot> IOError: [Errno 2] No such file or directory: '/bot/scripts/bf.py'
13:05:28 <oerjan> that was opened for reading
13:05:56 <ais523> ~exec exec "f=open('/bot/scripts/bf.py','w')\nf.write(bf3+bf4)\nf.close()"
13:05:57 <bsmnt_bot> NameError: name 'bf3' is not defined
13:06:03 -!- jix has quit ("Bitte waehlen Sie eine Beerdigungnachricht").
13:06:06 <ais523> ~exec exec "f=open('/bot/scripts/bf.py','w')\nf.write(self.bf3+self.bf4)\nf.close()"
13:06:19 <ais523> ~exec sys.stdout(os.listdir("/bot/scripts/"))
13:06:34 * ais523 came to the same conclusion as oerjan at about the same time
13:06:46 <ais523> ~exec exec "f=open('/bot/scripts/dof.py','w')\nf.write(self.bf7+self.bf8)\nf.close()"
13:06:50 <ais523> ~exec sys.stdout(os.listdir("/bot/scripts/"))
13:07:06 <ais523> ~exec self.raw_register_queue.pop()
13:07:07 <bsmnt_bot> AttributeError: IRCbot instance has no attribute 'raw_register_queue'
13:07:08 <ais523> ~exec self.raw_register_queue.pop()
13:07:09 <bsmnt_bot> AttributeError: IRCbot instance has no attribute 'raw_register_queue'
13:07:16 <ais523> ~exec self.raw_regex_queue.pop()
13:07:21 <ais523> ~exec self.raw_regex_queue.pop()
13:07:28 <oerjan> ~exec sys.stdout(self.bf7)
13:08:07 <ais523> I was about to warn you against that when you corrected your command
13:08:23 <ais523> ~bf ,[.,]!Definitely stopped this working?
13:08:47 <ais523> now all we have to do is reload it from the directory, somehow
13:10:31 <ais523> ~exec exec "f=open('/bot/scripts/bf.py')\nexec '\n'.join(f.readlines())\nf.close()"
13:10:31 <bsmnt_bot> SyntaxError: EOL while scanning single-quoted string
13:10:44 <ais523> ~exec exec "f=open('/bot/scripts/bf.py')\nexec '\\n'.join(f.readlines())\nf.close()"
13:11:18 <ais523> ~exec exec "f=open('/bot/scripts/dof.py')\nexec '\\n'.join(f.readlines())\nf.close()"
13:12:18 <ais523> (because outputting \0 with bsmnt_bot has no obvious effect, cat is actually one character shorter in DoFuck)
13:13:30 <ais523> oerjan: if you're interested in the code and don't want to flood the channel, do this:
13:13:38 <ais523> ~exec sys.stdout(repr(self.bf7))
13:13:38 <bsmnt_bot> "def bfarg(x,y):\n p=y.group(2)\n a=y.group(3)+unichr(0)\n o=''\n p=p+'!'\n t=[0]*30000\n i=0\n l=0\n while p[i]!='!':\n if p[i]==']' and t[l]!=0:\n c=1\n while c>0:\n i=i-1\n if p[i]==']': c=c+1\n if p[i]=='[': c=c-1\n"
13:13:41 <ais523> ~exec sys.stdout(repr(self.bf8))
13:13:42 <bsmnt_bot> " if p[i]=='+': t[l]=t[l]+1\n if p[i]=='-': t[l]=t[l]-1\n if p[i]=='<': l=l-1\n if p[i]=='>': l=l+1\n if p[i]=='.': o=o+unichr(t[l])\n if p[i]==',':\n t[l]=ord(a[0])\n a=a[1:]\n i=i+1\n sys.stdout(o)\nself.register_raw(r'\\S+ PRIVMSG (\\S+) :~dof ([^!]*)!?(.*)',bfarg)"
13:14:40 <ais523> Oh by the way, I invented a new joke language in my sleep last night. Look at this:
13:15:20 <ais523> I'll have to write Huh-language into bsmnt_bot instead, then.
13:16:10 <ais523> ~exec self.register_raw(r'\S_ PRIVMSG \S+ :~huh', lambda x,y:sys.stdout('Huh?'))
13:16:26 <ais523> ~exec self.register_raw(r'\S+ PRIVMSG \S+ :~huh', lambda x,y:sys.stdout('Huh?'))
13:17:52 <ais523> It's an esoteric language because it was already programmed into EgoBot before hand, just nobody noticed before
13:18:08 <ais523> I don't think it's Turing-complete
13:20:03 <bsmnt_bot> <__main__.IRCbot instance at 0xb7c9140c>
13:20:45 <ais523> what did you expect that line to do?
13:20:53 <oerjan> it ought to be possible to use execfile if the scripts is designed to modify bot
13:22:22 <ais523> ~exec self.raw_regex_queue.pop()
13:22:23 <ais523> ~exec self.raw_regex_queue.pop()
13:22:25 <ais523> ~exec self.raw_regex_queue.pop()
13:22:26 <ais523> ~exec self.raw_regex_queue.pop()
13:22:35 <ais523> ~exec execfile('/bot/scripts/bf.py')
13:22:45 <ais523> ~exec execfile('/bot/scripts/dof.py')
13:23:05 <ais523> oerjan: I think execfile works anyway in this case
13:23:32 <oerjan> hm... i suppose so, given the way ~exec is constructed
13:24:16 <bsmnt_bot> <__main__.IRCbot instance at 0xb7c9140c>
13:24:17 <bsmnt_bot> <__main__.IRCbot instance at 0xb7c9140c>
13:24:19 <bsmnt_bot> <__main__.IRCbot instance at 0xb7c9140c>
13:24:55 * oerjan wonders what confuses SimonRC
13:25:15 * ais523 wonders why SimonRC gave the same command three times
13:25:37 <oerjan> i didn't actually mean to print out the contents, just check that it had the value i thought
13:30:56 <SimonRC> ais523: to see if it gave the same answer
13:31:18 <ais523> I thought it might be something like that
13:33:28 <ais523> BTW, I take it MechaniQue is Turing-complete? From the wiki description, it looks like it would be easy to write a Minsky machine in it
13:34:04 <ais523> (It's also the only language other than INTERCAL I know of that uses NEXT for flow control; but it doesn't seem to have a FORGET statement or RESUME for values >1)
13:35:39 <oerjan> BASIC uses NEXT, although not in the same way.
13:36:56 <ais523> yes, BASIC uses NEXT as a punctuation mark (like } in C or cis in Perligata). A 'NEXT statement' to me would always imply the INTERCAL version
13:37:08 -!- sebbu has joined.
13:37:52 -!- bsmnt_bot has quit (Excess Flood).
13:38:14 <ais523> whoops, that was me, sorry
13:38:21 <ais523> but I thought it had been throttled?
13:38:56 <SimonRC> In the sense of strangling you.
13:39:48 <ais523> I asked it for all the attributes of the 'self' object and their values, without realising there were quite so many
13:40:56 <ais523> still, I discovered that 1174566977.4449899 was on its list of currently running commands
13:43:24 <ais523> so now the only bot we have running is clog?
13:45:35 <ais523> yes it is, I missed it
13:46:28 <oerjan> the good news is that for the first time, not all work was lost :)
13:47:18 <oerjan> i don't think the flood protection works for long lines.
13:47:34 <oerjan> only for many short ones
13:47:41 <ais523> it had split its message over three maximum-length lines and was kicked when it tried to write the fourth
13:48:15 <ais523> (and not all work has been lost for ages, I normally just copy the most-recently-working commands back out of the logs and type them in again)
13:55:38 <ais523> of course, #bsmnt_bot_errors isn't logged, so I might have to use a different method to save the experiments
13:57:41 <oerjan> like logging it yourself.
14:08:32 <ais523> strangely enough, my client's so old and/or primitive it appears not to have a logging function
14:08:46 <ais523> it doesn't matter though, I'll find a way (copy-and-paste has worked in the past)
14:10:13 <ais523> I know it's so primitive that I have to identify to nickserv by hand
14:13:56 <oerjan> can't you give a password on startup? that has been supported since the first IRC RFC.
14:15:14 <ais523> No. The only startup option is to run a JavaScript script, with scant information about how to use that to give a password or to message NickServ
14:15:35 <oerjan> you don't need to message nickserv. i don't.
14:15:57 <oerjan> the ordinary irc password is passed on by the server to nickserv.
14:16:07 <ais523> you have a choice: message NickServ or give a server password. If you can't give the password, messaging NickServ is given as a workaround
14:16:53 <oerjan> so basically you have a client that is _worse_ than some older ones.
14:17:21 <oerjan> are you using a URL for irc login?
14:17:51 <ais523> it's about 50-50 between using a URL and using /join
14:18:09 <oerjan> there might be a spot in the URL for login/password.
14:19:20 -!- ais523 has quit ("Testing").
14:20:26 -!- ais523 has joined.
14:20:45 <ais523> it didn't work; I got 'invalid IRC URL: (null)'
14:21:36 <ais523> I think it had problems trying to parse the resulting URL
14:24:15 <ais523> Microsoft's website about including username/password in URLs (it's about http, but it's the same for ftp and presumably irc too) is amusing
14:24:42 <ais523> http://support.microsoft.com/kb/834489
14:25:59 <ais523> it seems there are two workarounds for Microsoft disabling the feature: don't include a username-password pair (which is pretty obvious, really)
14:26:36 <ais523> and /instructing users/ not to include a username-password pair
14:28:54 <oerjan> i found a draft IRC URL scheme which conspicuously does not include a spot for password, just a request to prompt the user for one.
14:29:20 <ais523> I suppose putting a password in a URL is a security risk
14:29:36 <ais523> but probably not much more than sending it over IRC in plaintext
14:32:15 <ais523> heh, Microsoft's website also has step-by-step instructions for opening the box that the software comes in
14:47:23 -!- mtve has quit ("Terminated with extreme prejudice - dircproxy 1.0.5").
14:47:36 -!- mtve has joined.
15:41:20 -!- felixx has joined.
16:21:21 -!- ais523 has quit.
16:41:50 -!- jix has joined.
16:54:26 -!- ais523 has joined.
17:37:42 -!- felixx has left (?).
17:53:07 -!- oerjan has quit ("leaving").
18:05:51 -!- RodgerTheGreat_ has quit.
18:16:50 -!- RodgerTheGreat has joined.
18:31:30 -!- SevenInchBread has joined.
18:32:38 <SevenInchBread> ...have their been any efforts to convert the http://en.wikipedia.org/wiki/Voynich_manuscript into a computer-readable encoding?
18:33:56 <ais523> I thought that many traditional efforts to solve that assumed that the text itself was irrelevant and the code was encoded in the metadata
18:45:53 <ais523> SevenInchBread: Your new Dupdog-like language with a name I can't remember or type: do you think Wiki Cyclic Tag would be suitable as a language in it? Your new language provides the framework needed to make it Turing-complete by itself
18:50:32 -!- nazgjunk has joined.
18:51:27 <SevenInchBread> Well... it's not finished yet... so I have no clue how it will turn out... until someone writes me some interpreters. :)
18:52:47 -!- mule| has left (?).
18:55:47 <ais523> the problem is that you'll get syntax errors all over the place unless you harmonize all the interpreters. Perhaps only interpreters that don't have syntax errors should be used on any run, so programs can be simple using just 1, or complex using most of them?
18:56:42 <ais523> oh well, that means that you /will/ have to use every interpreter on any run, and so whenever a new interpreter's written it'll invalidate all previous programs written in the language
18:56:45 <lament> by the way a non-esoteric language (Forth) shares the syntax with kallisit.
18:56:59 <lament> (well, forth is fairly esoteric, but it's also fairly mainstream)
18:58:37 <ais523> It's impossible to write a syntax error HOMESPRING (although you can write something that causes the interpreter to generate a temporal paradox), but it's not really suited to how I expect kallisti will end up
18:59:45 <SevenInchBread> there will only be a certain set of interpreters.... it's a callaborative design - not the method in which you program in it.
19:00:54 <SevenInchBread> The basic idea was to create dupdog with a huge number of interpreters instead of two... but I'm too lazy to do that - thus I turned it into a collaboration. :)
19:08:30 <ais523> do it like TMMLPTEALPAITAFNFAL with the list of interpreters changing every day
19:08:44 <ais523> undefined behaviour is very useful in defining specifications for languages
19:10:55 * SevenInchBread is beginning to become weary of all the strictness entailed in computing...
19:13:36 <ais523> maybe we should have a forgiving esolang that can withstand small amounts of mistakes by its authors. It would make a change.
19:13:51 <ais523> s/its authors/people programming it/
19:15:30 <SevenInchBread> ...no one in "serious" programming language design seems to want this behavior... except like, Perl. I think it would be great to experiment with the idea and turn it into a workable idea.
19:16:35 <ais523> ideally it would even fix mistakes in the algorithm you were using
19:17:25 <RodgerTheGreat> the problem is that it would probably make it difficult to do things it "thinks" are a mistake, unless the syntax is very carefully designed
19:17:54 <SevenInchBread> right... one of the issues with Perl I think... the syntax is all weird because of all the corrective behaviors.
19:18:47 <ais523> you could just have some sort of 'I really mean this' delimiter that you could use over small sections of code
19:19:53 <RodgerTheGreat> I prefer the idea of designing syntax to eliminate programmer by either making mistakes more difficult to make or making it easier for the compiler to help the programmer find mistakes before the program can run
19:21:13 <SevenInchBread> eh, I don't care much about the later... but I use a lot of interpreted languages that don't do a lot of nifty compile-time checks.
19:21:36 <SevenInchBread> Just seems natural to me that an error occurs when it's tried...
19:22:33 <ais523> there was an interesting example in a book about compilers I read once; it was the quadratic formula with one closing bracket missing
19:23:00 <ais523> the book said that although it was obvious to a human where it went, it wasn't obvious to a computer, because there were 10 places it could legally be put and lead to a non-syntax-error
19:24:02 <SevenInchBread> hmmmm, that's a good point. Computers have a hard time with ambiguity.
19:24:18 <SevenInchBread> ...unless you give them rudimentary artifical intelligence.
19:25:14 <SevenInchBread> has anyone read about the Elephant language? It's got some weird ideas borrow from human intelligence and natural languages.
19:25:28 <RodgerTheGreat> perhaps set up a learning algorithm that'll pick up a coder's style and use that to guide corrections?
19:31:21 <ais523> it reminds me of TwoDucks, only it's actually theoretically possible to implement
19:33:29 <ais523> and it has a "don't" keyword, just like INTERCAL. Wow, it's almost esoteric!
19:34:01 <ais523> (I suspect it's don't is more practically useful than INTERCAL's (which is redundant except for comments), though).
19:38:26 <ais523> I don't know; IRP can be a pretty efficient language sometimes if you're skilled in writing it
19:39:36 <RodgerTheGreat> as long as you have one or more reliable interpreters, that's very true
19:39:56 <lament> Find the greatest prime number!
19:40:45 <lament> .......................where?...... i don't wanna know
19:41:00 <ais523> lament: there is no greatest prime number. Program complete in ...um... I've lost count.
19:42:03 <lament> Now, find the smallest prime number! :)
19:42:19 <ais523> lament: using one definition it's 2
19:42:38 <ais523> but there was a book I read once called 'Maths made difficult' which debated the usual definition of prime
19:42:56 <ais523> By the definition 'a prime is a number with two factors, itself and 1', the only prime is -1
19:42:57 <lament> apart from the plural "maths"
19:43:24 <ais523> It was a pretty good book; it was sort of maths from an esoteric perspective
19:44:08 <ais523> some of it was highly mathematical (it's the only place I've ever seen the formula for solving cubic equations written out _in full_)
19:44:40 <ais523> but some of it was frivolous (it had in a list of 'Objections to fractions' one based on a misquote from the Bible, and answered with a different misquote from the Bible)
19:45:11 <lament> still the word "maths" is ridiculous :)
19:45:55 <RodgerTheGreat> I generally take any example cited from a fictional work as instantly invalidating a proof.
19:46:00 <ais523> it's a common abbreviation in English as spoken in England, I think. (The North American abbreviation 'math' sounds strange to me.)
19:46:22 <lament> heh @ http://planetmath.org/?op=getobj&from=objects&id=1407
19:46:34 <ais523> RodgerTheGreat: I think what it did was take lots of arguments of different levels of seriousness, and 'refuted' them with the same level of seriousness
19:47:01 <ais523> Another fact from the same book: most people know that if you multiply two numbers, you get their product
19:47:04 <lament> ais523: just because it's common doesn't make it any less ridiculous :)
19:47:27 <ais523> and many people know that the numbers being multiplied are called the multiplier and multiplicand
19:47:44 <ais523> likewise, most people know about numerator and denominator for division
19:47:44 <lament> ais523: because 'mathematics' is a singular noun
19:48:05 <ais523> (they both end with 's', after all)
19:48:19 <ais523> however, in an addition the numbers are tha addend and the augend
19:48:29 <ais523> and in a subtraction the numbers are the minuend and the subtrahend, apparently
19:48:30 <lament> ais523: sure, but it's a strange way to shorten 'mathematics'
19:48:44 <SevenInchBread> >.> it's only strange if you've been using "math" for your whole life.
19:48:56 <ais523> lament: let's not turn IRC in to Usenet here
19:49:02 <lament> ais523: i suppose there's no good way to shorten an inherently bizarre noun (singular noun ending in s)
19:49:42 <ais523> let's shorten it to MM for the time being, then, to avoid the argument
19:49:51 <RodgerTheGreat> well, I always understood it as "mathematics" referring to a wide body of topics, each covering a specific type of "math"
19:49:59 <lament> RodgerTheGreat: well, then you understood it wrong
19:50:05 <lament> RodgerTheGreat: because that's not the etymology.
19:50:26 <SevenInchBread> more importantly.... why does it matter? and why must someone always be right or wrong?
19:50:51 <lament> actually wait, maybe you're right
19:51:15 <SevenInchBread> ...FROM NOW ON... AS IT IS THE NATURE OF OUR CHARACTERS.
19:51:33 <RodgerTheGreat> "google" is derived from a number's name. However, it's verb-based usage has only a tenuous connection with this meaning
19:51:58 <ais523> hey, numbers are functions in some languages (like my work-in-progress Underload)
19:52:14 <ais523> when applied to another number, it returns the argument to the power of the function
19:52:27 <ais523> and when applied to a string, it repeats the string that many times
19:52:42 <ais523> and when applied to a function, it acts like the corresponding Church numeral would
19:57:07 <lament> stupid google makes it hard to search for 'maths'
19:57:19 <lament> it thinks 'math' matches
19:58:39 <lament> no useful hits, though
19:58:57 <lament> ais523: how do brits shorten 'economics'?
19:59:05 <ais523> I just searched for "math" limiting it to pages in .uk domains and found no useful hits either
19:59:28 <ais523> lament: I don't think they do (I know an economist and they say 'economics' in full)
20:00:40 <lament> can't think of any singular nouns ending with s that get shortened....
20:00:46 -!- ais523 has quit ("in case I get caught up in an interesting conversation and miss my bus").
20:16:07 <lament> SevenInchBread: 'mathematics' and correspondingly 'maths' makes little sense to me because i'm not a native English speaker
20:16:28 <lament> SevenInchBread: and it's really bizarre for a singular noun have a plural ending.
20:17:04 <lament> SevenInchBread: it's something like 'mathematica' in all the other languages i'm familiar with.
20:17:25 <lament> apparently it's only plural in english and french
20:17:51 <SevenInchBread> I've never actually heard maths used until today... it sounded odd at first, but then in relation to "mathematics" it made sense.
20:19:29 <SevenInchBread> well... mathematics is, grammatically, singular... even though it looks plural.
20:21:56 <lament> exactly, it's grammaticaly singular, which is why it's weird for it to have a plural ending.
20:22:39 <lament> It's singular in both greek and latin.
20:24:19 <SevenInchBread> mathematics, economics, gymnastics, hueristics, physics, ergonomics, cybernetics, logistics, statistics
20:24:39 <lament> none of those make much sense to me.
20:24:40 <SevenInchBread> ....just a weird convention with... fields of study apparently.
20:25:21 <lament> i basically have to mentally change the final 's' to 'a'
20:27:19 <SevenInchBread> very infrequently does the breaking of a "regular" part of English grammar sound strange to me. :P
20:28:05 <SevenInchBread> ...because they're everywhere - the grammatical rules you follow for a word depends mostly on its origin.
20:29:47 <SevenInchBread> analysis sounds completely singular to me... where mathematics has a sort of semi-plural sound to it.
20:31:38 -!- RodgerTheGreat has quit.
20:32:17 <SevenInchBread> "Mathematics is the art of pedantic rigour" and "The mathematics are of vital importance" both sound fine to me (the second one sounds a bit more formal though)
20:35:09 <lament> SevenInchBread: yes, the grammatical rules do depend on the origin - and the origin is quite singular
20:35:25 <lament> analysis _is_ a singular noun with a singular ending
20:35:31 <lament> the plural would be analyses
20:35:48 <lament> 'is' is a common singular ending
20:37:30 <SevenInchBread> I define grammatically correct as semantically meaningful.
20:38:02 <SevenInchBread> if it conveys an idea - even a bizzare one - the grammatic formation doesn't matter much.
20:39:01 <SevenInchBread> ...even "colorful ideas sleep furiously" conveys an idea...
20:40:17 <SevenInchBread> http://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffalo_buffalo_buffalo_Buffalo_buffalo
20:46:06 <SevenInchBread> http://en.wikipedia.org/wiki/List_of_homophonous_phrases --strangely hilarious
21:12:39 <SevenInchBread> hmm... interesting... anyone ever heard of the Synthesis kernel?
21:12:52 <SevenInchBread> It's an extremly fast OS kernel that makes heavy use of self-modification.
21:36:15 <lament> well, considering you were told about it in this channel...
21:54:17 -!- nazgjunk has quit (Read error: 104 (Connection reset by peer)).
21:54:38 -!- nazgjunk has joined.
22:30:35 -!- jix has quit ("Bitte waehlen Sie eine Beerdigungnachricht").
22:35:01 -!- bsmnt_bot has joined.
22:35:09 -!- bsmnt_bot has quit (Remote closed the connection).
22:36:05 -!- bsmnt_bot has joined.
22:51:16 -!- SevenInchBread has quit (Read error: 104 (Connection reset by peer)).
22:52:50 -!- SevenInchBread_ has joined.
22:53:19 -!- SevenInchBread_ has changed nick to SevenInchBread.
22:57:30 <SevenInchBread> hmmm... I think it makes more sense to represent concatenations as trees... instead of just returning the combined result.
23:10:37 -!- UpTheDownstair has joined.
23:12:46 -!- UpTheDownstair has quit (Client Quit).
23:20:21 -!- sebbu2 has joined.
23:23:09 <bsmntbombdood> goedel's incompleteness theorem proves that humans are super-turing
23:23:13 -!- RodgerTheGreat has joined.
23:25:38 <bsmntbombdood> because axiomatic systems are equivalent to turing machines, and humans can derive results not possible from axiomatic systesm
23:27:33 -!- nazgjunk has quit (Connection timed out).
23:32:58 -!- GregorR has quit (Read error: 104 (Connection reset by peer)).
23:33:31 <lament> bsmntbombdood: what results?
23:34:21 <lament> that's a key element of your proof, unless your proof is supposed to be super-turing too
23:35:12 <lament> "Humans are super-turing, because i said so." - "But that is illogical!" - "That's right. Humans are above logic."
23:38:53 -!- nazgjunk has joined.
23:39:11 -!- sebbu has quit (Connection timed out).