00:40:50 <oklopol> i fast-forwarded the whole conversation as pointless... then at some point found out dupdog does a modulo when outputting making it sensible... but decided not to read again :P
00:44:31 <oerjan> Argh. Wait a minute. The Data.Sequence module cannot be used anyhow.
00:45:02 <oerjan> It stores lengths as Ints, making it useless for massively nested duplications.
00:46:25 * oerjan throws his beginning Dupdog implementation in the garbage bin.
00:46:50 <oerjan> ok, maybe not literally.
00:47:55 <lament> oerjan: that's pretty dumb :(
01:05:51 -!- sebbu has quit ("@+").
01:18:23 <oerjan> Hm, a BSD style license means you can bundle freely, doesn't it?
01:18:51 <oerjan> given the discussion previous today
01:19:08 -!- RodgerTheGreat has quit.
01:20:35 -!- RodgerTheGreat has joined.
01:28:16 <oerjan> I found a downloadable Data.FingerTree module in which you can select any Monoid as your size measure.
01:28:30 <oerjan> (Including the trivial ())
01:56:43 <oerjan> (it's the underlying representation for Data.Sequence)
02:10:20 -!- RodgerTheGreat has quit.
02:12:16 -!- RodgerTheGreat has joined.
03:06:48 -!- oerjan has quit ("leaving").
03:35:56 -!- oerjan has joined.
04:17:42 -!- ShadowHntr has quit ("End of line.").
04:26:05 -!- GreaseMonkey has joined.
04:28:46 -!- oerjan has quit ("leaving").
04:31:39 -!- GreaseMonkey has quit ("eek, pc going awol, back in 5").
04:42:50 -!- RodgerTheGreat has quit.
04:44:00 -!- RodgerTheGreat has joined.
05:13:44 <RodgerTheGreat> /// is interesting, but I'm not sure you could really implement arbitrary looping: http://www.esolangs.org/wiki/Slashes
05:14:52 <RodgerTheGreat> I think you could take a sequence like Code*, and then have "Code" contain instructions to rewrite "*" as "Code*"
05:15:12 <RodgerTheGreat> it'd be interesting to see if this language is turing complete...
05:15:41 <RodgerTheGreat> has anyone ever seen any TC proofs for string rewriting languages?
05:23:14 <RodgerTheGreat> if looping could be implemented, it might be possible to implement some kind of cyclic tag system from within a string rewriting language, proving computational usability... this could be interesting
05:26:59 <RodgerTheGreat> well, goodnight everyone- I might play with /// tomorrow...
05:27:12 -!- RodgerTheGreat has quit.
06:30:16 -!- Sgeo has quit ("Leaving").
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:22:28 -!- Sukoshi has joined.
09:23:07 -!- GreaseMonkey has joined.
09:53:55 -!- Sukoshi has quit ("待っているね、永遠の彼女。素晴しい楽園が二人だけの為に見付かった。忍ばせない。彼女).
10:05:13 -!- sebbu has joined.
10:24:16 -!- nazgjunk has joined.
10:28:54 -!- GreaseMonkey has quit ("Think you're a sore loser? Go to ED. They'll make you look angelic.").
10:50:43 <oklopol> RodgerTheGreat, Thue has a brainfuck interpreter in it, so cyclic tag should be trivial...
10:51:07 <oklopol> might read the logs though /megone->
11:48:53 -!- jix__ has joined.
12:12:10 -!- nazgjunk has quit (Read error: 104 (Connection reset by peer)).
12:12:31 -!- nazgjunk has joined.
12:27:01 -!- jix__ has changed nick to jix.
12:46:04 -!- nazgjunk has quit (Read error: 104 (Connection reset by peer)).
12:46:37 -!- nazgjunk has joined.
14:24:21 -!- nazgjunk has quit (Read error: 104 (Connection reset by peer)).
14:45:11 -!- oerjan has joined.
15:18:44 -!- nazgjunk has joined.
16:26:17 -!- tgwizard has joined.
17:35:54 -!- oerjan has quit ("leaving").
17:45:04 -!- nazgjunk has quit ("Leaving").
18:15:36 <oklopol> hehe reading english grammar, "in oldish english the infinitive can be used as the subject of a sentence" and thinking why did they demote functions to second-class citizens :P
18:16:00 <oklopol> (the grammar was finnish, it didn't say 'oldish')
18:20:45 -!- nazgjunk has joined.
18:28:38 -!- graue has joined.
19:04:20 -!- nazgjunk has quit ("Leaving").
19:50:50 -!- SevenInchBread has joined.
19:51:38 <SevenInchBread> wget -q --read-timeout=0.0 --waitretry=5 --tries=400 --background http://foo.bar/
19:55:55 <graue> wget runs on windows
20:13:45 <SevenInchBread> graue, is it on by default? I can use it from the command prompt
20:18:20 -!- nazgjunk has joined.
20:21:27 <SevenInchBread> DNS is (other than ISPism) the only part of the Internet I can think of that is mandatorially business-ized.
20:22:41 <oklopol> trues, is that what annouy you?
20:22:50 <oklopol> because i dislike the system itself
20:23:56 <SevenInchBread> it's kind of confusing... but I really don't like the commercial aspect.
20:25:33 <SevenInchBread> all the rest of the Internet proves that "you get what you pay for" is false... except this weird internet real-estate system that gives commercial entities more sway than the average socket connection.
20:26:56 <oklopol> mm i don't really care for the practical aspect so... i'll talk about my new language lopoda
20:29:44 <SevenInchBread> the entire basis for registrars is that it ensures you can trust the nameservers you're dealing with... but in all my experience anything on the internet involving money is usually more manipulative than free services.
20:39:44 <lament> shut up with your proposals and explain dupdog
20:39:57 <SevenInchBread> Let's create an experimental computer network... so we can fiddle around with ways to make the existing Internet better, without the worrys of standardization and convention
20:40:12 <SevenInchBread> dupdog is completely useless.... I can't even figure out how to do Hello, World! with it.
20:40:24 <lament> what i need explained is
20:40:45 <lament> you had at one point put up a buggy not-working implementation on the wiki page
20:41:04 <lament> in this implementation, when mfit output characters, it wrapped on 257
20:41:10 <SevenInchBread> yeah, that one didn't work... not sure why... haven't bothered delving further.
20:41:14 <lament> is that a typo, or you did you actually mean it to wrap on 257
20:41:26 <lament> because if it does wrap on 257, i have written hello world for it
20:41:37 <lament> and if it doesn't, it might be impossible to do so
20:42:25 <lament> http://pastebin.ca/397959
20:43:05 <oklopol> if it wraps on 257 there is no 257
20:43:17 <oklopol> 256 is the termination character
20:44:14 <lament> well, they spell hello world
20:44:20 <oklopol> SevenInchBread didn't you make this language? :D
20:44:40 <lament> the problem with esoteric internet is that we'd have to use the existing infrastructure which only understands IP
20:44:52 <lament> so we could only build stuff on top of IP, which is lame
20:45:04 <SevenInchBread> right right... but it would be easy to make a high-level emulation of a low-level internet.
20:45:07 <lament> unless we physically connect our computers ourselves
20:45:21 <oklopol> not really, since it could be easily converted to non-ip dependant
20:45:33 <SevenInchBread> for proof-of-concept more than any practical advantage.
20:45:47 <lament> so something like Tor?
20:46:25 <SevenInchBread> not necessarily... i was thinking more or less as a sandbox for making protocols and systems and testing how awesome they are.
20:46:53 <SevenInchBread> the problem with a huge international network is that a status quo of accepted standards emerges - even if they're not necessarily the best way to do it.
20:48:13 <lament> something tells me that, this being #esoteric, you perhaps don't have the optimal efficiency in mind
20:48:50 <SevenInchBread> ...yeah, don't care how fast it is... unless the point of the experiment is to test faster methods of communication.
20:49:17 <lament> well, all we need for this is protocol specification
20:49:54 <SevenInchBread> A protocol without an implementation is jsut a.... protocol.
20:50:32 <lament> well, a protocol specification + at least one person actually bothering to write an implementation.
20:51:44 <SevenInchBread> communication requires multiple parties that understand the protoocl.... thus, to test out new protocols, you need a network of talking heads that speak the same language.
20:52:55 <SevenInchBread> we can incorporate it into esoOS... when it's first released 20 years from now.
20:54:06 <lament> i have an idea! When you send a packet, you send it to a random address. If the computer at that address is not the recipient, it must pass the packet along to another random address - but only after adding some more stuff to the packet.
20:54:16 <lament> .....okay maybe that's not such a brilliant idea :D
20:55:28 <SevenInchBread> I like IP myself... not sure what you could do differently with it.
20:56:25 <SevenInchBread> maybe "esoteric wide web" is more a descriptive term for what I had in mind.
20:56:35 <lament> we could just use gopher
20:56:45 <lament> it's already extensively implemented, and it's still completely esoteric.
20:57:14 <lament> if you _really_ want to experiment with other low-level protocols, get into amateur radio.
20:58:57 <SevenInchBread> a protocol for defining protocols would be sweet... even though it would basically entail creating an entire programming language and sending it across lines... and it would be the biggest security hazard ever.
21:00:25 <lament> you know what's pretty cool? termcast
21:00:31 <lament> lets other people look at your terminal
21:02:04 <SevenInchBread> one of the common issues I see with networking is that.... while it's easy to talk to the other end, it's hard to create a persistent conversation.
21:03:30 <oklopol> it's all so bad one starts crying the first time they learn about it
21:03:55 <SevenInchBread> I'm not sure what I mean exactly.... I'm kinda thinking specifically of HTTP and other stateless protocols... and dynamic IP addresses. But I don't think getting rid of a dynamic address is such a hot idea.
21:04:48 <lament> it can be done, certainly
21:05:14 <SevenInchBread> but that takes away a lot of the awesomeness of the Internet.
21:05:43 <SevenInchBread> ...mainly, anonymity. It's why the internet is a cloud and not a city.
21:07:39 <lament> your IP is static to a fairly large extent.
21:08:09 <lament> that is, your registrar has some subnet or whatever
21:08:28 <lament> and you get an IP from their, and they have IPs assigned to them
21:08:35 <lament> and there's a central authority that manages it all
21:08:45 <SevenInchBread> ...straying back to programming languages for a second, I've got an idea for a language read concurrently by multiple interpreters with different meanings for different symbols... and a bitwise brainfuck with some basic ideas of quantum entanglement.
21:08:56 <lament> i could get to your ip by just randomly checking all the computers on your subnet
21:09:19 <SevenInchBread> wouldn't take not really... just all possible bit combinations.
21:09:46 <lament> i remember something like that
21:10:25 <SevenInchBread> quantum entanglement with bitwise brainfuck basically makes an event-programming-type thingy.
21:11:01 <lament> haven't i implemented that?
21:11:13 <lament> i even put it in egobot
21:11:50 <lament> no idea what the syntax would be
21:12:28 <lament> http://esolangs.org/wiki/Quantum_brainfuck
21:12:52 <SevenInchBread> ...so..... what's the idea behind the dupdog program? Found any properties that can be exploited?
21:13:17 <lament> yes, the fact that 257 is an odd number :)
21:13:43 <lament> the program is almost trivial, it's a sequence of NOPs and prints
21:13:46 <lament> there's no duplication
21:15:09 <oklopol> just a print every time the program size is the right size mod 257? and one char is cut off every cycle?
21:15:30 <lament> the only trick is that only mfit can print.
21:15:34 <oklopol> hmm, can you explain why 256 would've been bad?
21:15:47 <lament> so if the program is of even size, and mfit prints
21:16:00 <lament> then every time mfit runs, the program will still be of even size
21:16:11 <lament> since one character is removed each step, and new ones are never added
21:16:28 <lament> so if we wrap on an even number (256), mfit can never print a half of all characters
21:16:44 <oklopol> what's mfit do? or is it faster to read it?
21:17:18 <oklopol> link it if you have it open
21:17:29 <oklopol> i never remember the page :D
21:17:40 <SevenInchBread> dupdog was a strange attempt at playing with the semtantic interpretation of syntax.
21:19:22 <SevenInchBread> If I knew it has any sort of properties to it for creating abstractions, I'd add macro substitution into it...
21:21:36 <SevenInchBread> dupdog isn't very conceptually awesome... so I think my next attempt in that area will involving concurrently running interpreters on the same characters.
21:26:08 <oklopol> everybody makes one, SevenInchBread tells us what the interpreters can or can't do
21:26:14 <oklopol> and everyone willing makes one interpreter
21:26:27 <oklopol> and each is assigned the same char simultaneously
21:26:36 <SevenInchBread> SOUNDS LIKE A JOB FOR THE EDUCATED BLIND ESOLANG COMMITEE.
21:29:10 <oklopol> i used to be pretty confident about my english... i seem to know nothing about articles though now that i've read the grammar...
21:30:02 <SevenInchBread> oklopol, that sounds like a corny joke... but I haven't figured out the punch line.
21:31:15 <oklopol> i'm reading english grammar, there is no logic on the 'the' article when it comes to buildings, countries etc
21:31:43 <oklopol> and i thought i'd've known them all... just like that... but most come a suprise for me
21:32:52 -!- nazgjunk has quit ("Leaving").
21:34:09 <oklopol> 6 pages of rules for when to use 'the'
21:34:30 <oklopol> bsmntbombdood you've complained that before
21:34:44 <oklopol> and still i think it deserves it :)
21:37:36 <SevenInchBread> the unfortunate side-effect of living in a entirely-English region is not having any urgent need to learn multiple language.
21:38:27 <lament> oklopol: the articles are very very tricky
21:38:34 <lament> basically the only way to learn them is through use
21:40:12 <oklopol> yeah, i know, what discouraged me was i did so many errors... and i use this language more than my native one :\
21:40:38 <SevenInchBread> ....I know them, but I don't know the rules... using an article incorrectly simply doesn't "sound right".
21:41:47 <lament> that's what fluency is all about
21:43:59 <oklopol> well, i knew most of them... maybe my skillzorz will automatically improve with time
21:44:06 <oklopol> bsmntbombdood what don't you understand?
21:45:47 <oklopol> r <space> <square> <space> <square>R,
21:47:29 <oklopol> if you mean you don't understand how to code in that, i can't help you
21:47:38 <oklopol> because i can't figure it out
21:47:54 <oklopol> i guess that's what you meant
22:00:40 <SevenInchBread> hmmm... studying the morphology of a language sounds like a good way to make the language more intuitive... since it's based on the patterns recognized by fluent speakers.
22:00:53 <oklopol> SevenInchBread, i'm pretty sure you couldn't tell all of the cases of 'the' no matter how native you were
22:01:13 <oklopol> if the book is correct, british bridges aren't 'the' while american bridges are
22:01:34 <oklopol> so... you'd have some memory :D
22:02:32 <oklopol> that'd be Golden Gate Bridge
22:02:41 <oklopol> i was fishing under Golden Gate Bridge
22:03:30 <oklopol> but golden gate is american
22:03:33 <graue> hmm, "london bridge" but "the golden gate bridge"?
22:03:51 <SevenInchBread> I thought you meant British English used a instead of the. :P
22:04:49 <oklopol> graue, you can correct me if the book is wrong
22:05:00 <oklopol> but yes, that's what is says
22:05:51 <oklopol> except for british rivers that carry the name of the river they're built across... they're 'the'
22:06:16 -!- oerjan has joined.
22:07:20 <SevenInchBread> hmmm... being native to United Statesian English... hearing "Let's go fishing for large octupii below FooBar Bridge" does sound paticularly Britishy to me, if not some kind of non-American accent. :P
22:09:13 <SevenInchBread> hmmmm.... yeah, English articles are definetely weird.
22:11:10 <oklopol> sorry, i get some serious hard-on when correcting americans' english
22:12:55 <lament> i'm terrible with decency.
22:13:01 <lament> and half-grammatical with spelling.
22:14:19 <graue> "london bridge" and "the golden gate bridge" sound right to me
22:14:43 <SevenInchBread> English is.... a deformed fusion of Germanic and Romanic rules.
22:14:55 <oklopol> all natural languages suck
22:15:43 <SevenInchBread> I'd kill myself if natural languages took on the consistent design of artifical ones.
22:15:51 <lament> i don't think english has much romance grammar
22:16:42 <SevenInchBread> Most of the romance influence is in some word etymologies
22:17:59 <oklopol> well, a million per cent of english vocabulary is directly from french
22:18:08 <oerjan> i don't see why dupdog with even wrapping cannot work, you can use duplication to turn odd into even and both Mfit and Shanty can do that.
22:18:14 <oklopol> and i believe that's romance
22:21:53 <SevenInchBread> French is Germanic. Scanning over German I see tons of German words that resemble English words... more so than French words.
22:22:52 -!- RodgerTheGreat has joined.
22:23:11 <oklopol> i thought it was romance... what is romance? :P
22:23:18 <oerjan> so English has Germanic grammar with tons of Romance vocabulary, French has Romance grammar with tons of Germanic vocabulary
22:24:57 <SevenInchBread> Romance is Spanish, italian, Romania, Portugese, and....
22:27:07 <SevenInchBread> it decended from Latin.... but I can see much more Germanic influences in French than I can the other Romance languages.
22:28:41 <oerjan> France was rather thoroughly invaded by the Germanic Franks...
22:29:26 <oerjan> Germany and France used to be one kingdom for a while
22:30:30 <oerjan> an Reto-Romansch, Sardinian, probably Corsican
22:31:18 <SevenInchBread> so yeah... Romance grammar with tons of Germanic vocabulary.
22:31:22 <oerjan> (probably mispled some)
22:33:19 <RodgerTheGreat> what spurred a discussion regarding the origin of human languages?
22:33:31 <oklopol> SevenInchBread please don't confuse me like that
22:33:45 <oklopol> my systems get all mized upz
22:34:29 <oerjan> lessee it started with oklopol complaining how he didn't know all the rules for using "the"
22:34:42 -!- nazgjunk has joined.
22:35:47 <RodgerTheGreat> oh, and if we ever try doing the "EsoInternet" thing, the equivalent of IPs or routing codes or whatever should be a TC scripting language
22:37:50 <RodgerTheGreat> like, rather than knowing a remote computer's IP, you obtain a program that can traverse the network and find that specific computer
22:38:12 <RodgerTheGreat> and you need a different tiny program, as simple or complex as the coder makes it, to find each machine
22:38:53 <RodgerTheGreat> and doing tunneling or something TOR-like would just imply a different way of designing these traversal programs
22:40:16 <RodgerTheGreat> and it could be non-trivial to determine wether two programs pointed to the same machine if the scripts can do queries of some kind from each machine they meet along the way. :D
22:41:12 <oklopol> actually, it's a group of languages
22:41:19 <RodgerTheGreat> I'd probably be more interested in making a written language than a spoken one
22:41:48 <oklopol> i never even considered soken
22:41:56 <SevenInchBread> Kind of both... for me. I like unusual pronounciation rules and sounds.
22:42:11 <oklopol> since when designing, it never occured to me langs can be spoken :)
22:42:16 -!- sebbu2 has joined.
22:43:27 <SevenInchBread> I don't like the organizedness of artifical languages though... I'd try to make it purposefully natural by incomporating bits and pieces of other languages.
22:43:30 <oklopol> i was explaining my system to a friend today... he assumed i meant a spoken language
22:43:47 <oklopol> and i was like what the fuck are you saying when he said something about pronunciation
22:44:33 <RodgerTheGreat> I'd like to design a language that focussed on brevity and syntactic conciseness. Most human languages contain a great deal of redundant information.
22:44:52 <SevenInchBread> like, create an entire alternate history with multiple languages... tracking historical events to create natural-like influences and evolution.
22:45:27 <oklopol> the system has some interesting ideas, there are no nouns in the core language for one, you have noun packs, which you can master and they can then be used in the conversation if both speakers know them
22:45:36 * SevenInchBread likes the ambiguities... lots of words, with lots of ways to combine them
22:45:58 <RodgerTheGreat> and very simple glyphs that can be combined and superimposed, with intuitive modifications of expressed meaning
22:46:22 <SevenInchBread> yep... sounds like a lot of artifical languages... very modular and consistent.
22:47:27 <oklopol> well, that was just one thing... i'm not gonna share the whole spesification on an irc chan, and it's not done yet
22:47:44 <RodgerTheGreat> another interesting idea would be a written language with self-repairing syntax. Design it so that you can be missing half of the characters and reconstruct the whole meaning, without being explicitly redundant
22:47:52 <oklopol> the core language, common, is pretty much just a way to express logic
22:48:22 <RodgerTheGreat> so, you have the "control structure", and you're still working on the "command set"
22:48:43 <oklopol> yeah, that is already done implicitly in normal langs, the redundancy you mentioned
22:49:02 <SevenInchBread> hmm... to a certain extent you can do that informally with English and other natural languages.
22:49:16 <oklopol> sometimes you can understand a sentece with only the first few words of it, sometimes a sentence has multiple semantics
22:49:37 <oklopol> i just have guidelines i'm going to follow
22:49:56 <oklopol> there aren't any sequences of characters in the language
22:50:15 <oklopol> you can use whatever you want, probably words of your own previous native tongue
22:50:32 <oklopol> because the presentation of nouns is not important
22:50:37 <SevenInchBread> Humor, is highly linguistical. Good prose is often concisely ambiguous
22:50:41 <RodgerTheGreat> if you designed an entire language in a consistent manner, you could achieve a very nice human/machine readable manner for expressing ideas. Most programming languages are unsuitable because programming languages are designed for expressing algorithms
22:50:56 <oklopol> they are just referenced by indeices in the noun pack
22:51:38 <SevenInchBread> ...humor would be nigh impossible if every word and morphological structure had one clear meaning.
22:51:41 <oklopol> you convert them to whatever char sequences you want
22:52:05 <oklopol> SevenInchBread, i've never enjoyed wordplays that much
22:54:09 <oklopol> RodgerTheGreat, most comp langs are for algorithms, native langs are more like regexes, they define a structure, not a sequence or a clear action
22:55:38 <oklopol> i have so much to say i think i won't say anything anymore
22:56:02 <oklopol> i hate it when i think about something and find a conversation about it later
22:56:11 <oklopol> and all the info bangs into my head
22:56:43 <SevenInchBread> not always.... I'd relate natural languages more to... communication protocols. regexes express a definite pattern, algorithms define a definite sequence of actions... communication protocols send desired characteristics to a receiver, and leave the task of interpretation up to it.
22:57:38 <SevenInchBread> a language that says everything you want it to say and is interpreted exactly as it was intended is... boring.
22:58:28 <oklopol> SevenInchBread sending stuff has nothing to do with natural language
22:58:31 <RodgerTheGreat> I was wondering if this could be TC: http://www.esolangs.org/wiki/Slashes
22:58:39 <oklopol> you can just as well send code to a remote server
22:58:45 <SevenInchBread> oklopol, communication has everything to do with natural language.
22:58:51 <RodgerTheGreat> has anyone seen any TC proofs for string rewriting languages?
22:59:00 <oklopol> RodgerTheGreat thue has bf
22:59:59 <oklopol> you could as well say pressing buttons is somehow an intrinsic part of programming languages
23:00:02 <RodgerTheGreat> looping is a bit of a trick, but I think I can see a way to pull it off
23:00:43 <oklopol> and you can say bugs are the same as misinterpreting natural language
23:00:46 <SevenInchBread> what other purpose does a spoken word have than to be heard and understood.... what other purpose does a book have than to be read?
23:00:49 <oklopol> but it's not a critical part
23:01:30 <oklopol> i know exactly what the critical part (difference between natural and comp langs) so... my mind is made
23:01:52 <oklopol> what other purpose does code have than to be tokenized and run?
23:01:55 <RodgerTheGreat> I'm trying to think of a nontrivial task that cannot be bruteforced and requires no input, which would demonstrate the plausibility of /// being TC. Think a cyclic tag system would be possible?
23:03:59 -!- nazgjunk has quit ("gah sleep").
23:04:35 <oerjan> you could try the Ackermann function
23:04:43 <oklopol> it seems you can't duplicate anything
23:04:52 <oklopol> something is taken off every run, right?
23:05:09 <oklopol> i might've gotten it wrong : )
23:05:41 <RodgerTheGreat> if you use a two-step copy you could duplicate your main code over and over, simulating a loop
23:06:25 <SevenInchBread> if it's based off of string replacement... it's hard not to emulate thue.
23:07:02 <oerjan> it's self-modifying unlike Thue
23:07:25 <oklopol> can you show me an eternal loop?
23:08:41 <oerjan> Copying / and \ needs some cleverness.
23:08:52 <SevenInchBread> ....should probably be TC then... sounds like Thue - except for some minor rules about order of replacement
23:08:53 <RodgerTheGreat> if you do the copy all in one step, you get an infinite loop
23:09:12 -!- sebbu has quit (Connection timed out).
23:09:25 <oklopol> "and again escaped characters are treated as themselves."
23:09:39 <oklopol> why can't he just say \ escapes, use common sence :)
23:10:51 <SevenInchBread> It's very similar to Thue... just a bit hard to read the specification.
23:11:51 <RodgerTheGreat> you could halt in my example above by doing something to get rid of the % before you copy your main body again
23:13:19 <SevenInchBread> It's basically self-modifying Thue. Where the replacement rules are embedded in the input.
23:13:41 <RodgerTheGreat> so, a simple way of explaining the language is, /A/B/ replaces each instance of A with B, and \ works as an escape character
23:14:21 <RodgerTheGreat> and everything encountered outside /// is just printed
23:15:46 <SevenInchBread> yep... it looks like the writer tried to algorithmically reproduce what the interpreter does in English.
23:16:20 <SevenInchBread> Converting an algorithm into English results in a terrible explaination.
23:16:24 <oklopol> '/a/ \/a\/b\/ \/a\/b\/ /a'
23:16:51 <oerjan> well, the interpreter given was written after the language description (by me)
23:18:18 <RodgerTheGreat> oklopol: I think if you had "b/b/" it would just print out that string- it represents less than a complete /// expression, so that's where it halts
23:18:59 <oklopol> efefefefe /a/b/ c would evaluate /a/b/ though?
23:19:54 <oerjan> you forget that the substitution is repeated until the result no longer contains the source string.
23:20:45 <oerjan> so you cannot replace something immediately with something that contains it
23:21:08 -!- nazgjunk has joined.
23:21:22 <RodgerTheGreat> that's why I suggest the "two-step" process for duplicating source code
23:21:26 <oklopol> so /a/a/a is an infinite loop?
23:21:42 <SevenInchBread> so... you can do an infinite loop... but so far there's no way discovered to anything else while looping infinitely. :)
23:22:08 <RodgerTheGreat> no, you could easily do other things in the main code body
23:22:15 <oerjan> indeed those hardly count
23:23:42 <oerjan> as i said, copying / and \ will take some cleverness.
23:25:54 <oerjan> probably coding them as something else, and substituting them back. but while preserving the "something else" somehow to repeat the process.
23:26:41 <RodgerTheGreat> well, the 99bob example has some things in common with most quine designs
23:26:45 <oerjan> well you would expect to need a quine-like approach.
23:27:15 <RodgerTheGreat> 99bob is currently the least trivial code example we have on our hands
23:27:53 <RodgerTheGreat> perhaps we should work with the input extension and see if we can create a cat program as a proof-of concept of looping
23:29:45 <oerjan> a simpler idea: try to write a program that actually outputs something repeatedly.
23:30:57 <oerjan> i would expect the greatest problem is to get any nontrivial loop at all, so that should be enough for proof of concept.
23:31:47 <RodgerTheGreat> if we get a loop working, what else would we need for a TC proof?
23:32:49 <SevenInchBread> RodgerTheGreat, I like the double-step thing to... the main problem with effectively using substitution to representTC is that it's too eager.
23:34:05 <RodgerTheGreat> yeah- the main trick is limiting the scope of substitutions, which I think would depend on leaving a bunch of "flags" and things to create less aggressive replacements
23:34:19 <oerjan> i am suddenly wondering whether there is any way at all to remove a string from the program and then bring it back, repeatedly
23:35:04 <RodgerTheGreat> like replace / with # and \ with ? and store everything as a literal perhaps
23:35:19 <oerjan> but how can you avoid destroying all # and ? on the first iteration?
23:36:11 <RodgerTheGreat> then replaces won't touch it, but you can replicate it at will
23:37:09 -!- nazgjunk has quit ("tonight, we sleep in HELL!").
23:38:52 <oerjan> P -> aP somehow, for example
23:40:49 <oerjan> if a could contain escapes too, we would have achieved even more
23:41:30 <oerjan> but without would be a start
23:43:42 <SevenInchBread> with a sufficiently confounded syntax... you need no ponder escape characters.... at least practically.
23:44:09 <oklopol> what would '/ aa / \/a\a\/a\a\/a / aaaaa' become?
23:44:33 <oerjan> are the spaces included?
23:46:25 <oerjan> the substitution is aa -> /aa/aa/a so it will loop uselessly
23:57:54 -!- jix__ has joined.
23:58:37 -!- sebbu2 has quit ("@+").
23:58:53 <oklopol> meh, i'm so bad at making quines...