00:53:16 -!- calamari has joined.
01:15:25 -!- calamari has left (?).
01:25:40 -!- Sgeo has joined.
01:26:01 -!- sebbu has quit (Success).
01:57:57 -!- sarehu has joined.
02:17:13 -!- sarehu has quit (Read error: 110 (Connection timed out)).
02:50:45 -!- ehird` has quit ("K-Lined by peer").
03:13:17 -!- GreaseMonkey has joined.
03:15:47 * Sgeo is a beneficial mutation occuring in your lab.
03:19:12 <Slereah> Preparing THE HOLY WAR ON SCIENTOLOGY
03:21:10 <pikhq> Slereah: Would said mask bear any relation to Guy Fawkes? :p
03:21:17 <Sgeo> I'm going to eat then work on PSOX
03:21:51 <pikhq> Sgeo: BTW, for PSOX in PEBBLE2: you're in luck. I'm going to have an actual *type system* in there, I believe.
03:22:24 <Slereah> Although it will arrive in time for the second protest, in March.
03:22:43 <Sgeo> pikhq, did you note my changes to the File Descriptor system?
03:23:05 <Sgeo> With that, I need to eat
03:24:55 <RodgerTheGreat> I'm really in the mood to draw something. Any requests/ideas?
03:25:21 <pikhq> http://pastebin.ca/895846
03:25:30 <pikhq> I hope that, when I'm done, that will compile.
03:29:04 <pikhq> RodgerTheGreat: Thoughts on my idea?
03:29:28 <pikhq> Anyways: for you to draw: more of that series with the robotic revolution.
03:29:41 <pikhq> Actually, you should probably stick it on your nonlogic page, so that it's easier to find.
03:29:51 <Sgeo> pikhq, http://trac2.assembla.com/psox/changeset?new=trunk%2Fspec%2Fpsox.txt%4053&old=trunk%2Fspec%2Fpsox.txt%4038
03:29:58 <pikhq> (I know, I know, I could grep nonlogic dump and #esoteric logs. . . But I don't wanna)
03:30:10 <pikhq> Sgeo: Comments on my PEBBLE2 example?
03:30:29 <Sgeo> I don't understand PEBBLE2
03:30:49 <pikhq> What don't you understand?
03:30:49 <Sgeo> Oh, is that a proposed thing?
03:31:16 <pikhq> I've not implemented it; I've mostly been coming up with code examples while thinking of implementation.
03:31:39 <RodgerTheGreat> it looks like a pretty ambitious attempt at a macrolanguage
03:32:33 <pikhq> Which I think is appropriate for PEBBLE2.
03:32:44 <pikhq> The only hard part is going to be that expr command, I suspect.
03:33:26 <RodgerTheGreat> going with infix notation makes your job quite a bit more complex
03:34:10 <RodgerTheGreat> especially if you want to self-host- the only clean ways to do infix require converting to RPN or using recursion, and recursion does not lend itself well to BF
03:36:27 <pikhq> I'm not going to self-host.
03:36:44 <pikhq> How I'm thinking of doing the type-system and the allocation of variables does not lend itself to it.
03:37:17 <pikhq> I can either make it self-host, and have it be basically PEBBLE 1 done a bit more rigourously, or make it t3h awesome, and not have it self-host.
03:37:43 <pikhq> (I *really* don't want to do garbage collection from within BF)
03:38:00 <pikhq> Although making it RPN or Polish may help me out a bit.
03:38:27 <pikhq> Probably Polish because, after all, I already have commands like ":= $i 5" and "+= $i 5". . .
03:38:44 <pikhq> Making the expression parser handle those commands in the same way is merely consistent.
03:41:23 <pikhq> PEBBLE2 should end up making all PEBBLE code a hell of a lot cleaner as a result.
03:42:05 <pikhq> I'm really looking forward to making PFUCK about as trivial to write as a C version would be. :p
03:53:22 -!- immibis has joined.
04:50:17 <RodgerTheGreat> alright, I'm drawing these really fast, but I think you guys will still enjoy them
05:01:12 <pikhq> Would you class having a copy of your GPG revocation certificate and a secure way of encrypting your data as paranoid?
05:02:17 <pikhq> (I do neither ATM: I just have the means to)
05:03:24 * pikhq encodes his grocery list with tax records as the one-time pad. :p
05:05:49 <pikhq> Not really. I set all that up in half an hour.
05:06:18 <pikhq> . . . Perhaps a bit of free time.
05:06:56 -!- immibis has quit (Read error: 110 (Connection timed out)).
05:07:23 <pikhq> Of course, if/when I finish PEBBLE2, then we can say that it's a truly insane amount of free time used. :p
05:07:35 <pikhq> Hell, the same applies to PEBBLE 1.0. . .
05:07:42 <pikhq> I *did* spend the better part of a year on that.
05:07:53 <RodgerTheGreat> you're talking to the guy that spent a good chunk of 8 months writing a Java-based Pokemon clone.
05:08:17 <pikhq> You're talking to a guy who's been tempted to do the same in a different language.
05:08:39 <pikhq> I'd say we're about equal on the free time front.
05:09:11 <pikhq> I have yet to enter college, so. . .
05:11:03 <pikhq> You'd be amazed at how much work teachers think they can assign just because it's an 'honors class'.
05:11:17 <RodgerTheGreat> college devours your free time like no other. I was only able to work on CRPG during my vacation time- I would spend the first few days lying around, playing with my cats and watching TV, recharging. Then, I would code like a madman, laying down the ideas that had been splintering away in the back of my head for months.
05:11:34 <pikhq> When I have an actual college class (calc II) that has less homework than high school classes, you know something's up.
05:11:59 <pikhq> (either my teachers are insane, or that college sucks. :p)
05:13:53 <Sgeo> Can I give this line to a computer-illiterate person?
05:13:54 <Sgeo> "Downloading files SHOULD be safe, provided you do NOT open or run them."
05:14:10 <Sgeo> She stopped opening attachments and reading emails from non-@farmingdale.edu
05:14:10 <pikhq> Sgeo: Ask one of them.
05:14:41 <Sgeo> I'm trying to help tell this person what is and isn't safe..
05:14:51 <Sgeo> Are PDFs safe? >.>
05:15:20 <RodgerTheGreat> they do, interestingly, encapsulate several TC languages
05:15:21 <Sgeo> "application/pdf == Safe for Abode Reader to read, bad for Acrobat 4.0 and 5.0 (I think)"
05:15:27 <pikhq> PDFs are little more than Postscript that feels fairly uppity.
05:15:38 <Sgeo> "image/jpeg == Once problematic, should be safe if Windows is kept up-to-date"
05:15:50 <Sgeo> "application/msword == MS Word document, safe if your macro security
05:15:50 <Sgeo> settings are safe"
05:15:58 <Sgeo> "application/x-msdos-program == BAD BAD BAD"
05:16:31 <pikhq> Oh, yeah. . . I forget that this whole 'Microsoft' thing seems to think that executing arbitrary executable code in arbitrary files (even non-executable ones) is a good idea. . .
05:19:38 <Sgeo> .PIFs would be application/?
05:21:41 <pikhq> appliccation/x-pif, I think.
05:24:54 <Sgeo> "application/(anything else) == I can't predict every attachment, probably not safe"
06:11:37 -!- Sgeo has quit (Remote closed the connection).
06:55:07 <RodgerTheGreat> two more pages: http://rodger.nonlogic.org/images/axon/
07:17:04 -!- oerjan has joined.
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:49:24 -!- GreaseMonkey has quit ("Hi Im a qit msg virus. Pls rplce ur old qit msg wit tis 1 & hlp me tk ovr th wrld of IRC. and dlte ur files. thx.").
09:29:56 -!- oerjan has quit ("leaving").
10:03:16 -!- sebbu has joined.
10:24:33 -!- puzzlet has quit (Remote closed the connection).
10:24:41 -!- puzzlet has joined.
10:57:02 -!- puzzlet has quit (Remote closed the connection).
10:57:10 -!- puzzlet has joined.
11:11:14 <SimonRC> http://www.enterprise-architecture.info/Images/Documents/symeph%5B1%5D.pdf
11:11:15 <SimonRC> I particularly recommjend page 11
11:11:53 -!- sebbu2 has joined.
11:20:09 <SimonRC> in fact, their whole website looks a bit TDWTF-worthy
11:27:53 -!- sebbu has quit (Read error: 110 (Connection timed out)).
12:16:22 -!- helios24 has joined.
12:23:36 -!- Slereah has quit (Remote closed the connection).
12:32:23 -!- Slereah has joined.
13:24:03 <oklopol> expr isn't that hard, infix -> rpn is trivial, and rpn -> brainfuck is not that hard
13:30:51 -!- jix has joined.
13:37:17 -!- helios24 has quit ("Leaving").
14:43:01 -!- RedDak has joined.
15:07:16 -!- timotiis has joined.
15:12:11 -!- cepera has joined.
15:19:54 -!- cepera has quit ("freenode.net").
15:31:50 -!- ais523 has joined.
16:03:04 -!- Slereah has quit (Read error: 104 (Connection reset by peer)).
16:03:43 -!- Slereah has joined.
16:20:07 -!- Sgeo has joined.
16:25:08 <Sgeo> I might release an alpha version of the PSOX ref. interp. soon!
16:25:57 <Sgeo> Anyone other than me excited?
16:26:08 <Slereah> Depends, you've got some drugs?
16:40:44 <Sgeo> I decided that a lot of parsing should occur in @argtypes
16:48:15 <Sgeo> Itsy Bitsy Teenie Weenie Yellow Polka-dot Clarification
16:48:44 <Sgeo> ^^ used that as a commit message
16:49:15 * ais523 performs the sort of action which usually results in coffee spilling over a computer keyboard, but doesn't due to an empty mouth
16:59:42 <Sgeo> I'm too obsessed with that now to implement PSOX :/
17:06:24 * ais523 uses single-character smilies
17:09:41 <Sgeo> Alpha version will implement the pseudodomains and domain2
17:24:00 -!- RedDak has quit (Remote closed the connection).
17:24:52 * Sgeo feels like everything's close to completion
17:46:49 -!- RedDak has joined.
17:48:22 -!- Dagide has joined.
17:48:53 -!- RedDak has quit (Client Quit).
17:49:03 -!- Dagide has quit (Read error: 104 (Connection reset by peer)).
17:49:19 -!- RedDak has joined.
17:49:31 -!- Hiato has joined.
17:50:34 <Hiato> alright, I retract my previous statement and insert an "Hello all :)"
17:53:11 * ais523 always writes Hello, world! with the comma
17:54:11 * Hiato realises that this is some kind of string handling syntax as of yet unknown to him
17:54:29 <ais523> it's Perl and sed syntax for substituting regexps in strings
17:54:41 <Slereah> s/A/B = substitute A for B
17:54:41 <ais523> s/a/b/ changes the first a in some string to a b
17:54:48 <ais523> s/a/b/g changes all as to bs
17:54:59 <Hiato> what is the g for then?
17:55:02 <ais523> on IRC, the string that's operated on normally has to be determined from context
17:55:08 <Slereah> It's the same as *B, but * requires the context to be obvious.
17:55:10 <ais523> without the g, it only makes one change
17:55:57 <ais523> "This sentence contains many nonsklarkish English flutzpahs, but the overall plugganzip can be glorked from context"
18:01:59 <Sgeo> Hiato, excited about a possible alpha release of PSOX today?
18:02:31 <Hiato> very much so, I was not aware of this aforementioned release, but know I am ;)
18:03:55 <Sgeo> Do custom domains need to be working for this release?
18:05:54 <Hiato> that is you decision entirely, it would be nice, but hey ;) Also, now is supper, so I'll be back after I've my fill..
18:07:01 -!- jix has quit ("CommandQ").
18:11:46 -!- oerjan has joined.
18:13:55 <Sgeo> Might be an alpha release of PSOX.py soon!
18:16:22 <Sgeo> It will be just pseudodomains and domain2(System)
18:26:21 -!- oerjan has set topic: fix $ (++) "THE TOPIC IS NOT " . show.
18:28:07 <Hiato> ok, I'm back. Sgeo: any news?
18:28:33 <Sgeo> Just pseudodomains and domain2(System), not custom domains yet
18:28:42 <Sgeo> and you'll get it via SVN maybe
18:30:15 -!- helios24 has joined.
18:30:47 -!- olsner has quit (Read error: 110 (Connection timed out)).
18:33:15 <Sgeo> Currently, I'm writing a lot of code w/o testing :/
18:34:10 <Hiato> I would offer to test, but then again, you don't want an idiot smashing randomly at your code ;)
18:35:31 <Sgeo> I'll do testing, and hopefully any and all bugs will make themselves obvious
18:42:12 <Sgeo> Testing now, trying to fix a bug
18:42:16 <Sgeo> Don't know what's causing it
18:44:15 <ais523> what effects does the bug have?
18:44:32 <Sgeo> Nothing is happening
18:44:47 <Sgeo> When I do Ctrl-C, I find it's at the point where it retrieves a line from the client
18:45:42 -!- olsner has joined.
18:45:46 -!- olsner has quit (Read error: 104 (Connection reset by peer)).
18:45:53 <Sgeo> Actually, that's in the loop, so maybe I should throw some prints in
18:47:32 -!- olsner has joined.
18:47:43 <Sgeo> It's not getting past that line at all
18:47:58 <Sgeo> Maybe I should test with a Python program not a BF one
18:52:24 <Sgeo> That's a strange bug, it's misreading the version number..
18:52:28 <Sgeo> But that's a separate issue
18:52:57 <Sgeo> Oh, I commented something out without knowing what it does
18:53:28 <Sgeo> Ok, fixing that line helps
18:56:54 <Sgeo> Oh, it's not working with the BF file at all
18:57:03 <Sgeo> It's like the BF interp never flushes its stdout
18:57:16 <Sgeo> In the Python file, I'm flushing stdout manually
19:00:55 -!- n0nsense has joined.
19:02:42 <oerjan> no n0nsense here please. we should all speak perfectly cromulently.
19:03:15 <n0nsense> I wrote a bot in Tcl with commands to evaluate brainfuck^^
19:10:25 <lament> so much useless software in this world.
19:11:25 <Sgeo> Is PSOX useless?
19:11:42 <RodgerTheGreat> well, there are two ways software can be useful- useful to the world, and useful to the maker. Even if nobody else is ever going to see the code, it can still be a good learning experience.
19:11:49 * ais523 saw a suggestion once that it would be possible to perform parallel computing by sending specially-crafted HTTP requests to a huge number of computers in such a way that whether they were accepted/rejected/bounced/replied-to/whatever performed the computation
19:12:22 <ais523> it would be great if there were some way to write a massively parallel BF interp like that, but unfortunately I doubt it's possible
19:12:33 <ais523> and besides, you'd probably get sued by people for wasting their bandwidth
19:13:34 -!- RedDak has quit (Remote closed the connection).
19:14:12 <RodgerTheGreat> it's a beautiful thing to realize that almost any environment can be used to perform computation
19:14:58 <ais523> before I discovered esolangs, a related pursuit I used to try was to see what environments could be programmed
19:15:10 <ais523> I used to test them with a simple noughts-and-crosses program that never lost if going first
19:15:31 <ais523> that program is easy to write because you can write it as four independent threads if necessary that never need to interact
19:15:46 <ais523> so not a demonstration of TCness or even computational usability, but still impressive
19:15:52 <lament> never need to interact?
19:16:02 <lament> they don't share resources?
19:16:13 * ais523 was talking in a theoretical sense
19:16:16 <RodgerTheGreat> I'd assume they share input events and the ability to draw to the screen
19:16:18 <ais523> you can program it in Paint, for instance
19:16:33 <ais523> by using two slightly different colours and bucket-fill as the input mechanism
19:16:44 -!- n0nsense has left (?).
19:17:02 <ais523> i.e. clicking on any particular location for your move causes the appropriate move to be echoed by Paint because they're connected by a thin line in a different colour
19:17:18 <ais523> and the point is that there are four move-pairs that can be triggered in any order
19:17:44 -!- danopia`school has joined.
19:19:58 <ais523> http://filebin.ca/tesmth/OANDX.BMP
19:20:25 <ais523> that's it with two shades of blue which are hopefully different enough for you to see how it works
19:20:45 <ais523> hmm... maybe I should make a PNG version. It's years since I last looked at that file
19:22:51 <ais523> looking at the dir it's in, it seems I have a winhlp32 and Powerpoint version as well
19:23:02 <ais523> or maybe the original winhelp
19:23:13 <ais523> because I used to use Windows 3.1 around that time
19:23:30 <Hiato> woow, asi523 - that's ingenious!!
19:23:37 <Hiato> beautifully thought out :D
19:23:46 <Sgeo> ..What looked like a bug was me forgetting how the input function worked...
19:24:56 -!- danopia has quit (Connection timed out).
19:25:08 <Hiato> would I be correct in saying that it is not possible to win?
19:25:39 <ais523> in fact that's why that algorithm was chosen
19:25:47 <ais523> simple, yet impresses people who don't understand programming
19:26:07 <Hiato> indeed, even for those who think they do ;)
19:27:04 <Sgeo> What are we talking about?
19:27:17 <ais523> an image I mentioned about 7 minutes ago
19:27:32 <ais523> that gives the illusion of being a noughts-and-crosses program for Paint
19:28:26 <Hiato> hrm... I'm thinking of messing with this to implement a draw and lose sign (same construction technique I think) - though I'm not sure it's possible because paint is a definite overwrite procedure, if's and dependencies are not really possible.... perhaps lighting up lose/draw letter by letter may work, but then again, there is no way to say for sure what order the squares where played in...
19:29:22 <ais523> I thought about that too, but never seriously tried
19:29:32 <RodgerTheGreat> you can think of the floodfill as a much more limited version of wireworld, in a sense
19:30:07 <lament> photoshop fill, which fills similar colors (up to a given threshold) as well as exact match, can be used to implement logic gates
19:30:22 <Hiato> yeah, that is an awesome idea :D
19:30:24 <RodgerTheGreat> Hiato: and you can do a form of sequencing logic- lemme see if I can come up with an example...
19:30:49 <Hiato> I'd like to see this
19:33:22 <Sgeo> Well, for some reason it's skipping some input..
19:33:38 -!- Tritonio_ has joined.
19:33:40 <Sgeo> Maybe I should see what the server is sending
19:35:10 <ais523> hmm... does that need two fill colours for user input to work?
19:35:14 <RodgerTheGreat> if all four cyan paths have been filled with black, the pathway from bottom to top can now be filled with a different color
19:35:18 <Sgeo> The input command is sending stuff all at once
19:35:27 <ais523> however, you can assign them to left-click and right-click
19:35:48 <RodgerTheGreat> and I'm pretty sure it's impossible to do anything complex with a single color
19:36:09 <Hiato> damn, I see know, that is actually very very clever
19:36:24 <RodgerTheGreat> but the above is essentially an arbitrary-input AND gate. :)
19:36:31 * Hiato goes to work on the lose sign
19:37:04 <RodgerTheGreat> and OR gates are naturally trivial... so you can do combinational logic... holy shit
19:37:28 <Hiato> which can be overcome
19:37:31 <ais523> and the fact that changing one input can affect another input
19:37:39 <Sgeo> That's because when I was trying to fix things, I changed the client a bit
19:37:43 <Hiato> vy using as many layers as necessary with different colours
19:37:44 <ais523> whereas in most programming languages changing an input only affects the outputs
19:37:56 <Hiato> and spacing in between
19:38:16 <RodgerTheGreat> ais523: I think it should be possible to overcome that with multiple colors
19:38:48 <Hiato> pretty much what I said :P
19:40:13 <Sgeo> The cat program WORKS!
19:40:56 <ais523> cat from file to file, presumably
19:40:56 <lament> with multiple colors, you can certainly have logic gates, but it's too much work for the user
19:40:59 <Sgeo> Actually, the Python client that is a client that does cat is
19:41:01 <ais523> otherwise PSOX isn't really needed
19:41:08 <Sgeo> ais523, I'm just testing
19:41:22 <Sgeo> But this client in Python is the first PSOX program ever run successfully
19:41:35 <Hiato> hrmm... this requires many and gates
19:41:42 <Sgeo> Don't know why BF isn't working
19:42:20 <Hiato> running into wire crossing :(
19:43:35 <Sgeo> Another BF program is sending stuff, but this one isn't
19:45:48 <RodgerTheGreat> lament: I'm not sure- I think it might be possible to do everything with sort of a "two-color logic"
19:45:57 <lament> still too much work for the user
19:46:24 <Hiato> hrmm, I'm still on the gates, perfecting and working :P
19:47:22 <ais523> incidentally, this reminds be of another esolang I was working on
19:47:47 <ais523> what I can remember is that it was a cellular automaton in 2D
19:48:08 <ais523> that cells only ever changed to be brighter colours, never darker
19:48:17 <ais523> and that http://pastebin.ca/896573 is a TC proof by emulating rule 110
19:48:24 * oerjan is reminded that 2-coloring is easy while 3-coloring is NP-complete
19:48:25 <ais523> but I've completely forgotten how the lang worked and never documented it
19:48:28 <Hiato> interesting, what happens after white
19:48:42 <ais523> Hiato: then the cell can't change at all
19:48:49 <ais523> in fact, most cells never change more than once, if ever
19:48:54 <Hiato> I see, a shame this wanted documented
19:48:58 <ais523> it was an experiment in irreversible programming
19:49:04 <Hiato> I see... sounds interesting
19:49:13 <Sgeo> The client isn't sending stuff!
19:49:19 <Sgeo> Why isn't the client sending stuff!
19:49:23 <ais523> however, the only way to recover the spec now is likely to reverse-engineer it from the one surviving program
19:49:34 <oerjan> Sgeo: block buffering perhaps?
19:49:57 <ais523> are you flushing the output after every write?
19:50:13 <Sgeo> Maybe the person who wrote the bf interp is an idiot and . sets the current cell to 0
19:50:16 <ais523> if you aren't, pipes between programs sometimes will store up a lot of data and then output it all at once, for efficiency reasons
19:50:27 <oerjan> a file can be unbuffered, line buffered or block buffered. with the last output is only sent when the buffer is full.
19:50:34 <Sgeo> ais523, can't force client to flush
19:50:38 <Sgeo> Except with newlines
19:50:41 <ais523> oerjan: or there's an EOF signal
19:51:14 <oerjan> and only line buffering implies flushing on newlines
19:51:36 <ais523> on Unices a single Control-D will flush standard input if there's at least one unflushed character in it
19:51:46 <ais523> or end-of-file if all characters have been flushed
19:51:53 <ais523> (and, of course, the terminal controls stdin)
19:54:05 <Sgeo> How do I make the bf interp flush on newlines?
19:54:21 <ais523> what lang is it written in?
19:54:38 <Sgeo> I don't know, it's the interp that Ubuntu has
19:54:51 <ais523> Ubuntu has two bf interps
19:55:01 <Sgeo> the one done with command `bf`
19:55:44 * ais523 downloads its source code
19:56:18 <ais523> it autoflushes after every character
19:56:24 <ais523> so flushing isn't the problem
19:56:49 <ais523> line 294 of bf.c, if you're interested
19:58:13 <RodgerTheGreat> Hiato: counters are also really easy: http://nonlogic.org/dump/images/1202500500-count.png
19:58:54 <ais523> RodgerTheGreat: using cyan/black alternation as the input?
19:59:11 <RodgerTheGreat> I think a cyan/black/cyan/black "pumping logic" could be used to create arbitrarily complex machines. The real challenge will be trying to come up with self-resetting gates
19:59:13 <ais523> I seem to remember discovering that pattern as well, although I never used it for anything useful
19:59:31 <ais523> a self-resetting gate is impossible
19:59:31 <Hiato> is that functional?
19:59:47 <ais523> because once two sections of colour have fused together by becoming the same colour, there's no way to reverse it
19:59:58 <ais523> and without fusing colour sections there's no obvious way to do logic
20:00:02 <Hiato> PS: currently I have a mismatched, but functional winning device that works for the following config:
20:00:19 <Hiato> yes, I must agree with asi523 here...
20:00:56 <RodgerTheGreat> and anything more complex than two-color logic would be unfeasible from a user perspective
20:01:16 <Hiato> that's the exact problem I am having
20:01:29 <Hiato> the winning counter works once, but only once
20:01:40 <Hiato> without user intervention, I cannot reset it.
20:01:52 <ais523> Hiato: just tell them to return-to-save at the end of a game
20:02:20 <Hiato> then thing is: It isn't even a correct way to win
20:02:36 <Hiato> my rules are: if you occupy the top right spaces, you win (right angle in a row)
20:03:01 <Hiato> yeah, I was. This is for the "CPU"
20:03:21 <Hiato> would we call this emergent behaviour? (Out of interest)
20:04:02 * ais523 used to call it "alternative programming"
20:04:10 <RodgerTheGreat> I call it "lifting". We're taking the logic embedded in the flood-fill tool and raising it to a level at which we can make use of it.
20:04:36 <Hiato> I like all of these terms ;)
20:06:57 <Hiato> http://filebin.ca/gonsrd/wingate.JPG
20:07:06 <Hiato> it's pretty much what I am using, nothing special
20:07:17 <Hiato> a minor variation on RodgerTheGreat's idea
20:07:20 <Sgeo> I ran the bf cat program for PSOX successfully!
20:07:21 <lament> you could choose a better format than jpg, you know.
20:07:35 <Hiato> but it's not for actual testing
20:08:12 <ais523> Sgeo: what was wrong beforehand?
20:08:21 <Sgeo> I still do not know, actually
20:08:40 <Hiato> the thing here is: Once the switch is sent through that activates one of the conditions, I still need to send one more black "pulse" to make it pass through the completed gate
20:08:42 <Sgeo> Probably I kept messing up the BF code
20:09:14 -!- Tritonio__ has joined.
20:10:16 -!- Tritonio_ has quit (Read error: 110 (Connection timed out)).
20:10:25 <Hiato> this is very difficult...
20:11:36 <Hiato> my verdict: not practical ( :P ), but plausible. It requires such thorough planning so as to make no "wires" cross and seeing as gates are not re-usable, I cannot consider this TC ;)
20:11:57 <ais523> lack of data storage implies that it's not TC
20:12:09 <ais523> lack of reversibility implies it's not even bounded-storage or FSA
20:12:19 <ais523> on the other hand, it may still be marginally useful for computation in some circumstances
20:13:12 <Hiato> lets see, how's about the term "disposable computation" :P
20:15:11 <ais523> grr... I must really try to figure out what the spec for that rule-110-implementing language was
20:15:36 <Hiato> ok, I officially give up on the win/lose/draw signs. I'm convinced it's not possible, as besides all the aforementioned problems, you cannot make an automated even that needs more than 1 "pulse", so I can't send a win or lose through, even if the gates somehow worked without wire crossing/bleeding of gates from one to another.
20:16:22 <RodgerTheGreat> I don't think it's not possible, but I will agree it is not a simple problem.
20:16:25 <ais523> at a guess, and partly influenced from memory:
20:16:40 <ais523> the cells of the program are coloured in colours
20:16:52 <ais523> (the comment would seem to imply that)
20:17:05 <ais523> if a R/G/B component of a cell is 0, that component can never increase
20:17:27 <ais523> a cell can only change if for each colour component that is non-0 in that cell, some neighbouring cell is brighter in that component
20:17:55 <ais523> and all the cells that do change set each of their components to the maximum of any neighbouring cell's component in that colour
20:17:55 <Hiato> RodgerTheGreat: I'm pretty darn sure that it is not possible, for the simple reason that you cannot automate the displaying of anything, as one event to complete gate, then you need another to fuse sign colours together. IE: after the user/paint has made three in a row, there needs to be one more event to trigger the signs.
20:18:33 <ais523> the program is tiled across the plane infinitely, to provide the TCness
20:18:39 <Hiato> that sounds complicated, yet fascinating
20:18:42 <ais523> and user input changes the colours somehow, to set things going
20:19:51 <ais523> it's sort of like floodfill, but more interesting
20:19:53 <Hiato> im interested though, asi523, how would one convey a message from one place to another if, with each use, the pathway will effect it's neighbours and it will reach the end of its life soon (255,255,255)
20:19:59 <Sgeo> I finished coding the system domain, except for mapping (which should be a2), and need just to test it
20:20:05 <ais523> the tiling is infinite
20:20:20 <ais523> so you start the program running at just one point on the infinite plane, and the paths go on for ever
20:20:24 <Hiato> but surely a given path can only be used x number of times
20:20:35 <ais523> yes, but you only need to use it x number of times
20:20:51 <ais523> there's a reason why the program emulates a 1D cellular automaton
20:21:00 <ais523> I think it runs from top to bottom evaluating the automaton
20:21:24 <Hiato> aha, here comes the conundrum: if it is TC complete, then it may well never halt, but your programme will halt on the xth iteration of a specific path
20:21:25 <ais523> and after a certain number of steps all cells above a certain point become irrelevant and that's the first moment that cells below a certain another point can change
20:21:35 <ais523> Hiato: there are an infinite number of paths
20:22:13 <Hiato> I was under the impression that you had to programme this thing though (in some abstract sense, like COnway's game of life). I think I see now that it builds itself :)
20:23:17 <Hiato> I think though that I am still misunderstanding you...
20:23:20 <ais523> of course, this is all deduction and a slight bit of memory, so some of the details may be wrong
20:23:30 <ais523> maybe the only solution will be to write an interpreter and see if the program runs
20:24:37 <Hiato> I must say, that above all else, Cellular Autonoma is probably the single most fascinating thing that I have seen yet on a computer
20:33:55 <GregorR> Strange that you'd misspell something that fascinating to you :P
20:34:52 <Hiato> anyway, asi523: this is pretty similar (well vaguely) to what you were proposing [if only just] http://en.wikipedia.org/wiki/Cyclic_cellular_automaton
20:34:54 <oerjan> no, Cellular Autonoma really _are_ even more fascinating. they just haven't been invented yet :D
20:35:35 <oerjan> Hiato: well you could also start spelling ais523 right...
20:35:39 <GregorR> http://www.codu.org/pics/displayimage.php?album=4&pos=6 // I wrote a raycaster that takes real elevation data and satellite imagery - I'm proud of it 8-D
20:36:23 <Hiato> blarg, it's not my night
20:37:19 <Sgeo> Ok, the regex processor is working, the current problem is in the types
20:37:21 <Hiato> GregoR: impressive :D
20:37:44 <ais523> Hiato: that link is interesting, but seems somewhat different from my irreversible programming language
20:38:46 <Hiato> true, but it is vaguely reminiscent in the sense of neighbour value comparisons to determine the next state. At least that is the parallel I drew
20:39:07 <Sgeo> Why can't this thing handle something so simple like FNUM(3)
20:39:50 <GregorR> Hiato: I have the source up somewhere, I'm not officially releasing anything because a friend of mine claimed he could make it 100x faster, so I gave it to him :P
20:40:05 <GregorR> If he does make it 100x faster, I'm turning it into a plugin for FlightGear :P
20:40:20 <Hiato> oh yeah, awesome. Sure, no worries :)
20:42:02 -!- ais523 has quit ("going home").
20:43:35 -!- RedDak has joined.
20:44:50 <Sgeo> Ok, not all of the information is being given to the type to process
20:48:57 <Sgeo> So it isn't quite argtypes's fault
20:50:34 <Sgeo> And the root cause of the problem: An * that didn't belong
20:50:53 <Sgeo> At least SEEINTERNAL now shows a lot of stuff
20:51:33 <Sgeo> Ok, now I think it's all ready woohoo
20:52:06 <Sgeo> Should I just wrap it up in the current directory structure and add a README.txt
20:55:27 <Sgeo> Here's the changeset: http://trac2.assembla.com/psox/changeset/59/
20:56:19 <Sgeo> Hiato, you there?
20:57:15 <Sgeo> The bug I was talking about most, was send(the_func(the_domain, *argtuple)) instead of send(the_func(the_domain, argtuple))
20:58:54 <Hiato> I'm not sure what you are on about, but I'll be glad to help if I can
21:05:15 <Sgeo> Ah crud. Do the .svn folders ever contain sensitive stuff?
21:05:37 <Sgeo> Because I made a .zip archive and it included .svn
21:05:45 <Sgeo> Or maybe I should remove them manually?
21:07:48 * Sgeo gets rid of some .svn and backup files that got in there.
21:07:54 <Sgeo> Then... RELEASE of a1!
21:09:59 <Sgeo> Oh, I left a test function in the system domain
21:14:17 <Sgeo> http://sgeo.diagonalfish.net/esoteric/PSOX-1.0a1.zip
21:14:33 <Sgeo> oerjan, oklopol anyone else interested?
21:22:17 <GregorR> .svn doesn't contain anything sensitive, all the login data et c is contained in ~/.subversion
21:23:43 <Sgeo> Anyone looking at the release?
21:25:22 <Sgeo> RodgerTheGreat, RedDak , Hiato?
21:26:07 <Sgeo> http://sgeo.diagonalfish.net/esoteric/PSOX-1.0a1.zip
21:26:34 -!- cherez has joined.
21:27:00 <Sgeo> http://sgeo.diagonalfish.net/esoteric/PSOX-1.0a1.zip
21:27:49 <Sgeo> Is ANYONE looking at it?
21:28:12 <Sgeo> http://esolangs.org/wiki/PSOX
21:28:47 <Sgeo> Admittedly, the code is ugly
21:30:01 <lament> so it's a protocol? where's the spec?
21:30:48 <Sgeo> It's in there, as well as http://trac2.assembla.com/psox/browser/trunk/spec
21:31:07 -!- Hiato has quit (Read error: 110 (Connection timed out)).
21:31:23 <Sgeo> RodgerTheGreat, did you run the example?
21:31:44 <RodgerTheGreat> I also find it interesting that you support math in the PSOX API- it'd definitely come in handy for languages like INTERCAL and the like
21:32:10 <Sgeo> the BF examples in the impl work while some example in ex doesn't work, I think
21:43:35 -!- RedDak has quit (Remote closed the connection).
21:43:46 <Sgeo> http://sgeo.diagonalfish.net/esoteric/PSOX-1.0a1.zip
21:43:59 <Sgeo> (does not support custom domains and domains other than 0-2)
21:45:10 <pikhq> This weekend, I plan on working on PSOX.
21:45:31 <Sgeo> Maybe I should have added custom domains before release
21:45:50 <Sgeo> Although I guess I can just tell people that SVN supports custom domains once it happens
21:46:57 <Sgeo> But I have a feeling that custom domain support would be rather simple..
21:48:58 <Sgeo> If you want to test writing your own domain, you can write it as a builtin for now, actually
21:49:35 <Sgeo> You can make it builtin 3 for instance. The thing that adds in the builtins doesn't check odd or even
21:50:12 <Sgeo> And when custom domains are implemented, they can override the builtins when a client executes the mapping function
21:52:54 <Sgeo> Where's ehird?
21:53:19 <Sgeo> PSOX can no longer be considered to be vaporware
21:53:37 <pikhq> Sure it can. Just not truthfully.
21:57:22 <Sgeo> Maybe I shouldn't have zipped the release up, just left it in SVN
21:58:08 <Sgeo> Anyone interested in an alpha version would be interested in the most up-to-date version possible
22:01:03 <Sgeo> How could such a thing POSSIBLY store body parts, or even nomod scripts for that matter...
22:01:08 <Sgeo> erm, wrong channel
22:05:58 -!- danopia`school has quit (SendQ exceeded).
22:08:37 <oerjan> indeed, get the newest version of PSOX that can store body parts
22:09:01 -!- ehird`_ has joined.
22:10:06 <Sgeo> Guess what you can't call vaporware anymore..
22:10:07 <Sgeo> http://sgeo.diagonalfish.net/esoteric/PSOX-1.0a1.zip
22:10:31 <Sgeo> (doesn't support custom domains or domains other than 0-2
22:12:27 <Sgeo> ehird`_, you awake?
22:12:49 <Sgeo> ehird`_ pinged out on another network we share
22:13:21 -!- ehird__ has joined.
22:13:25 -!- danopia has joined.
22:13:35 <Sgeo> did you see my messages?
22:13:56 -!- ehird__ has set topic: ((lambda (f) (lambda (x) (f (lambda () (x x)))) (lambda (x) (f (lambda () (x x))))) (lambda (f) (display "THE TOPIC IS NOT ") (write (f)))).
22:14:12 <ehird__> nothing after 'its still vaporware'
22:14:13 <Sgeo> http://sgeo.diagonalfish.net/esoteric/PSOX-1.0a1.zip
22:14:35 <Sgeo> I never saw an 'its still vaporware'
22:15:04 <ehird__> Sgeo: its still vaporware.
22:19:50 <ehird__> you know what would be cool
22:19:54 <ehird__> a lazy erlang/prolog hyrbid
22:23:34 <ehird__> also, evaluation terminates if something else
22:23:35 <ehird__> "hi" will never be printed.
22:23:45 <ehird__> if(x,y,z) -> not(λ -> x,y,x), z. !! this would work if you didn't need the return value.. but generally you do
22:23:54 <ehird__> if(x,y,z) → not(λ → x,y,x), z.
22:27:41 -!- ehird`_ has quit (Read error: 113 (No route to host)).
22:29:43 -!- ehird__ has quit ("K-Lined by peer").
22:29:52 -!- ehird`_ has joined.
22:31:31 <ehird`_> if(x,y,z) → not(λ → x,y,x), z. !! is it possible to make this return y or z?
22:32:09 <ehird`_> (evaluation of the predicate terminates after the first false statement, and the whole predicate becomes false, btw.)
22:34:04 <Sgeo> ehird`_, you have absolutely no interest in PSOX?
22:44:43 <Sgeo> What did you need me to say "foo" for?
22:45:14 <pikhq> Sgeo: I will be using PSOX as an optional feature in the PEBBLE2 standard library once I get that off the ground.
22:45:26 <pikhq> (probably togglable via some switch like --enable-feature psox)
22:48:16 <Sgeo> Did you poke around the released alpha yet?
22:49:59 <ehird`_> pikhq: You have any idea for 'if'? :P
22:54:27 <ehird`_> if(x,y,z) → or(and(x,y),z). !! well this would work, but it depends on 'or' and 'and' ;)
22:56:11 -!- oerjan has quit ("Good night").
22:59:16 -!- puzzlet_ has joined.
23:01:14 -!- puzzlet_ has quit (Read error: 104 (Connection reset by peer)).
23:05:15 -!- puzzlet_ has joined.
23:06:45 -!- sarehu has joined.
23:12:01 -!- puzzlet has quit (Read error: 110 (Connection timed out)).
23:18:53 <RodgerTheGreat> hey guys, I finished a new game! http://rodger.nonlogic.org/games/ICEBreaker/
23:24:30 <Sgeo> ehird`_, why? It does seem to be
23:26:08 * Sgeo wonders what this unsaved change in an example BF file was..
23:26:16 * Sgeo hopes it wasn't anything important..
23:27:10 <ehird`_> Who wants to join the TURKEY BOMB IMPL_EMENT ATION PR_OJECT?
23:30:35 <RodgerTheGreat> how the hell would you implement turkey bomb? It's a drinking game!
23:31:09 <ehird`_> DRINKING GAME -> PROGRAMMING LANGUAGE -> DRINKING GAME -> PROGRAMMING LANGUAGE&DRINKING GAME
23:31:19 <ehird`_> Also, ais523 was working on one, at one point.
23:31:26 <ehird`_> He had a trick to make it turing complete but I forget what it was
23:31:42 <ehird`_> One thing I like about his: it used ^L page feed to seperate players
23:33:02 <GregorR> All programming languages are drinking games if you do it right.
23:33:27 <ehird`_> #turkeyb, since we can't name the binaryTURKEY_BOMB :p
23:35:38 <ehird`_> (its option to enable turing completeness will, of course, be called TURING BOMB)
23:36:19 <ehird`_> you can never tell whether the drinking will stop!
23:42:01 <ehird`_> Either I have to contain a tag in each object (interferes a bit), or I have to have first-class types.
23:42:08 <ehird`_> tag is probably the best way.
23:42:15 <ehird`_> void *TURKEY_BOMB = &TURKEY_BOMB;
23:50:15 -!- Slereah has quit (Read error: 104 (Connection reset by peer)).
23:54:41 <ehird`_> } TURKEY_BOMB[] = { { 12, (struct _TB *)TURKEY_BOMB } };
23:59:37 <ehird`_> RodgerTheGreat: Hmm. ZILCH cannot be void because it needs a tag :P
23:59:50 -!- timotiis has quit ("leaving").