00:14:36 -!- tusho has quit.
01:24:22 -!- fungot has quit (Read error: 145 (Connection timed out)).
01:26:00 -!- Sgeo has joined.
01:27:33 -!- fizzie has quit (Read error: 145 (Connection timed out)).
01:35:45 -!- olsner has quit ("Leaving").
02:02:05 -!- HanDongSeong has joined.
02:10:40 -!- ihope has joined.
02:12:33 <ihope> I want to come up with a very simple set of Thue rules that are... Thue-complete.
02:13:31 <ihope> So that any Thue program can be converted into a Thue subprogram under these rules that runs with just as much nondeterminism, just as quickly.
02:15:13 <ihope> And after doing that, I want to engineer bacteria that rewrite strands of nucleic acid according to those rules.
02:19:02 <Sgeo> Me on a test: "Client, server, and circuit are NOT the three hardware components of a network, because client and server are properties of the software, not hardware. However, that was probably the anticipated response"
02:42:57 -!- bsmntbombdood has joined.
03:48:10 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | afk eating cake and drinking coffe.
04:46:20 -!- bsmntbombdood_ has joined.
04:46:27 -!- bsmntbombdood has quit (Read error: 110 (Connection timed out)).
05:02:13 -!- oerjan has joined.
05:14:12 -!- Sgeo has quit (Remote closed the connection).
05:25:14 -!- HanDongSeong has quit ("염장").
06:01:18 -!- oerjan has quit ("leaving").
06:34:39 -!- puzzlet_ has changed nick to puzzlet.
06:44:16 -!- ihope has quit (Read error: 110 (Connection timed out)).
06:58:19 -!- kar8ng1 has joined.
07:00:34 -!- fizzie has joined.
07:02:28 -!- fungot has joined.
07:07:19 -!- CO2Games has joined.
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:59:45 -!- GreaseMonkey has joined.
09:48:10 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | how would the code for C-INTERCAL be compiled on that? you would need to convert to tri-graphs right?.
09:58:44 <puzzlet> honestly speaking i don't get the topic
10:17:51 <fizzie> Trigraphs are a relatively rare C feature to cope with systems whose character sets do not include the characters C needs. For example you can write "??/" to get a "\", so you can use "??/n" as a newline in a string constant.
10:19:26 <fizzie> GCC does those when called with -trigraphs or -std=c<something>.
10:19:46 <fizzie> So I assume the topic is about some system with a limited character set, or something.
10:20:38 <fizzie> An EBCDIC system, it seems. Not unsurprising.
10:21:31 -!- kar8ng1 has quit ("Leaving.").
10:22:58 <fizzie> I remember there being an amusing snippet about trigraphs in the GCC documentation, but can't seem to find it now.
10:24:23 -!- fungot has quit (Remote closed the connection).
10:45:38 -!- Mony has joined.
11:02:37 -!- GreaseMonkey has quit ("So, how much do you love noodles?").
11:37:51 -!- ihope has joined.
12:26:47 -!- Mony has quit (Read error: 110 (Connection timed out)).
12:27:04 -!- kar8ng1 has joined.
12:27:26 -!- kar8ng1 has left (?).
12:29:09 -!- kar8ng1 has joined.
12:32:53 -!- kar8ng1 has left (?).
12:39:04 -!- ais523 has joined.
13:03:37 -!- tusho has joined.
13:04:20 <tusho> yeah, i absolutely won that one
13:04:36 <ais523> as I didn't start typing until after you'd finished
13:04:36 <tusho> joined 2008-09-23 13:03:37 +0100
13:04:39 <tusho> sent message two seconds later
13:13:16 -!- KingOfKarlsruhe has joined.
13:17:55 -!- oerjan has joined.
13:18:09 <AnMaster> ais523, hm I think I managed to implement the select operator in erlang too
13:18:22 <AnMaster> got some tests? the ones I tried seem to work
13:18:57 <ais523> hmm, I don't have tests offhand but I can just try putting some random big numbers into intercalc
13:20:38 <ais523> '#12345$#12345'~#58990 is CXIII (or 113)
13:20:46 <ais523> a random test case I used on INTERCAL
13:21:19 <AnMaster> ais523, um that would be select(mingle(12345, 12345), 58990) ?
13:21:44 <AnMaster> 7> intercal:select(intercal:mingle(12345, 12345), 58990).
13:21:54 <AnMaster> 5> io:format("~.2B~n", [intercal:select(2#11100, 2#10101)]).
13:22:22 <AnMaster> yet I get 807 for your instead of 113
13:22:26 <ais523> hmm... try with a number more than 16 bits long
13:22:42 <AnMaster> 1> intercal:mingle(12345, 12345).
13:22:55 <ais523> something really simple like intercal:select(2#11001100110011001100, 2#1010101010101010101010)
13:23:04 <ais523> AnMaster: let me check, I calculated it but the answer came out in Roman numerals
13:23:47 <ais523> 25165122525 is the actual answer from the mingle
13:24:00 <AnMaster> 9> io:format("~.2B~n", [intercal:select(2#11001100110011001100, 2#1010101010101010101010)]).
13:24:06 <ais523> that select looks right
13:24:11 <ais523> so my guess is it's mingle playing up again
13:24:27 <AnMaster> ais523, btw select uses big endian, didn't work otherwise
13:24:42 <ais523> there should just be some option to use maths-endian
13:25:05 <AnMaster> ais523, hm? There is native too
13:25:14 <AnMaster> 10> io:format("~.2B~n", [intercal:mingle(12345, 12345)]).
13:25:25 <ais523> what's 12345 in binary?
13:25:38 <AnMaster> 11> io:format("~.2B~n", ).
13:25:53 <ais523> ok, that mingle's wrong
13:26:10 <ais523> and I'm having problems figuring out wtf it's done to the bit order
13:26:19 <ais523> all your versions of mingle seem to output the right bits in the wrong order
13:26:38 <AnMaster> ais523, it is the same one that we thought worked yesterday
13:26:44 -!- oklofok has quit (Read error: 113 (No route to host)).
13:26:48 -!- moozilla has quit (zelazny.freenode.net irc.freenode.net).
13:26:53 -!- oklofok has joined.
13:27:15 <AnMaster> ais523, you said it should be 25165122525?
13:27:28 <AnMaster> 10111011011111101010100101111011101
13:27:38 <ais523> Roman numerals are hard to read
13:27:48 <ais523> and the answer should definitely have 1s and 0s in pairs
13:28:15 -!- moozilla has joined.
13:28:37 <ais523> \C\C\L\IdclxiiCCLXXV in CLC-INTERCAL format
13:28:48 <ais523> C-INTERCAL format requires a fixed-width font and is two lines high
13:28:51 <oerjan> you could use some spread function as an intermediate, that just intersperses 0 between all bits
13:29:03 <ais523> I'll use tty.i and a debugger, at least that will get the output in decimal
13:29:37 <oerjan> and then mingle(x,y) = 2*spread(x) + spread(y)
13:29:52 <ais523> C1: Expression is 0xf000fc3
13:30:02 <ais523> oerjan: that's what C-INTERCAL does
13:30:04 <oerjan> where 2* and + could easily be done with bit operations instead
13:30:07 <ais523> except it does it with bit operations
13:30:13 <AnMaster> ais523, which is 1111000000000000111111000011
13:30:25 <ais523> yes, that last answer is the right one I think
13:30:27 <AnMaster> *closer* to mine but not that close
13:30:57 <ais523> ah, I know how I misread it
13:31:04 <ais523> accidentally read lx as 51 rather than 6
13:31:16 <ais523> and likewise LXX as 52 rather than 7
13:31:22 <ais523> wow, my Roman Numerals have got rusty...
13:31:22 <AnMaster> my mingle gives 1032975 not 251662275
13:31:33 <ais523> AnMaster: it's outputting the bits in the wrong order
13:31:54 <AnMaster> ais523, but wrong order *how* this time?
13:32:17 <ais523> I don't know, it would be so much easier if the bits were separate colours so we could track them...
13:32:23 <tusho> #ESOTERIC NEWS BULLETIN
13:32:23 <ais523> Deewiant: yes, I know, I was referring to the span of digits
13:32:31 <tusho> AIS523 RESIGNS AS MAINTAINER OF C-INTERCAL
13:32:33 <oerjan> AnMaster: do you have a paste for that mingle code?
13:32:36 <tusho> CITES "ROMAN NUMERAL PROBLEMS"
13:32:37 * ais523 imagines beepy music in the background
13:32:57 <tusho> APPARENTLY ERIC S. RAYMOND, THE ORIGINAL WRITER, IS TAKING OVER
13:33:00 <ais523> tusho: nah, I was working on the build system further this morning
13:33:11 <tusho> IN AN UNRELATED NOTE, C-INTERCAL USAGE DROPS 70% GOING TO CLC-INTERCAL.
13:33:12 <ais523> it now works with -j2 (and presumably with -jN for all N)
13:33:19 <tusho> USERS CITE "ESR" PROBLEMS.
13:33:19 <oerjan> AnMaster: i mean a link
13:33:22 <AnMaster> oerjan, http://rafb.net/p/lrtTEi50.html
13:33:28 <ais523> also I think it works on cross-compile
13:33:30 <tusho> THANK YOU FOR WATCHING. WE'LL BE BACK IN A FEW HOURS.
13:33:40 <ais523> I have a cross-compiler here but not an ARM processor to run the results on
13:34:21 <ais523> and actually I like that select
13:34:40 <ais523> it could be optimised by exploiting the fact that Acc1 is always 0, though
13:35:00 <ais523> but presumably its length matters?
13:35:05 <AnMaster> ais523, problem is knowing the length yes
13:35:15 <AnMaster> you need to know how much to extract at the end
13:35:21 <oerjan> ouch erlang is harder than i hoped
13:35:32 <AnMaster> ais523, seems size() returns whole bytes only
13:35:55 <AnMaster> a better way would maybe be incrementing a number to record number of bits
13:36:03 <AnMaster> but I'd like to fix mingle first
13:37:12 <tusho> oerjan: might be just AnMaster's code...
13:37:24 <oerjan> oh <<>> is a kind of bit field record syntax, i guess
13:37:38 <AnMaster> oerjan, it is binary/bitstring syntax
13:37:40 <tusho> excuse the fact that it's obscure and impossible to read
13:37:55 <ais523> the obvious version would just be the first one you pasted, with mingle(T1, T2, <<H1:1, H2:1, Acc/bitstring>>)
13:38:03 <ais523> that looks like it should work, but IIRC it didn't
13:38:12 <ais523> hmm... maybe I should install Erlang to test stuff
13:38:14 <AnMaster> Correct: 1111 0000 0000 0000 1111 1100 0011
13:38:14 <AnMaster> Mine: 1111 1100 0011 0000 1111
13:39:06 <tusho> Why is "I'm curious, How many redditors listen to NPR (at least one show per day)", a self.reddit post, #1 on reddit.
13:39:13 <tusho> It's not a link to anything. It's just a bunch of people answering a question.
13:39:20 <tusho> Why is it #1 on a site supposed to be links to interesting stuff?
13:39:36 <ais523> AnMaster: it's not nibble-basis at all, it's swapped the first 16 bits with the second 16 bits, but left-justified for some reason
13:39:49 <ais523> if I pad to the right with 0s (note this is not normally a mathematically sensible operation):
13:40:02 <ais523> correct: 1111 0000 0000 0000-1111 1100 0011 0000
13:40:16 <ais523> yours: 1111 1100 0011 0000-1111 0000 0000 0000
13:40:30 <ais523> but why is it padding /rightwards/?
13:40:35 <ais523> that changes the value of a number
13:40:45 <ais523> and numbers don't intrinsically have a particular number of bits
13:41:05 <oerjan> maybe it's using some endianness on the level of bytes?
13:41:18 <oerjan> which conflicts with the one inside bytes
13:41:20 <ais523> oerjan: it would have to be middle-endian to swap 16-bit blocks
13:41:24 <ais523> and nothing is middle-endian nowadays
13:41:34 <tusho> just for no reason
13:41:38 <ais523> I think the one of the PDP computers was the most recent to do that
13:41:44 <tusho> http://www.google.com/search?q=erlang%20middle-endian
13:42:05 <oerjan> could that :16/little-unsigned qualifier be messing things up?
13:42:17 <AnMaster> oerjan, the default is indeed big
13:42:29 <AnMaster> oerjan, however it didn't work at all with big before
13:42:49 <oerjan> ah, so you put Int1 and Int2 as little-endian but the _result_ is interpreted as big
13:43:09 <AnMaster> 24> io:format("~.2B~n", [intercal:mingle(12345, 12345)]).
13:43:30 <AnMaster> ais523, doesn't make any sense though :P
13:43:48 <oerjan> oh wait you have :32/little-unsigned on the result pattern
13:44:00 * ais523 downloads Erlang from the Ubuntu repos
13:44:31 <AnMaster> ais523, erl -man erlang for example for that module, for language ref http://www.erlang.org/doc/reference_manual/part_frame.html
13:45:44 <ais523> according to the docs it should work even with no endianness specifier at all
13:45:46 <AnMaster> 29> io:format("~.2B~n", [intercal:mingle(16#1234, 16#4321)]).
13:47:42 <AnMaster> http://rafb.net/p/h1LInl74.html
13:48:13 * ais523 tries to think up a sensible way to calculate 16#1234 mingle 16#4321
13:48:51 <ais523> all INTERCAL calculators take input in decimal though
13:48:58 <ais523> so I simply used bash to convert to decimal in the end
13:48:59 <oerjan> well you can do each hex digit pair separately
13:49:08 -!- bsmntbombdood_ has quit (Read error: 113 (No route to host)).
13:49:25 <ais523> C1: Expression is 0x120d0e21
13:49:30 <AnMaster> ais523, or in bash: echo $(( 0x1234 ))
13:49:37 <ais523> AnMaster: yes, I used the bash echo method
13:50:58 <AnMaster> well I think my last mingle works now
13:51:20 <AnMaster> 42> intercal:select(intercal:mingle(12345, 12345), 58990).
13:51:27 <AnMaster> and that was your original one too
13:51:39 <ais523> you could have a go at the other three INTERCAL-72 operators as well if you like, to finish the set
13:51:54 <AnMaster> ais523, I don't really understand how they work
13:52:14 <ais523> ok, using lowercase letters for bits:
13:52:14 <AnMaster> ais523, I implemented select from http://esolangs.org/wiki/INTERCAL#Expression_syntax
13:53:06 <ais523> V2#abcdefghijklmnop is <<p|a,a|b,b|c,c|d,d|e,e|f,f|g,g|h,h|i,i|j,j|k,k|l,l|m,m|n,n|o,o|p>>
13:53:09 <ais523> where | means bitwise or
13:53:16 <ais523> and V is intercal's unary OR operator
13:53:20 <ais523> well, the 16-bit version
13:53:37 <ais523> they're overloaded in INTERCAL, there's both the 16-bit version and the 32-bit version, which follow the same pattern
13:54:16 <AnMaster> ais523, so the last bit is used to or the first bit?
13:54:33 <ais523> and placed in the topmost bit of the result
13:54:37 <ais523> each bit is used exactly twice
13:54:51 <ais523> numbers don't really have first and last bits
13:54:55 <ais523> topmost == most significant
13:55:07 <ais523> I'm not sure which end of the number the most significant bit is at in Erlang
13:56:17 <AnMaster> 3> <<H:1,_/bitstring>> = <<2#10000000>>.
13:56:51 <AnMaster> ais523, so p is unchanged at the end then?
13:57:25 <ais523> you get p|a in the most significant bit and o|p in the least significant
13:57:59 <oerjan> this should be easy if erlang has ordinary bit operations on bitstrings
13:58:21 <oerjan> could use a higher order function
13:58:22 <AnMaster> oerjan, you have bitshift and such on numbers
13:58:37 <oerjan> i'm talking about or, and and xor
13:59:10 <AnMaster> well exists on integers but not on full binaries
13:59:30 <AnMaster> you could convert it into a list I guess
13:59:33 <oerjan> right maybe no use for bitstrings then
14:00:23 <AnMaster> oerjan, well the bit you extract at the head of the binary is a number
14:00:34 <AnMaster> so bxor or such would work on that
14:01:09 <oerjan> yay, Irregular Webcomic's presidential election is _really_ improving now
14:02:26 <oerjan> AnMaster: you just need to rotate, and then use an integer bit operation with the original number
14:02:39 <ais523> ah, of course oerjan is an accomplished INTERCAL programmer
14:03:00 <ais523> so knows what e's talking about here, yes, that's a good way to describe it
14:03:33 <AnMaster> there is bit shift but not bit rotate
14:03:52 <oerjan> that is easy to do with bitstrings i presume
14:03:55 <ais523> bit rotate's easy enough to implement with bitmasking and shifts
14:04:03 <ais523> but with pattern-matching bitstrings it's still easier
14:05:29 <AnMaster> oerjan, it would be rather inefficient. rotating like that. Erlang got single assignment. but the tail in the binary can still work without copies thanks to it just being a pointer into the binary.
14:05:37 <AnMaster> rotating would for various reasons upset that
14:05:54 <AnMaster> for more info I suggest reading http://www.erlang.org/doc/efficiency_guide/binaryhandling.html
14:05:57 <oerjan> oh so it's a linked list sort of thing?
14:06:08 <ais523> AnMaster: pointers to single bits of memory are likely to be less efficient than copying a single 32-bit integer
14:06:11 <tusho> #ESOTERIC NEWS BULLETIN: Erlang takes over a day to do a trivial task done in other languages in a few minutes even for people who don't understand it, even though it has special data structures for strings of bits and such.
14:06:12 <AnMaster> oerjan, not a bitstring no, Erlang have Lisp style lists too
14:06:26 <ais523> that efficiency guide is presumably for large binaries
14:06:38 <ais523> tusho: there is a wtf here, but you've completely mistaken where it is
14:06:40 <AnMaster> ais523, hm true for that part.
14:06:52 <tusho> ais523: either erlang or AnMaster, I'm guessing
14:06:53 <AnMaster> well I'm not sure what oerjan meant
14:07:09 <ais523> tusho: <ais523> AnMaster: pointers to single bits of memory are likely to be less efficient than copying a single 32-bit integer
14:07:30 <ais523> and it's not a wtf with Erlang, it's a wtf due to generalising documentation without thinking about the edge cases
14:07:52 <tusho> ais523: however - it's taken over a whole day
14:08:00 <tusho> to write a function that is what, 10 lines or something now?
14:08:07 <tusho> and still doesn't work? even though it's doing something really simple?
14:08:13 <tusho> it doesn't really make me think well of erlang...
14:08:13 <ais523> it took several months to get the C versions of the INTERCAL operators right
14:08:24 <ais523> and working efficiently
14:08:29 <tusho> ais523: efficiently isn't a goal here
14:08:34 <ais523> OTOH, Erlang doesn't have a ternary type
14:08:41 <tusho> I wrote my c one in like 5 minutes after I clarified everything
14:08:50 <ais523> so it'll have to general-case the bases other than 2, just like C-INTERCAL does
14:10:03 <tusho> ais523: so you think that it's OK for erlang to cause this much hoo-hah over implementing mingle even though it has a special bitstring type thing with pattern matching designed for stuff like this, even though I wrote a C version in 5 minutes (mostly starting doing it manually then realising I could generate it) after getting exactly what it does
14:10:40 <ais523> tusho: I wrote an Erlang version (with the wrong syntax) in about 3 minutes, and could have fixed it easily
14:10:53 <oerjan> ok what's wrong with rot1(Int) -> <<X:15, H:1>> = <<Int:16>, <<R:16>> = <<H:1, X:15>, R. if i got the syntax right
14:10:54 <ais523> AnMaster wanted to use a loop for some reason, I think he's invented a bignum mingle
14:10:57 <tusho> ah, so it's just more "AnMaster is terrible at erlang"
14:11:13 <ais523> oerjan: that looks more or less right, any mistakes in it are due to me not knowing Erlang
14:11:31 <AnMaster> ais523, yes it should work with a few changes for bignum
14:12:19 <ais523> oerjan: what would your definition of rot13(Int) be, by the way?
14:13:01 <oerjan> just replace 15 and 1 by 13 and 3 or possibly the other way around
14:13:05 <AnMaster> oerjan, and well it works for 16 bit integers sure. :)
14:13:26 <ais523> oerjan: I was actually wondering how much you would redefine the function
14:13:31 <ais523> to suit the change in name
14:13:58 <oerjan> well ints are not character strings...
14:14:06 <AnMaster> anyway it would be rot13_16 or something, since you operate on 16 bits there
14:14:10 <oerjan> and also, i don't know erlang
14:14:15 <ais523> it's not necessarily an integer, it's just called Int
14:14:23 <oerjan> can the part after : be a variable?
14:14:40 <AnMaster> ais523, strings are a list of integers. or a binary
14:14:55 <oerjan> if it can then it would be easy to write a general one
14:16:30 <AnMaster> 2> "abcd" =:= [$a, $b, $c, $d].
14:17:15 <oerjan> erlang doesn't have pattern guards like haskell does it? might be nice for a rot13 function
14:18:13 <AnMaster> foo(A) when is_integer(A) and (A > 100) -> A*A.
14:19:18 <ais523> ah, now that's neater in Prolog: foo(A,Rv) :- is_integer(A), A > 100, Rv is A*A.
14:20:07 <AnMaster> ais523, actually you can use , there iirc, but with and it is more readable IMO
14:20:32 <ais523> Prolog uses , for and and ; for or too
14:20:37 <ais523> and you can write out the words
14:20:43 <ais523> they're so common though that nobody ever does
14:20:57 <ais523> foo(A,Rv) if is_integer(A) and A > 100 and Rv is A*A.
14:21:04 <ais523> just looks like a long run-on sentence, anyway
14:21:08 <ais523> so the symbols are more usual
14:21:24 <AnMaster> ais523, wouldn't work, iirc that will be (is_integer(A) and A) > 100
14:21:33 <AnMaster> which will cause issues since true > 100 makes no sense
14:21:41 <ais523> and and or have lower precedences than anything
14:22:38 <AnMaster> true is larger than 100, this is defined to make sorting in list with mixed things easier
14:22:42 <AnMaster> number < atom < reference < fun < port < pid < tuple < list < binary
14:23:00 <ais523> in C true is normally less than 100, not always though
14:23:28 <ais523> well you can compare strings to integers in C too
14:23:38 <ais523> for instance "true" > 100 is almost certainly true
14:23:41 <ais523> not a good idea, though
14:23:50 <AnMaster> yes but that will be the first char's value
14:24:03 <ais523> that would be *"true" > 100
14:24:15 <ais523> it's comparing the integral value of the pointer to the number
14:24:24 <fizzie> You are likely to get a warning for that anyway.
14:24:49 <AnMaster> ais523, http://rafb.net/p/0OeXKS19.html
14:25:20 <AnMaster> - - <-- something went wrong in the paste there
14:25:21 <ais523> ah, so and and or are operators in Erlang
14:25:28 <ais523> they're keywords in Prolog
14:26:01 <AnMaster> ais523, operators yes, but you can't make a function called or.
14:26:15 <ais523> heh, in Algol 68 you can call a function anything
14:26:48 <AnMaster> well you may be able to by quoting it, depends on if it treats functions at atom names or as something else
14:27:04 <AnMaster> since you *can* make atoms that doesn't start with a lower case char
14:27:25 <ais523> in algol 68 you got to choose whether you quoted the keywords or the identifiers or both
14:27:34 <ais523> also, in Prolog you can have arbitrary atoms too
14:27:43 <AnMaster> but that is kind of confusing, so normally not used
14:27:49 <ais523> they use the same syntax as strings for things that don't start with lowercase letters
14:27:58 <ais523> but atoms and strings are freely coercible so it doesn't matter
14:28:14 <ais523> and the compiler/interp will just pick one depending on which it feels would work better
14:28:20 <AnMaster> > atom_to_list('This is a Valid erlang Atom').
14:29:26 <AnMaster> (atom_to_list/1 is from the erlang module, which contains BIFs, that is built in functions, some of them are "auto imported" and thus doesn't need the module specifier erlang:)
14:31:38 <ais523> AnMaster: do you know of any systems which need cflags set in the environment to be able to compile?
14:32:33 <AnMaster> ais523, well depends. If I want to compile a 32-bit binary on my 64-bit multilib system then yes
14:32:52 <ais523> I'm thinking about the C-INTERCAL build
14:32:56 <ais523> at the moment it configures twice
14:33:08 <AnMaster> ais523, well doesn't automake have some AM_CFLAGS or such if you want to add custom ones?
14:33:11 <ais523> once with the settings you give, for the stuff it ends up with
14:33:25 <ais523> and once with a blank environment for the OIL compiler and other things only used during the build
14:33:46 <ais523> that way cross-compile works, it builds a OIL cross-compiler that runs natively and then cross-compiles everything else
14:34:37 <fizzie> Configure twice, build once, like that old old saying goes.
14:35:38 <ais523> http://code.eso-std.org/c-intercal/configure.ac, http://code.eso-std.org/c-intercal/buildaux/buildconfig.ac and http://code.eso-std.org/c-intercal/buildaux/Makefile.am, if you're interested in how the build system works at present
14:35:49 <AnMaster> ais523, I know that need custom LDFLAGS though
14:35:53 <ais523> personally I think it's more streamlined and nicer to use than even Automake at default
14:36:07 <AnMaster> don't remember the details but pretty sure I read about that once
14:36:08 <ais523> AnMaster: I don't think Automake even pays attention to LDFLAGS
14:36:13 <tusho> dnl Yes, that is a valid email address. If your mailer doesn't support nested
14:36:14 <tusho> dnl comments, then get a better mailer.
14:36:16 <tusho> I wonder if gmail does?
14:37:04 <ais523> also, dnl is the second most insane idea for making comments work in a language I know, after INTERCAL's
14:37:09 <tusho> The email address "ais523" is not recognised. Please fix it and try again.
14:37:47 <tusho> I think its using the comment as a name
14:38:30 -!- oklofok has quit (Read error: 113 (No route to host)).
14:38:33 -!- oklopol has joined.
14:40:09 <tusho> yay, soon tusho.net will be an openid
14:40:32 <ais523> tusho: what do you think of my build process by the way?
14:40:37 <AnMaster> ais523, you said <<p|a,a|b,b|c,c|d,d|e,e|f,f|g,g|h,h|i,i|j,j|k,k|l,l|m,m|n,n|o,o|p>>, so creating a list with p added at the start then taking every pair in it should work or?
14:40:40 <tusho> ais523: I don't understand it.
14:40:43 <ais523> AnMaster: yes, it should
14:40:48 <ais523> tusho: it's way neater than gcc's
14:40:59 <ais523> it claims to be written in autoconf, but it isn't really
14:41:15 <ais523> it uses autoconf macros for a few things like options, most of it's written directly in shell script though
14:42:22 <tusho> wtf, someone just claimed http://intertwingly.net/ looks like 1996 :D
14:47:56 <oklopol> umm, 1996 was new like ten years ago
14:48:43 <ais523> ugh, I'm getting more spam in German
14:49:19 <tusho> ais523: i love that spam
14:49:31 <tusho> read it out loud and imagine an old german guy with a moustache and glasses saying it
14:50:25 <AnMaster> 4> intercal:unary_or(16#1234).
14:51:11 <ais523> wait, was that output in decimal?
14:51:21 <ais523> give the output in hex, then I can try to work it out in my head
14:51:26 * oerjan got this spam in "norwegian" today. i _really_ wonder if spammers actually get hits on this insanely badly translated stuff...
14:51:40 <ais523> oerjan: what was it, BabelFish norwegian
14:52:19 <ais523> nah, should be 367C I think
14:52:49 <oerjan> what was worse was it tried to claim to be from the university (some phishing stuff)
14:53:07 <tusho> they just do it to avoid spam filters
14:53:08 * ais523 wonders why phishers are so bad at spelling
14:53:12 <tusho> and put the actaul content in an attached image
14:53:18 <ais523> you would think it would be a bit of a giveaway...
14:53:28 <oerjan> it wasn't just spelling, but horrible grammar too
14:53:40 <ais523> use the images that are attached to spam as CAPTCHAs
14:53:49 <ais523> that way the spammers end up fighting an arms war against each other
14:53:54 <AnMaster> ais523, 367C is 1B3E but with an extra 0 bit at the end
14:54:36 <ais523> AnMaster: show your code?
14:55:06 <AnMaster> http://rafb.net/p/fvPJUc40.html
14:55:43 <oerjan> tusho: oh that may be it. i read my mail on nvg through an xterm so i don't see attachments automatically
14:55:44 <oklopol> ais523: what do you mean by taht?
14:56:05 <ais523> oklopol: spammers obfuscate their attached images to stop anti-spam systems detecting them
14:56:18 <ais523> so you use the spammer's own technology to stop spam
14:56:29 <ais523> by either making their CAPTCHAbots detect the anti-spam
14:56:39 <ais523> or else having captchas which are too hard for the spammers to crack
14:56:45 <tusho> mine the obfuscated images
14:56:47 <ais523> it must be one or the other as it's the same problem from opposite sides
14:56:48 <tusho> and put them on websites
14:56:54 <tusho> then, the website spammers
14:56:58 <tusho> will have to break those
14:57:01 <oklopol> i thought you somehow meant the spammers would directly have to decrypt their own images or something.
14:57:02 <tusho> so that we can steal the breakage
14:57:03 <tusho> to detect their spam
14:57:08 <tusho> they won't be able to break it
14:57:09 <ais523> oklopol: no, each other's images
14:57:11 <ais523> but it comes to the same thing
14:57:12 <tusho> and website spam will drop
14:57:12 <oklopol> but right, as captchas somewhere else
14:57:25 <tusho> ais523: they're different breeds, is the problem
14:57:27 <tusho> email & website spammers
14:57:31 <tusho> your thing can kill the latter
14:57:34 <tusho> but the former will just keep going
14:57:43 <ais523> you think the email spammers would win?
14:57:52 <tusho> also: when spammers break captchas
14:57:53 <ais523> and lots of people would like it if you could kill website spam at a stroke
14:57:55 <tusho> they never reveal HOW...
14:58:02 <ais523> personally I think the website spammers would win
14:58:10 <ais523> as there's a limit to how ingenious captchas could get
14:58:11 <tusho> ais523: but you'd have to tell websites things like
14:58:18 <AnMaster> ais523, any idea why you want an ending 0 bit there
14:58:22 <tusho> Ch!ea p via.gra www.cheapviagracialis.inFo
14:58:33 <tusho> that... is not good user experience, nobody wants to enter that on a website about kittens
14:58:41 <ais523> also, what's the capital F on the TLD for?
14:58:48 <tusho> ais523: avoiding OCR
14:58:54 <tusho> also: getting the punctuation and spacing and everything right will just annoy the user
14:59:01 <tusho> they'll just not do it
14:59:13 <tusho> 'ash3yda7' is about as much as they'lle nter
14:59:23 <ais523> have you seen how silly some captchas are at the moment?
14:59:29 <tusho> and people complain
14:59:33 <ais523> which is probably easier for computers than humans
14:59:37 <tusho> rapidshare fixed their cat-captcha
14:59:42 <tusho> after people exploded into RAGE
14:59:48 <ais523> they weren't the only people using it I think
14:59:52 <ais523> it got quite popular once
15:00:04 <tusho> some letters have cats printed on top of them
15:00:05 <ais523> oklopol: AnMaster: lots of distorted letters as usual
15:00:08 <tusho> you have to only enter the ones with cats on them.
15:00:13 <ais523> but you only type the ones with faint images of cats printed on them
15:00:14 <tusho> also, the cats kind of overlap other letters
15:00:17 <tusho> and they don't even look like cats.
15:00:18 <tusho> and they're faint.
15:00:19 <ais523> there are ones with dogs printed on them instead
15:00:28 <tusho> and you can't enter them
15:00:33 <ais523> and they aren't even drawn with lines, just with dots forming a pattern
15:00:36 <tusho> pretty much the stupidest thing ever, which is why they fixed it
15:00:49 <oklopol> perhaps one of vjn's domains should be just about captchas
15:00:58 <ais523> the ironic thing is that they hardly obfuscated the letters themselves
15:01:08 <AnMaster> that sort of thing may be too hard for humans
15:01:19 <ais523> AnMaster: as I said, most likely easier for computers than humans
15:01:46 <ais523> AnMaster: as for your unary or, why on earth are you using a loop?
15:01:55 <AnMaster> ais523, did you test that unary or on 0x1234 in intercal or?
15:01:57 <ais523> it makes it pretty hard to figure out exactly what's going on
15:02:03 <ais523> AnMaster: I did it in my head
15:02:22 <AnMaster> ais523, what do you suggest using instead? To be usable with bignum
15:02:22 <ais523> unary or and and are the easiest of the operators to do in one's head
15:02:29 <AnMaster> ais523, the wrapper isn't bignum
15:02:29 <ais523> AnMaster: bignums don't have a top bit anyway
15:02:45 <ais523> so it's kind-of hard to see what the operators would do
15:02:56 <AnMaster> ais523, I thought top bit set rounded upwards to whole byte :)
15:03:30 <ais523> AnMaster: if you make that whole pair of bytes it'll be quite similar to INTERCAL's dynamic typing
15:03:40 <ais523> although the way typing actually works in INTERCAL is that there are two data types
15:03:49 <ais523> most operators have fixed prototypes
15:04:02 <ais523> but select returns either 16-bit or 32-bit depending on how many 1 bits are in its rightmost operand
15:04:46 <ais523> (N.B. C-INTERCAL doesn't actually do all that, but cheats; there's a case in INTERCAL-72 that messes up in C-INTERCAL by design, as a result, and is far too ingrained to fix)
15:05:39 <AnMaster> ais523, if you read that file you will see that the "private" functions should work for any binaries. and the wrappers are fixed size
15:05:48 <AnMaster> there is a plan behind that yes
15:05:58 <ais523> are you planning to invent bignum INTERCAL, though?
15:06:19 <AnMaster> ais523, remember the ICAL fingerprint Mike Riley made?
15:06:33 <ais523> oh dear, did it forget to specify the bitwidths?
15:06:39 <AnMaster> ais523, now efunge have bignum cells so...
15:06:46 <tusho> ais523: NIETCRLA would have a bignum mode
15:06:56 <tusho> (my hypothetical INTERCAL interpreter, see if you can get the name)
15:06:58 <AnMaster> ais523, I think it assumed fixed size cells
15:07:06 <tusho> (Pronounced nyet-cur-lah)
15:07:12 <ais523> tusho: I do, it's one of the possible variants on 64 bit middle-endian
15:07:14 <AnMaster> ais523, in fact efunge is the only bignum befunge-98 that I know about
15:07:18 <oklopol> tusho: it's quite a powerful encryption
15:07:40 <AnMaster> ais523, so thus i need bignum versions
15:07:47 <ais523> AnMaster: the problem is that many idiomatic INTERCAL expressions only work with fixed-size cells
15:08:28 <ais523> DO .5 <- '?"'&"':2~:5'~'"'?"'?":5~:5"~"#65535$
15:08:28 <ais523> #65535"'~'#65535$#0'"$#32768'~'#0$#65535'"
15:08:28 <ais523> $"'?":5~:5"~"#65535$#65535"'~'#0$#65535'"'
15:08:28 <ais523> "$"':5~:5'~#1"'~#1"$#2'~#3
15:08:31 <AnMaster> ais523, well if I did it for cfunge it would end up as 32-bit or 64-bit depending on compile time options. You mean programs won't work then?
15:08:44 <ais523> AnMaster: look at all the hardcoded constants in the above
15:09:23 <ais523> the original syslib version
15:09:32 <AnMaster> ais523, actually ICAL do seem to specify bitwidth
15:09:38 <ais523> (I later determined that part of the code was redundant, but everyone had been too confused to work it out)
15:09:56 <AnMaster> ais523, however I would like to make bignum versions anyway, bignum intercal would be fun
15:10:00 <ais523> the point is that INTERCAL has 16-bit and 32-bit data types, pretty much no matter what it's working on
15:10:16 <ais523> if you do make bignum INTERCAL, I suggest that you have a bignum consisting of all 1s
15:10:29 <ais523> that you can mingle with a bignum consisting of all 0s so as to get alternating 1s and 0s
15:10:32 <tusho> ais523: NIETCRLA would also have C-INTERCAL and CLC-INTERCAL modes, because I'm crazy like that
15:10:34 <oklopol> wouldn't that be kinda pointless though
15:10:34 <AnMaster> ais523, that would probably be possible in haskell
15:10:48 <tusho> hmm.. I like the name ITRALCEN better
15:10:52 <tusho> see if you can work THAT on out...
15:10:56 <oklopol> oh, have one bignum like that
15:10:59 <AnMaster> ais523, however erlang doesn't have infinite lists since it wasn't made by mathematicians :P
15:11:05 <ais523> you have to be able to do something like that in order to get most expressions to work
15:11:29 <ais523> tusho: you have half the characters backwards, alternating between the two sides
15:11:35 <oklopol> it's simple to make that kind of infinite bignum work
15:11:39 <tusho> ais523: no, just like this
15:11:48 <oklopol> it will have a finite amount of data at any point, and the normalization of representation is simple
15:11:54 <tusho> INTERCAL -> take the first and second letter -> type them -> put the cursor between them -> repeat with rest of strong
15:11:55 <oklopol> basically just cut it into blocks
15:12:05 <ais523> tusho: comes to the same thing, just a differnet way to describe it
15:12:11 <tusho> IN ITEN ITRCEN ITRALCEN
15:12:13 <ais523> oklopol: yes, that would work I think
15:12:26 <ais523> tusho: that looks vaguely like the Latin motto for something
15:12:38 <ais523> it's just missing a couple more vowels to work visually
15:13:41 <AnMaster> 9> io:format("~.2B~n", [intercal:unary_or(16#1234)]).
15:13:41 <AnMaster> 11> io:format("~.2B~n", [16#367C]).
15:13:59 <AnMaster> ais523, so are you sure about that last 0?
15:15:05 -!- ais523 has quit (Remote closed the connection).
15:15:23 -!- ais523 has joined.
15:16:07 <ais523> [Tue Sep 23 2008] [15:14:05] <ais523> yes
15:16:09 <ais523> [Tue Sep 23 2008] [15:14:29] <ais523> what's unary_or(1) with your implementation
15:16:22 -!- oerjan has quit ("leaving").
15:17:30 <oklopol> so, intercal or, (x<<1)|x?
15:17:43 <AnMaster> ais523: 14> io:format("~.2B~n", [intercal:unary_or(1)]).
15:17:43 <AnMaster> 10000000000000000000000000000001
15:17:43 <ais523> oklopol: pretty much, except it's a rotate not a shift
15:17:53 <ais523> AnMaster: are you rotating in the wrong direction?
15:18:00 <ais523> it should be ...0000000011
15:18:45 <oklopol> lambda x:[a|b for a,b in zip(x,x[1:]+x[:1])]
15:19:13 <ais523> oklopol: using addition for or?
15:19:25 <oklopol> >>> (lambda x:[a|b for a,b in zip(x,x[1:]+x[:1])])()
15:19:57 <ais523> I thought [a|b was the start of a list comprehension
15:19:58 <oklopol> i guess it may have looked like a haskell list comprehension
15:20:18 <ais523> what's the result of the zip, there
15:22:22 <AnMaster> ais523, I worked it out by hand and I guess I'm rotating wrong yes
15:23:45 <AnMaster> changing endbit to the other end didn't work hm
15:24:24 <oklopol> >>> [((a,b),"==>",zip(a,b)) for a,b in [(lambda r:([r() for i in xrange(10)],[r() for i in xrange(10)]))(lambda:int(__import__("random").random()*50))]]
15:24:24 <oklopol> [(([5, 17, 0, 21, 30, 46, 11, 34, 21, 13], [35, 37, 12, 26, 28, 45, 5, 21, 0, 38]), '==>', [(5, 35), (17, 37), (0, 12), (21, 26), (30, 28), (46, 45), (11, 5), (34, 21), (21, 0), (13, 38)])]
15:24:32 <AnMaster> ais523, the *least significant bit* should be used to or the most significant right?
15:24:56 <ais523> and the result should be in the most significant
15:25:13 <ais523> the least significant should also be used to or the second-least-significant with the result in the least significant
15:25:22 <ais523> so yes, you were right
15:25:22 <oklopol> ais523: i'm sure that answered your question in a confusing way
15:25:37 <ais523> oklopol: I know what zip does
15:25:44 <ais523> it was the [1:] [:1] I was wondering about more
15:25:52 <ais523> about whether it was working
15:25:57 <ais523> AnMaster: sorry for giving you wrong advice
15:26:01 <ais523> maybe you were right all along
15:26:05 <ais523> I'll run it through ick
15:26:15 <oklopol> l[:i] means the list up to i, exclusive, l[i:] means from i up, inclusive
15:26:45 <ais523> yes, 1000....00001 for #V1
15:27:09 <ais523> now I remember why unary or is such a pain to use
15:27:14 <ais523> let me check my #1234 too
15:27:25 <ais523> what's 16#1234 in decimal, again?
15:28:03 <ais523> the answer's 6974 in decimal
15:28:11 <AnMaster> anyway if my or work then it should be trivial for xor and and
15:28:15 <ais523> (sorry, I cheated and used wimpmode to speed things up a bit)
15:28:34 <ais523> ok, so it is right all along
15:28:41 <ais523> you need two wrappers by the way, for 16-bit and 32-bit or
15:28:53 <ais523> unless you look at the length of the bit field to work out which to use, I suppose
15:29:11 <AnMaster> ais523, well the loop variant works for any given size of binary
15:29:14 <tusho> http://tusho.net/blog/ <- yay pagination works!
15:29:24 <AnMaster> ais523, I need *two wrappers* is what you mean
15:29:28 <ais523> <Google> "Did you mean tsusho MARY?"
15:29:42 <ais523> AnMaster: yes, that's what I said
15:30:10 <tusho> <Google> "Did you mean tsusho MARY?" <- what
15:30:12 <AnMaster> ais523, I could make a bignum one, just need to find out how large binary is needed
15:30:22 <ais523> tusho: I was trying to find the documentation for MKRY
15:30:27 <ais523> Google didn't know where it was thuogh
15:31:39 <tusho> ais523: tip - if you want something I made, tusho.net slash name-of-thing should work well in future :P
15:31:58 <tusho> heh, my blog is the first result for tusho blogs but not tushob log
15:32:19 <ais523> probably Google doesn't know about MKRY because you didn't link it anywhere
15:32:33 <ais523> except there should be a link from ircbrowse, but maybe it's nofollowed
15:32:53 <tusho> ircbrowse won't link it
15:32:55 <tusho> it'll just have the URL
15:33:07 <tusho> anyway, I'll link to it from my blog
15:33:11 <ais523> ircbrowse links URLs IIRC
15:33:17 <tusho> ais523: do you like the pagination, by the way?
15:33:32 <tusho> if there's more than 20 pages (more than 50 posts) it'll be like this:
15:33:45 <tusho> 1 2 3 4 5 6 7 8 9 10 ... 457 458 459 460 ->
15:35:04 <tusho> ais523: http://tusho.net/blog/mkry
15:35:08 <tusho> google should pick it up sometime or other
15:35:31 <ais523> tusho: I so hope you get a comment on that written by some non-#esotericer
15:35:44 <tusho> ais523: i probably will, but it'll be "lolwut"
15:36:15 * tusho changes the footer a bit to link to tusho.net
15:37:16 <AnMaster> ais523, my last version is http://rafb.net/p/q0Ns9N24.html
15:38:14 <oklopol> the prediction was correct, i didn't get it.
15:38:14 <ais523> AnMaster: you say "32-bit" in your spec for the unaries but mention no bitcount anywhere for them
15:38:36 <oklopol> like, pushes the ascii value of '.' a few times?
15:38:55 <oklopol> well just once according to the stack thingie
15:39:07 <oklopol> hmm, actually there's an ellipsis
15:39:15 <AnMaster> ais523, the code was common so I abstracted it out :P
15:39:26 * oklopol waits for the truth to appear
15:39:34 <tusho> oklopol: its obvious,ehehehehe...
15:39:44 <ais523> AnMaster: I'd suggest removing the 32-bit restriction from the docs, if you've removed it from the code
15:40:04 <AnMaster> ais523, I didn't it was just abstracted out into the code at line 47
15:40:09 <tusho> oklopol: mikeriley says 'ehehehehehhhhehehehehe' a lot
15:40:11 <tusho> and does things like this
15:40:19 <tusho> ehehehehehehhe,,,,i don't know........
15:40:25 <tusho> his fingerprints are never well specced
15:40:29 <ais523> AnMaster: ah, spotted it now
15:40:31 <tusho> they're always as obscure and hard to understand as MKRY is
15:40:44 <tusho> each one pushes 3-15 copies of what they say
15:40:48 <tusho> C pushes 3-15 ','s to the stack
15:40:52 <tusho> D pushes 3-15 '.'s to the stack
15:40:56 <oklopol> tusho: i didn't realize it was a joke of some sort :)
15:40:58 <AnMaster> ais523, arguably the internal functions should be documented too ;P
15:41:00 <tusho> and E pushes 3-15 e/h (randomly chosen each time) to the stack
15:41:10 <tusho> it lets you efficiently and easily construct mike riley sentences
15:42:20 <AnMaster> ais523, btw would you say that GCC's error messages for syntax errors often are a bit cryptic and don't point out where the error really is?
15:42:30 <ais523> sometimes, sometimes they're useful though
15:42:34 <tusho> ais523: why did you want the spec? do you want to implement an intercal version?
15:42:48 <AnMaster> ais523, I had: "./intercal.erl:41: syntax error before: '('"
15:42:52 <ais523> I just wanted to remind myself of what it was
15:42:58 <AnMaster> now I think that is a bit less specific
15:43:09 <tusho> ais523: ah was it mentioned?
15:43:14 <ais523> AnMaster: INTERCAL's error messages are really non-sepcific
15:43:17 <ais523> tusho: no, I just remembered it
15:43:31 <ais523> the line number they give is normally the line after the error, but not always
15:43:47 <tusho> oh dear... i've sacrificed a bit of lynx usability
15:43:54 <tusho> the paginator is all scrunched together with no spaces
15:43:58 <AnMaster> ais523, well in erlang's case the errors are mostly not ordered after line numbers
15:44:05 <tusho> I guess I'll just have to put the spaces back in but wrap them with <span class="space">
15:44:08 <tusho> so I can display:none them
15:44:14 <AnMaster> so a lot of the time it is the third error or so pointing out where the issue really was
15:44:36 <tusho> this'll be ugly, though...
15:44:36 <AnMaster> ais523, since the first few are generally along the lines of "function used not defined"
15:45:10 -!- oerjan has joined.
15:46:22 <tusho> <p id="footer"><a href="/">tusho.net</a><span class="space"> </span><span class="pagination"><span class="space"> </span><span class="current-page">1</span><span class="space"> </span><a href="http://tusho.net/blog/page/2">2</a><span class="space"> </span><a class="next-page" href="http://tusho.net/blog/page/2" title="→">→</a></span></p>
15:46:27 <tusho> that is sure darn ugly :D
15:46:38 <tusho> but hey, it works nicely in lynx now
15:46:44 <tusho> and isn't exactly unsemantic
15:46:46 <ais523> AnMaster: I've seen worse, though; one system I used to program in simply told you if there was an error or not
15:46:48 <tusho> they are, after all, spaces.
15:47:01 <ais523> when there was I used to find its location using binary search and comments
15:48:10 <ais523> AnMaster: scripting language for a game
15:48:10 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | huh?.
15:48:19 <ais523> also, good timing optbot
15:50:09 <tusho> ITRALCEN would give detailed, precise and helpful error messages
15:50:17 <tusho> more than just about any other compiler out there
15:50:28 <tusho> that's about as INTERCALLY as the incredibly obscure, unhelpful ones of most INTERCAL compilers, I think
15:50:51 <AnMaster> In http://rcfunge98.com/rcsfingers.html#ICAL wtf is the "n" argument to R?
15:51:21 <ais523> number of entries deep to resume
15:51:31 <ais523> in INTERCAL you can return from a procedure other than the one you're currently in
15:51:49 <ais523> R is equivalent to 1-F1R, if it helps
15:51:52 <AnMaster> <tusho> more than just about any other compiler out there <-- have you seen clang or icc? They show the code in the errors/warnings and underline the relevant part of the line
15:52:01 <AnMaster> don't know how you would beat that
15:52:16 <tusho> python uses an up arrow on the right spot
15:52:19 <AnMaster> tusho, yes I know several other languages do that
15:52:21 <ais523> or Perl, which uses a ^ underneath?
15:52:22 <tusho> but that's a pretty basic diagnostic...
15:52:25 <tusho> and yeah, i would beat that
15:52:28 <tusho> ais523: that's what i said
15:52:44 <ais523> tusho: except you said Python and I said Perl...
15:52:47 <AnMaster> icc sometimes even make suggestions for how to fix iirc
15:52:55 <tusho> AnMaster: so does ghc, but its suggestions are useless
15:53:02 <tusho> mine would use its semantic analysis tools to suggest how to fix
15:53:04 <tusho> and not just based on the error
15:53:08 <tusho> but based on the code and the error
15:53:22 <AnMaster> that would actually be pretty interesting
15:53:27 <tusho> actually, because of the tons of help it gives you it'd probably be more useful to run it as a g ui
15:53:32 <tusho> so that you could expand/contract the warnings & errors
15:53:33 <ais523> tusho: it shouldn't need the code at all, it should be able to make the program work correctly from a copy of the sepc
15:53:35 <tusho> to avoid a flood of info
15:53:43 <tusho> ais523: hahahah, well no not quite :P
15:54:21 <AnMaster> aren't there languages like that, you tell it what you want to do, and it figures it out
15:54:37 <AnMaster> or at least have been ideas of that
15:54:43 <ais523> AnMaster: arguably Prolog is like that to start with
15:54:51 <ais523> but really you have to know the internals to use it properly, !
15:54:53 <tusho> AnMaster: declarative ones
15:55:16 <tusho> ITRALCEN is pretty much the other side of intercallyness as opposed to the compilers of today
15:55:39 <ais523> tusho: as opposed to typical INTERCAL compilers or typical programming language compilers?
15:55:50 <tusho> it's precise and helpful and featureful and everything, trying to surpass even conventional languages, and yet the language itself is still very INTERCALLy so it's helping you with tons of detail... to code an awful, awful wreck
15:55:51 <AnMaster> tusho, are you going to implement it?
15:55:57 <tusho> AnMaster: yes, sometime
15:56:04 <tusho> i haven't got all the details sorted out in my head
15:56:07 <ais523> tusho: I have plans for something like that too
15:56:16 <tusho> ais523: yes, but that's a different language
15:56:21 <tusho> which you've told me about
15:56:27 <tusho> i mean, mine isn't "easy intercal"
15:56:29 <tusho> it's kind of sinister
15:56:33 <tusho> it's polite, helpful, and all that
15:56:35 <tusho> and yet it's driving you insane
15:56:44 <tusho> because it's calmly and politely helping you to code *in INTERCAL*
15:56:50 <ais523> anyway I've had a huge revolutionary idea to revolutionise INTERCAL expressions the same way its flow control is actually really nice
15:57:18 <ais523> I was also thinking about what a dedicated INTERCAL IDE would be like
15:57:31 <ais523> I decided it would leave certain features locked until you got enough experience points to be able to use them
15:57:34 <AnMaster> tusho, you need some animated icon that pops up bubbles like: "You seem to be writing a hello world program, do you want help?"
15:57:38 <tusho> AnMaster: exactly!
15:57:41 <AnMaster> you could call it... hmm.... Clippy!
15:57:49 <tusho> if ITRALCEN included an IDE it would syntax highlight EVERYTHING
15:57:59 <tusho> while you type and move your mouse tooltips would go into excruciating detail
15:58:01 <ais523> actually mine was going to too
15:58:07 <tusho> tons of panels to the side letting you analyze and refactor your code
15:58:13 <ais523> except syntax highlight not based on syntax
15:58:13 <tusho> full integration with the compiler! project support!
15:58:24 <tusho> and it'd KILL YOU because it would be so over the top and you WOULD BE WRITING INTERCAL
15:58:30 <ais523> the highlight instead showed how often you'd used that particular construct
15:58:40 <ais523> so for instance if you use a variable a lot its name becomes bright green
15:58:42 <tusho> it'd probably be a more horrifying experience than the bare bones tools for INTERCAL of today
15:58:48 <ais523> whereas it's a dull red if it's the only time you've ever read it
15:59:07 <tusho> ais523: like heat vision for INTERCAL!
15:59:24 <ais523> the trick is that that would actually catch lots of typos
15:59:33 <ais523> heh, you could make it just work on character neighbours
15:59:40 <ais523> and then it would be language agnostic
15:59:56 <ais523> sort of like a Markov chain, but instead of generating nonsense it checks to see how much Markov-sense your code makes
16:00:08 <ais523> if the whole thing's flourescent green you've cut-and-pasted too much
16:00:20 <ais523> actually, that tool would have saved me several hours on a RL project once
16:00:22 <AnMaster> ais523, kate can be set to pop up auto completion based on existing words in the document
16:00:31 <tusho> AnMaster: so can every editor ever
16:00:31 <ais523> where I accidentally duplicated part of a table of constants
16:00:44 <tusho> it does it in one line of vimrc
16:00:49 <tusho> you don't even have to enable it
16:00:50 <tusho> it's there by default
16:00:56 <ais523> someone should make a VID which is like vi, but worse
16:01:01 <oklopol> ais523: that's a genius idea
16:01:30 <tusho> ais523: not exactly hard
16:01:35 <AnMaster> tusho, actually it may be enabled by default in kate
16:01:36 <tusho> and yes i mean bill joy vi
16:03:00 <tusho> ugh, my bad posture is starting to be uncomfortable
16:03:04 <tusho> after all these years...
16:03:47 <ais523> tusho: according to the CotC website you're still supine
16:04:33 <tusho> i think i may need a new chair
16:16:43 -!- kar8ng1 has joined.
16:17:33 <oklopol> there may be porn of it, but google definitely does not index it.
16:18:24 <oklopol> human+computer, to be exact
16:27:20 -!- jix has joined.
16:43:52 -!- KingOfKarlsruhe has quit (Remote closed the connection).
16:55:33 <tusho> ais523: you mentioned that dnl was crazy
16:55:55 <ais523> discards everything up to the next newline, including the newline
16:56:37 <ais523> so it's a cross between comment and line continuation
16:56:42 <ais523> also you can generate them at runtime
16:56:47 <ais523> which can lead to chaos on occasion
17:10:48 -!- Mony has joined.
17:13:24 -!- Sgeo has joined.
17:20:19 -!- ais523 has quit (Remote closed the connection).
17:20:35 -!- ais523 has joined.
17:24:54 -!- CO2Games has quit ("And I invented doors, no joke!").
17:28:14 <ais523> CO2Games: no you didn't
17:53:04 <tusho> CoE: CO2Games did not invent doors.
17:57:10 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
18:02:29 <AnMaster> ais523, yes I wonder about dnl, # comments work in configure.ac too
18:02:46 <ais523> AnMaster: # is even weirder in m4
18:02:54 <ais523> it echos everything from the # to the end of the line into the output
18:02:58 <ais523> without interpreting any of it
18:03:04 <ais523> so # comment dnl comment 2
18:03:11 <ais523> actually copies # comment dnl comment 2 into the output
18:03:16 <tusho> http://tusho.net/ is now an openID
18:03:38 <ais523> even people who like m4 more or less outright admit it's an esolang
18:03:45 <AnMaster> ais523, well that will result in a bash comment
18:04:59 <ais523> autoconf used to be all written in sh
18:05:28 <ais523> they used m4 to automate the process of generating configure scripts
18:05:37 <ais523> as it's apparently simple to start with
18:05:49 <ais523> just a case of writing substitutions for each command
18:06:05 <ais523> there's a lot of depth to it though, in that over time you discover increasingly weird ways to abuse its evaluation engine
18:06:23 <ais523> and power m4 users sort of actively encourage that sort of thing
18:06:25 <AnMaster> ais523, yes I wrote a few autoconf macros myself
18:06:41 <ais523> it would be to someone with your personality, I imagine
18:07:06 <AnMaster> what is that supposed to mean?
18:07:14 <tusho> AnMaster: you only like boring, sane things
18:07:25 <ais523> AnMaster: different people have different styles
18:07:42 <ais523> you tend towards things which are relatively sane and work more or less the same way each time
18:07:53 <ais523> and as a result your programs tend to be saner than most of the people's here
18:08:03 <ais523> presumably tusho thinks this is a bad thing
18:08:14 <ais523> also, befunge is sane for an esolang
18:08:21 <tusho> yeah befunge is pretty pedestrian
18:09:07 <AnMaster> ais523, and I would like Feather if you ever got it done ;P
18:09:24 <ais523> I have headaches just thinking about it
18:09:28 <ais523> probably that's why I never get it done
18:09:32 <tusho> yeah AnMaster thinks he likes feather
18:09:37 <ais523> it's brilliantly elegant despite the headaches, though
18:09:43 <tusho> but i suspect he'd ravage every single part of it if it was actaully made
18:10:17 <AnMaster> ais523, I like thinking about the weird parts of core funge
18:10:39 <ais523> AnMaster: in m4 those sorts of interactions are what you do to actually run programs, normally
18:11:21 <ais523> although I don't think it's abuse so much in this case
18:11:41 <ais523> also there are tricks involving dynamically changing the syntax of the language to do stuff, but they tend to be looked down upon
18:11:57 <AnMaster> ais523, my autoconf macros didn't really use edge cases
18:12:11 <tusho> Hooray, pavatar works too!
18:12:20 -!- olsner has joined.
18:12:21 <AnMaster> and well Feather would change the language
18:12:35 <ais523> Feather would change everything
18:12:41 <tusho> http://www.pavatar.com/check?hp=http%3A%2F%2Ftusho.net%2F <-- it wurkz
18:12:45 <ais523> from its own point of view it can change the universe
18:12:47 * tusho waits for someone to ask what that is, they always do
18:12:52 <ais523> although I suspect the universe could have other ideas
18:13:02 <AnMaster> ais523, could you even make any meaningful programs in Feather without changing the language?
18:14:01 <ais523> generally speaking you don't change the language as such but add features to it
18:14:01 <AnMaster> but that wouldn't be fun of course
18:14:10 <tusho> ais523: you should make all the core primitives modify the actual language, by changing the expression you're evaluating into its result
18:14:14 <tusho> then EVERYTHING would modify it
18:14:23 <ais523> and those features seamlessly interoperate with the other features that other libraries had added
18:14:42 <ais523> that way, for instance, you can load a debugger after the fact, and then step through the code up to the point you're at at the moment
18:14:48 <ais523> sort of like JIT debugging but even crazier
18:14:53 <AnMaster> ais523, well how can you add features that require external communication? that is not included in "turing complete". For example file io, calling external code. or whatever
18:15:19 <AnMaster> but I may be wrong since you failed to explain how feather actually works ;P
18:15:26 <ais523> I've been thinking a lot about that
18:15:34 <AnMaster> ais523, hello world program in feather please?
18:15:47 <ais523> AnMaster: I can't write anything in Feather yet, not even a NOP, really
18:16:13 <ais523> but output would ideally have to be rewound as you went back and redone as you went forwards
18:16:20 <AnMaster> ais523, what language do you plan to implement feather in? or the boot strap feather in
18:16:33 <ais523> AnMaster: I'm not sure, I was discussing this with tusho a while back
18:16:42 <tusho> ais523: perhaps squeak
18:16:44 <tusho> just for the irony
18:16:47 <ais523> probably eventually C but I want to start with something functional
18:17:21 <ais523> functional should be easier
18:17:31 <ais523> Unlambda has all the right features, actually, but is a pain to program in]
18:17:41 <AnMaster> ais523, if I know the language (and/or plan to learn it, like scheme (which I plan to learn when I get time)) I could probably help with parts that need to be done
18:18:06 * tusho notes down: Fork Feather at nearest oppertunity
18:18:28 <ais523> one of the main problems with Feather is I feel it needs an interp before it can be properly specced
18:18:38 <ais523> to see what's impossible and what's merely implausible
18:18:41 <ais523> implausible I can do, sometimes
18:19:22 <AnMaster> ais523, what are the features you need of unlambda?
18:19:34 <ais523> very first-class functions, call/cc
18:19:54 <AnMaster> ais523, and that kind of prevents a lot of the IO stuff
18:19:56 <ais523> some sane and preferably extensible way to interface with the user would be nice, too
18:20:00 <AnMaster> ais523, scheme or some other lisp maybe then?
18:20:12 <ais523> also ideally call/cc should rewind output
18:20:23 <ais523> but yes, Scheme is near the top of my list of potential languages for this atm
18:21:07 <AnMaster> ais523, rewind output could be hard, what if the output is sent to something like a printer?
18:21:34 <ais523> my intended semantics are to just do nothing if output is rewound and then the same output is reoutput
18:21:46 <ais523> but if the output is rewound and different output comes out, just print the new output
18:22:03 <ais523> that'll be closest to what the user expects, I think
18:23:23 <oerjan> feather should clearly be written in Coq or Agda
18:23:39 <AnMaster> ais523, oh btw didn't you ask some time ago about languages that can seralize functions? And that unlambda was one of them
18:23:59 <ais523> Underlambda is defined to be able to serialise functions
18:24:16 <ais523> there's no technical reason why not but it might be inefficient
18:24:21 <AnMaster> ais523, erlang can seralise functions
18:24:35 <ais523> certainly if you're using the Lisp method of storing in a list and using eval to call, it's trivial
18:24:36 <oerjan> maybe it could compile to Parrot
18:24:39 <AnMaster> in fact it is used internally in between distributed erlang nodes
18:24:41 <ais523> however lambdas tend to be better
18:24:52 <ais523> oerjan: is that a genuine language yet?
18:24:59 <AnMaster> <<131,112,0,0,3,129,1,213,76,172,89...>>
18:25:21 <ais523> presumably the resulting serialisation doesn't mean anything to anyone but Erlang?
18:25:52 <AnMaster> ais523, indeed. Erlang of same major release even
18:26:11 <AnMaster> which happens every few years or so iirc
18:26:11 <ais523> that's enough for Underlambda, anyway
18:26:23 <ais523> presumably it can deserialise the result back into a function again
18:26:54 <ais523> (Underlambda's great in that you can merely serialise a continuation if you happen to want to save an image)
18:29:39 <oerjan> Lazy Bird perhaps? although it may not have continuations
18:30:00 <ais523> continuations are very very important to the way I plan to implement Feather
18:30:24 <ais523> in fact I was thinking about writing my own language in C, merely implemented to optimise continuations for the unusual use they'll see in a Feather program
18:32:29 <Slereah2> Lazy Bird doesn't have continuation.
18:32:55 <Slereah2> At least I can't think of a way to do it.
18:33:12 <Slereah2> Mainly because I still don't know what it is
18:33:21 <tusho> Slereah2: Does it have c?
18:33:31 <tusho> Then not unless you do CPS.
18:33:52 <ais523> Slereah2: it basically involves never ever returning from a function
18:33:57 <tusho> ais523: parrot is the perl6 VM.
18:34:07 <ais523> instead you give a function pointer as an argument saying where to transfer control after the function runs
18:34:11 <AnMaster> <ais523> presumably it can deserialise the result back into a function again
18:34:15 <ais523> then you get rid of the call stack
18:34:19 <AnMaster> assuming it is a fun() style fun
18:34:19 <ais523> because you don't need it
18:34:30 <AnMaster> if it is a MFA style fun it is just a reference
18:34:42 <AnMaster> MFA style fun = fun mymodule:myfunc/2
18:34:54 <AnMaster> which is more or less like a C function pointer
18:35:05 <tusho> ais523: by the way, do you have any idea how to make scp -r handle deletes too?
18:35:17 <ais523> no, not off the top of my head
18:35:22 <ais523> scp is likely the wrong tool to be using
18:35:30 <tusho> what should i use instead, then?
18:35:35 <ais523> you could delete the entire tree at the other end then scp
18:35:42 <ais523> that would come to the same thing
18:35:56 <ais523> why is that ouch, you're overwriting all the files over there anyway
18:36:05 <ais523> (that's /why/ it's the wrong tool to be using)
18:36:05 <tusho> does rsync do deletes?
18:36:19 <ais523> one of my friends uses svn for that purpose, strangely enough
18:36:48 <oerjan> somewhat of a subversion
18:37:31 <AnMaster> tusho, it does if you give it the right command line option
18:37:52 <AnMaster> then it removes files that doesn't exist on the target side
18:38:05 <AnMaster> err which doesn't exist on source side*
18:38:32 * tusho gets an evil idea - make his homepage FOAF instead of HTML and use an xml stylesheet to show it to browsers - like http://hublog.hubmed.org/foaf.xml
18:40:04 <tusho> http://www.foaf-project.org/
18:40:09 <tusho> person metadata & relationship thing
18:40:27 <tusho> http://hublog.hubmed.org/foaf.xml is a foaf file
18:41:12 <tusho> essentially, if I do that, tusho.net will be the most semantic homepage evar
18:41:32 <tusho> i have YADIS directing to my openid, a Pavatar header, ...
18:42:30 <tusho> no search on my site, though I might add it to my blahg
18:42:53 <ais523> what about microformats?
18:43:15 <tusho> ais523: unneeded, microformats are for HTML, all the info will be in the FOAF profile
18:43:25 <tusho> since tusho.net is just going to be my identity page
18:43:31 <tusho> no microformats on tusho.net
18:43:35 <tusho> no microformats on tusho.net/blog
18:43:38 <tusho> but i don't see any that i could use
18:46:03 <AnMaster> <tusho> ais523: unneeded, microformats are for HTML, all the info will be in the FOAF profile <tusho> no microformats on tusho.net/blog <-- that is because it is html there will be no microformats?
18:46:27 <tusho> AnMaster: "*tusho gets an evil idea - make his homepage FOAF instead of HTML and use an xml stylesheet to show it to browsers - like http://hublog.hubmed.org/foaf.xml"
18:46:30 <tusho> Ideas do not apply themself immediately.
18:58:49 <tusho> so: almost all done
18:59:04 <ais523> it's quite rare to get helpful channels nowadays for some reason
18:59:17 <ais523> although the ones I'm in tend to be helpful
18:59:22 <ais523> probably why I stick around in them
18:59:29 <ais523> even #esoteric is helpful sometimes
18:59:35 <tusho> also, RDF and FOAF are both good usecases of XML, and also very easy to use
18:59:39 <ais523> fungot can give some amazing advice
18:59:41 <tusho> which i think is excellent
18:59:46 <ais523> although it isn't here right
18:59:51 <ais523> I'll have to ask optbot instead.
18:59:52 <optbot> ais523: yes. this is church numeral stuff.
19:01:59 <tusho> optbot's secrets are revealed at last!
19:01:59 <optbot> tusho: Why isn't ` apply? :-P
19:03:13 * tusho tries to figure out how to use foaf:based_near
19:03:13 <ais523> obviously optbot isn't written in Unlambda, then
19:03:13 <optbot> ais523: the thing you told Deewiant
19:05:19 <AnMaster> ais523, yes I tend to ask programming questions in esoteric
19:05:28 <AnMaster> because no one asks: "why to you need that"
19:05:44 <ais523> but normally afterwards
19:05:52 <ais523> and out of curiosity rather than pointing you to a better way
19:06:04 <ais523> well, if there is a better way, we'll let you decide if you want to use it
19:06:05 <AnMaster> of course because the whole point is doing it in weird ways
19:06:34 <AnMaster> ais523, I still like the idea of bignum intercal but I don't think I would be able to implement it
19:06:35 <ais523> if I came here and said "I'm writing a machine-code program, but I want all the characters in it to be printable", nobody would bat an eyelid
19:06:41 -!- KingOfKarlsruhe has joined.
19:07:07 <ais523> and yes I have done that
19:07:21 <ais523> it's the concept of a self uu-decoding application
19:07:32 <ais523> to distribute uudecode over Usenet to people who don't have it
19:07:40 <ais523> only a DOS version though, I don't think you can do that in ELF
19:07:49 <AnMaster> ais523, can't they access ftp?
19:08:06 <ais523> but again, why care if there are better ways?
19:09:09 <AnMaster> but the point is to try to figure out what almsort would be!
19:09:26 <AnMaster> (iirc that was how befunge was created, befunge being a typo)
19:10:33 <AnMaster> "The word "Befunge" started life as a typographical error for the word "before", typed by Curtis Coleman at 4AM on a BBS chat system."
19:10:40 <AnMaster> http://esolangs.org/wiki/Befunge#Etymology
19:10:43 <ais523> that always confused me
19:10:58 <ais523> but befunge and before are pretty different to type
19:11:14 <AnMaster> ais523, dvorak? (not sure if that matters=
19:11:20 <ais523> maybe it was a non-QWERTY keyboard?
19:11:24 <AnMaster> ais523, would you say = was an easy typo for ) and ?
19:11:44 <AnMaster> ais523, on a Swedish keyboard it is
19:11:54 <ais523> = is two keys away from ) and doesn't need shift on a UK layout
19:12:03 <ais523> so yes, I think it must be some non-UK/US layout
19:12:52 <AnMaster> see how that is easy to typo on a Swedish keyboard
19:13:25 <ais523> top row on a UK keyboard, with shift
19:13:45 <ais523> US keyboards are !@#$%^&*()_+ IIRC
19:13:54 <AnMaster> well if you include the first char before 1 then it is
19:13:58 <oerjan> dvorak doesn't look very plausible for before -> befunge either
19:14:05 <ais523> ¬ is shift-backquote for me
19:14:11 <ais523> and backquote's to the left of 1
19:14:29 <ais523> yes, it isn't even in ASCII (nor is £ for that matter but I can understand why that's on a UK keyboard)
19:14:43 <AnMaster> http://en.wikipedia.org/wiki/Image:KB_Sweden.svg
19:15:00 <AnMaster> ais523, oh and back quote is ` you mean?
19:15:17 <ais523> comes in useful when typing Unlambda programs
19:15:28 <AnMaster> ais523, well don't do that on Swedish keyboard then
19:15:43 <AnMaster> pretty bad for programming really
19:15:58 <ais523> most programming languages are US/UK-keyboard-centric
19:16:04 <ais523> although APL needed a keyboard of its own IIRC
19:16:15 <fizzie> (The Finnish keyboard is identical to the Swedish one, just to add one more point of data.)
19:16:16 <tusho> ais523: what is the http header to tell a browser to display the file, not download it?
19:16:21 <tusho> e.g. for xml with a stylesheet
19:16:25 <AnMaster> ais523, erlang works pretty well, a few uses for altgr only
19:16:33 <fizzie> Although I've made my dead keys undead.
19:16:35 <ais523> although IIRC it's different in IE for no apparent reason
19:17:06 <tusho> Is it content-disposition?
19:17:18 <AnMaster> tusho, application/xml+xhtml iirc?
19:17:19 <ais523> although isn't that for MIME type?
19:17:27 <tusho> AnMaster: no, no, you misunderstand
19:17:40 <tusho> safari & firefox etc see "application/xrds+xml" and think "lolwut? best download"
19:17:45 <tusho> I want to keep that mime type - it is correct
19:17:48 <tusho> but tell the browser to display it
19:17:56 <fizzie> Content-disposition: is at least the one people use to do "forcedly download this file instead of displaying", so I guess it could work the other way around too.
19:18:00 <ais523> maybe content-disposition is the mime type browsers think it ought to be
19:18:04 <tusho> Content-Disposition: inline
19:18:08 <AnMaster> tusho, got some example site where it works?
19:18:22 <fizzie> Yes, and "Content-Disposition: attachment" for the "download it you stupef browser" one.
19:18:25 <AnMaster> now I wonder how to set that in lighttpd config
19:18:48 <ais523> incidentally, what happens if you put a META HTTP-Equiv in an HTML file saying it isn't actually HTML after all?
19:19:10 <tusho> Beh. Despite Content-Disposition: inline, my browser insists that downloading it is a good idea.
19:19:21 <AnMaster> ais523, logically it should cause an infinite loop
19:19:28 <tusho> It isn't being set
19:19:45 <AnMaster> ais523, since it will not read the HTTP-Equiv since it isn't html
19:19:53 <tusho> <html><meta http-equiv=content-type value=text/plain><p>hello world</p></html>
19:19:57 <AnMaster> then it will see that it is html from the server headers
19:19:57 <tusho> should display the source in text/plain
19:20:05 <tusho> if I know my http & html specs correctly
19:20:08 <AnMaster> then it will see that it really was
19:20:19 <ais523> tusho: you so have to try that now
19:20:22 <tusho> AnMaster: no, it's logical
19:20:27 <tusho> you determine content type by:
19:20:36 <tusho> 2. if header says html, check html head
19:20:41 <tusho> 3. if the head mentions something, it's actually that
19:20:50 <tusho> that's how you can use value=application/xhtml+xml
19:20:54 <AnMaster> tusho, on 3 *it should backtrack to 1*
19:20:55 <tusho> even though your server says text/html
19:20:56 <ais523> bet you it doesn't actually work in practice with text/plain
19:21:04 <tusho> AnMaster: then you couldn't use Content-Type _at all_
19:21:08 <ais523> no actual money, though
19:21:13 <ais523> because I'm not certain
19:21:23 <AnMaster> tusho, um I set my webserver to serve them as application/xhtml+xml... if the extension is .xhtml
19:21:23 <tusho> Header onsuccess set Content-Disposition inline
19:21:26 <tusho> surely that should work...
19:21:32 <tusho> AnMaster: not everyone controls their host
19:21:43 <oerjan> ais523: but is it at least genuine monopoly money?
19:21:49 <tusho> AnMaster: HAHAHAHAHHAHAHAHAHAHAHA
19:22:06 <tusho> Of course I can. Will I unless needed? No,.
19:22:11 <tusho> Plus, it works for other headers.
19:22:25 <AnMaster> tusho, well of course, and you could actually use ruby or whatever instead
19:22:35 <tusho> Or I could use .htaccess, because this works:Header onsuccess set X-Pavatar http://tusho.net/avatar
19:22:41 <tusho> and therefore so should this:Header onsuccess set Content-Disposition inline
19:22:44 <fizzie> At least Firefox still renders the page "<html><head><meta http-equiv="content-type" value="text/plain"></head><body><p>foo</p></head></html>
19:22:46 <tusho> so I should fix the problem, whatever it is.
19:22:55 <tusho> fizzie: Report a bug :P
19:23:03 <ais523> no, I think fizzie's results may be right
19:23:12 <ais523> there are some things which can't be set in an http-equiv
19:23:27 <ais523> content-type with arbitrary values seems to be a good candidate for that treatment
19:23:44 <fizzie> But you can do http-equiv=content-type, value=text/html; charset=foo.
19:23:51 <tusho> Incidentally, http://www.w3.org/RDF/Validator/ARPServlet.tmp/servlet_34128.png is pretty.
19:24:44 <ais523> "I was born at an early age"?
19:25:08 <tusho> I forgot to sync the .htaccess, hah
19:25:12 <fizzie> The HTML spec is pretty vague about metadata. For HTML 4.01, "User agents are not required to support meta data mechanisms. For those that choose to support meta data, this specification does not define how meta data should be interpreted."
19:25:29 <oerjan> tusho: 
19:26:01 <tusho> Content-Disposition: inline
19:26:02 <tusho> and it still downloads
19:26:15 <ais523> oerjan: you should at least link it
19:26:25 <ais523> [http://en.wikipedia.org/wiki/Template:Fact citation needed]
19:26:34 <AnMaster> tusho, "Warning: implode() [function.implode]: Bad arguments. in /home/tusho/www/tusho.net/id/MyID.php on line 1569"
19:26:53 <tusho> AnMaster: What did you do?
19:26:56 <AnMaster> tusho, oh and http://tusho.net/id/ gives a directory index
19:27:09 <tusho> What went wrong now...
19:27:14 <AnMaster> tusho, http://tusho.net/id/ gives a directory index, it was listed in that image. so I clicked in MyID.php
19:27:19 <AnMaster> tusho, then I got that message
19:27:20 <tusho> yes okay, shut up, I'm fixing it
19:27:31 <ais523> hmm... it seems now they have a page [[w:Wikipedia:Citation needed]] to link people to
19:27:37 <ais523> *hmm... it seems now they have a page http://en.wikipedia.org/wiki/Special:Search?go=Go&search=Wikipedia:Citation_needed to link people to
19:27:58 <AnMaster> ais523, why doing it through search?
19:28:07 <ais523> AnMaster: my client does that automatically
19:28:08 <AnMaster> ais523, also I seen that before
19:28:25 <AnMaster> I mean wikipedia using it on that page
19:28:57 <AnMaster> http://tusho.net/id/ -> "Internal Server Error"
19:29:11 <tusho> AnMaster: Hi. Which part of "Shut up, I'm fixing it" did you not understand? Love, tusho.
19:29:28 <ais523> Love? Between you and AnMaster?
19:29:35 <tusho> ais523: Sarcasm, more like.
19:29:46 <AnMaster> <tusho> yes okay, shut up, I'm fixing it
19:29:51 <AnMaster> tusho, and it was a different problem :P
19:30:03 <tusho> Just a problem in the fix.
19:30:28 <AnMaster> "phpMyID is NOT compatible with Suhosin or other hardened PHP systems (Debian users take note)." <-- wtf, I wouldn't run php without Suhosin on a server
19:30:58 <AnMaster> ais523, http://www.hardened-php.net/suhosin/
19:31:41 <tusho> AnMaster: it works fine with suhoshin.
19:31:51 <tusho> however, if you read, it tells you why
19:31:56 <tusho> suhoshin disables a core feature that is needed to do openid
19:31:58 <ais523> heh, their site has a dotted underscore on every instance of "PHP" which pops up a tooltip saying "Hypertext Preprocessor"
19:33:13 <ais523> it wasn't really the repetition that got me so much, but the fact that the tooltip is wrong
19:33:42 <AnMaster> ais523, what does PHP stand for then?
19:34:01 <ais523> I think it may be "PHP Hypertext Preprocessor" nowadays, not sure though
19:34:37 <ais523> "PHP: Hypertext Preprocessor" according to Wikipedia
19:35:24 <fizzie> AnMaster: Can't really find much from the XHTML spec either -- I mean, the legal syntax is there, but it doesn't seem to say much about the semantics of meta.
19:36:30 <fizzie> In contrast, HTML 5 draft does specify that if there is a http-equiv="content-type" thing, it must consist (case-insensitively) of the literal string "text/html;", then optional spaces and "charset=encoding"; it can't be used to change the type to something non-HTML.
19:36:52 <ais523> well obviously, if it isn't HTML it doesn't conform to the HTML 5 spec
19:37:27 <ais523> think of it this way: would it make sense to specify in the spec for INTERCAL, say, "If an INTERCAL program contains the string BLUE HAIR NOW, it is actually not an INTERCAL program"?
19:38:01 <ais523> but the spec would then be contradicting itself
19:38:06 <fizzie> I think it's a nice almost-paradoxy thing that should optimally cause something more interesting than just working the way common sense says it does.
19:39:04 <oerjan> Which is perfectly appropriate for an INTERCAL spec, within limits.
19:39:35 <ais523> hmm... retrograde comments would be really useful
19:39:45 <ais523> which allow you to comment backwards to the start of the program
19:39:52 <ais523> so you can put anything you like before them
19:39:57 <ais523> the last one in a program is the only one that counts
19:40:05 <ais523> because the others are commented out
19:41:07 <AnMaster> ais523, err, parsing would be painful :/
19:41:21 <ais523> scan backwards for the first comment delimiter you find
19:41:25 <ais523> then discard everything before that
19:41:28 <ais523> as a preprocessing phase
19:43:07 <oerjan> that leaves only the question of keyword. i suggest EXTERMINATE
19:46:57 <AnMaster> if this channel what DnD what alignments would you say the different people here have?
19:47:38 <fizzie> Usually by default it will replace the first occurrence only since you didn't add the 'g' modifier.
19:47:46 <ais523> it's the first appearance by default, you have to do something like s/^(.*)what/$1was/ to get the second
19:47:57 <AnMaster> fizzie, indeed, but I wanted to be extra clear
19:48:43 <KingOfKarlsruhe> optbot: <msg id="lua"><![CDATA[function x() print ('abc') end]]></msg>
19:48:43 <optbot> KingOfKarlsruhe: 'if' is a function
19:49:05 <optbot> ais523: except copy-on-write, but that just delays the blit
19:49:19 <ais523> hmm... that was probably me who said that
19:49:23 <ais523> when discussing how fork() works
19:49:27 <oerjan> optbot: your wisdom is beyond compare
19:49:28 <optbot> oerjan: you could not infect OpenBSD boxes with it, though
19:49:44 <ais523> heh, that combination makes sense
19:49:49 <oerjan> optbot: you only run on Linux then?
19:49:50 <optbot> oerjan: also the world's best imperative language.
19:49:59 -!- fungot has joined.
19:50:11 <ais523> OpenBSD is secure against everything, even optbot's wisdom
19:50:21 <ais523> fungot: anything to add to this?
19:50:22 <fungot> ais523: its only as difficult as it might be that i only just learned it
19:50:55 <oerjan> fungot is definitely chaotic here
19:50:57 <fungot> oerjan: fnord fizzie damned if i know
19:51:18 <ais523> fungot: but are you good or evil?
19:51:19 <fungot> ais523: fnord these fnord i learned the term. people usually don't stop.
19:51:47 <fungot> fizzie: every language an fsa recognizes can easily be poked to change their habits, even in fnord stores.
19:52:02 <fizzie> I don't think we have many fnord stores around here. :/
19:52:18 <oerjan> it uses the evil word (albeit defectively), the answer should be obvious.
19:52:29 <ais523> fizzie: you should so fill up fungot with logs from #nethack
19:52:30 <fungot> ais523: of course, but i
19:52:47 <AnMaster> tusho: "Warning: implode() [function.implode]: Bad arguments. in /home/tusho/www/tusho.net/id/MyID.php on line 1569" still
19:52:59 <fizzie> ais523: Hmm, maybe I should also be there too if it's that good.
19:52:59 <tusho> AnMaster: don't load that page then, it's an internal page, so shut up
19:53:00 <fungot> KingOfKarlsruhe: chicken has something like that
19:53:02 <tusho> it's not linked anywhere
19:53:14 <tusho> http://tusho.net/id/ is my OpenID.
19:53:25 <tusho> MyID.php is phpMyID, an OpenID implementation.
19:53:33 <tusho> you're accessing it directly so it's not getting the config
19:53:48 <ais523> tusho: what username and password should I put in there?
19:54:00 <tusho> if you don't know it, well, why are you trying to log in as me?
19:54:09 <ais523> so it's a login just for you
19:54:26 <tusho> I put tusho.net in an OpenID box, it gets, from the headers, a link to http://tusho.net/yadis
19:54:29 <fungot> oerjan: let's go start a flamewar to cure your boredom),
19:54:32 <tusho> which contains a link to http://tusho.net/id/
19:54:44 <tusho> so it uses that as my openid server
19:54:49 <tusho> and that's how i can login as tusho.net
19:55:29 <ais523> fungot: YOU ARE A RUBBISH BOT! HAH! EAT FLAMES,,,,,,,,,,,,,,,,eheheheh...
19:55:30 <fungot> ais523: has scheme only 26 levels of recursion
19:55:49 <fungot> ais523: feel free to type the slash.
19:55:59 <fungot> ais523: why don't you try it out with some language settings might convert ( quote red)
19:56:46 <tusho> ^ an xml file rendered via xsl generating html
19:56:58 <oerjan> fungot: a box has in the vicinity of London subway stations like a banana
19:56:58 <fungot> oerjan: no ignorance at all, really
19:57:00 <tusho> hoorah for advanced semantic web technology
19:58:09 <oerjan> tusho: is your web page sentient yet?
19:58:14 <AnMaster> oerjan, btw I think we all are more or less chaotic in here, or we wouldn't like esolangs ;P
19:58:20 <tusho> oerjan: no, but approaching it
20:01:27 <ais523> we're all chaotic by AnMaster's standards, but AnMaster is positively lawful compared to most of us
20:01:40 <AnMaster> ais523, I say you are lawful in certain ways
20:01:45 -!- Mony has quit ("À vaincre sans péril on triomphe sans gloire...").
20:01:57 <ais523> I'm pretty lawful in some ways and chaotic in others
20:03:23 <tusho> http://tusho.net/ now looks pretty cool, too!
20:03:39 <tusho> now just to add my IM accounts and i'm done
20:03:54 <AnMaster> tusho, does the openid have to be an url?
20:04:08 -!- ais523_ has joined.
20:04:36 <AnMaster> http://openid.net/get/ says "URI" under "Roll Your Own"
20:05:01 -!- ais523 has quit (Read error: 104 (Connection reset by peer)).
20:05:13 <tusho> http:// uris are uris
20:05:18 <tusho> it just has other constraints too
20:05:22 <tusho> besides, it couldn't validate urn: openids
20:05:24 <tusho> the point is it has to communicate
20:05:28 <tusho> so that it can validate that you are you
20:05:39 <tusho> (specifically, that you own the openid you entered)
20:12:08 <AnMaster> tusho, what happens if you ever change domain?
20:12:16 <AnMaster> will that block the accounts you have?
20:12:23 <AnMaster> I prefer normal username and password then
20:12:35 <tusho> Everything expires.
20:12:40 <tusho> Username and password are no more durable.
20:12:47 <tusho> Your OpenID doesn't have to be in the same place.
20:12:59 <tusho> It can change as long as the URI containing the pointer to it stays the same.
20:13:01 <ais523_> well, I set my PGP key to a two-year expiry
20:13:08 <ais523_> because I don't expect to control my email address beyond then
20:13:23 <tusho> You could switch from a livejournal ID, say, to myopenid.com, to a self-hosted one
20:13:25 <tusho> and keep the same URI
20:13:31 <AnMaster> and then I will add a new id and revoke the old id, on the same key
20:13:33 <tusho> as long as that URI stays up
20:13:35 <tusho> as long as you change the location it points to as a openid
20:13:39 <tusho> AnMaster: most services let you do that
20:13:44 <tusho> you login with an openid and create an account
20:13:50 <tusho> later, you can associate a different openid with that
20:13:54 <tusho> and then login with that instead
20:14:05 <tusho> Accounts on each servers, but only one login.
20:14:18 <tusho> Admittedly, it'd be a hassle to change them all, but hopefully that URI doesn't change often enough for that to be a problem...
20:14:25 <AnMaster> tusho, what about fallback login with user/pass
20:14:39 <tusho> Although yes, most services allow that too.
20:14:50 <AnMaster> tusho, because I think I will control my user/pass way better than a openid
20:15:03 <tusho> Why? It's all your obligation.
20:15:10 <tusho> A domain gets sold if you let it expire yourself.
20:15:21 <AnMaster> tusho, I may not have the money
20:15:32 <tusho> AnMaster: Fine, so switch your openid on things just before the domain expires.
20:15:45 <AnMaster> tusho, no I just prefer normal user/pass
20:15:54 <tusho> Ahh. "It's better because it's better."
20:16:23 <tusho> Not the ones you're mentioning, though.
20:16:36 -!- ais523_ has changed nick to ais523.
20:16:38 <AnMaster> tusho, it is *one* of the flaws
20:16:47 <AnMaster> tusho, involving a third party is another
20:16:58 <tusho> Third parties are not involved.
20:17:08 <tusho> Only if you explicitly do so by using an openid host.
20:17:22 <oerjan> Notice: Undefined variable: commenter_portrait_url in user/themes/tusho-blogs-2/post.php line 52
20:17:29 <AnMaster> tusho, yes but the majority of the users don't have their own servers or their own domains
20:17:40 <tusho> AnMaster: So then they have to use a third-party, just like they would have to for any service.
20:18:09 <AnMaster> tusho, not if the site they want to login on uses user/pass
20:18:18 <tusho> Yes - the third party is the site.
20:18:26 <AnMaster> tusho, no that is the second party
20:18:33 <tusho> AnMaster: a year ago reddit's database got hacked.
20:18:38 <tusho> Turns out they stored passwords in plaintext.
20:18:42 <tusho> So everybody's password was stolen.
20:18:49 <tusho> They did it so they could offer a 'forgotten your password' feature.
20:18:49 <AnMaster> tusho, well that would affect *one* site
20:18:56 <tusho> Sometimes, third parties are more trustable than second parties.
20:19:15 <AnMaster> tusho, and forgotten your password should reset pass to a random string
20:19:19 <tusho> I know i'd trust an established openid host by the company that created openid and is endorsed by tons of people & organizations
20:19:27 <tusho> over some random site I want to use.
20:19:47 <AnMaster> tusho, a company can go bankrupt
20:19:57 <AnMaster> or have server issues and be down when you need to login
20:19:59 <tusho> Everything can fail.
20:20:05 <tusho> It's a question of what fails the least.
20:20:12 <AnMaster> while user/pass just depends on the site you want to login on
20:20:21 <AnMaster> it won't block logging in on other sites
20:21:17 <tusho> I am too busy to continue this, as always I'll point you to #openid for further discussion.
20:21:30 <AnMaster> tusho, a hacked password affects *one* site
20:21:38 <AnMaster> always use different passwords for different sites
20:21:57 <ais523> well there was that big controversy in the US recently where someone 'hacked' Sarah Palin's email, by guessing the answer to her password reset question IIRC
20:22:00 <tusho> (I just looove it when AnMaster does something in direct ignorance of my last message one message after)
20:22:12 <tusho> ais523: It was a guy from /b/, incidentally.
20:22:17 <ais523> yes, I'm not surprised
20:22:28 <AnMaster> ais523, yes but it would not affect all her other services. Only her email
20:22:41 <AnMaster> ais523, while with openid it would allow access to *many* services
20:22:51 <ais523> hmm... OTOH, OpenID would prevent there being easily guessable reset questions in the first place
20:22:59 <ais523> so it's all a matter of balances everywhere
20:23:06 <AnMaster> ais523, reset option shouldn't be easy
20:23:49 <AnMaster> it should require sending to your email which I admit is hard for an email account, but then I suggest a phone call is needed from the isp to the customer to reset pass for that
20:23:54 <AnMaster> it is the only sane way for email
20:24:23 <AnMaster> and not voip phone but real landline phone
20:24:54 <tusho> AnMaster: that'd fuck me over a few years ago, then
20:25:00 <tusho> admittedly, i agreed to being 13 or over...
20:25:14 <tusho> but i was incredibly shy (even more so than now) and, er, my voice was obviously not 13 :P
20:25:33 <AnMaster> tusho, well no solution is perfect
20:25:45 <tusho> a solution is don't forget your password.
20:25:52 <AnMaster> but then I'm sure that could be explained for a human
20:26:21 <ais523> tusho: well people can still fire off the password reset questions even if the account's password is forgotten
20:26:31 <ais523> as many times as they like, often...
20:26:33 <tusho> ais523: because they're trivial for anyone
20:26:51 <ais523> in some places I've had to set up reset questions
20:26:56 <ais523> but they let me pick my question
20:27:04 <ais523> so I made it hard, obscure and without context
20:27:14 <tusho> i make my question 'what is your mothers maiden name'
20:27:18 <tusho> and the answer iudfshsdiufhlsdfijaklsdhaksehu3iqey2378hny78qwjycd89asi8dmnuqwe
20:27:25 <ais523> it was something not very important, so I designed it so my family could guess it and nobody else could
20:27:30 <ais523> you know, just in case...
20:27:38 <AnMaster> tusho, I hope that isn't the real one ;P
20:27:47 <tusho> AnMaster: depends on my entropy at the time...
20:27:54 <tusho> if you seed my random-finger-basher the right way...
20:28:12 <ais523> also, the security questions normally have restrictions on them
20:28:15 <fizzie> I think Gmail has a sneaky feature in that the password reset question is only enabled if the account has not been used for five days; so if you actively use your account, no-one can use the reset feature; and if you really forget the password, you just have to wait a few days.
20:28:28 <ais523> which means the right answer doesn't get through their weak-password-detector
20:29:20 <AnMaster> tusho, actually that is rather predictable, a human will distribute about half the key presses on the left and the right side, and on average every other key press alternate between right and left side iirc
20:29:34 <tusho> AnMaster: what-everrrrrrrrrrrrr
20:30:24 <Deewiant> AnMaster: you still need the keyboard layout ;-)
20:30:35 <AnMaster> also two key presses after each other tends to be different keys
20:30:42 <tusho> Error loading stylesheet: An unknown error has occurred ()
20:30:56 <Deewiant> tusho: more so than "Error: no error"
20:31:29 <AnMaster> oh and repeating the key like abab is rather common
20:31:53 <AnMaster> <tusho> and the answer iudfshsdiufhlsdfijaklsdhaksehu3iqey2378hny78qwjycd89asi8dmnuqwe <-- no 'aa', but a few 'abab style repeats
20:32:01 <tusho> sruq894ue98tjmf98ujirkuioaua89w4k09e8rsah7ye8a95jye6897hyyujmriouj89048dk89es7kru8txinjcv dhjkzsu98ra7ik8sdiopufjdnvm hsektufy978
20:32:18 <AnMaster> tusho, more random in that aspect ;)
20:32:36 <Deewiant> just "head -b100 < /dev/urandom"
20:32:46 <ais523> | uuencode if you want it printable
20:33:02 <AnMaster> I use head of urandom then filter it through tr
20:38:46 <tusho> ais523: any ideas why http://tusho.net/ doesn't render?brb
20:39:08 <AnMaster> http://rafb.net/p/quAcgw35.html
20:39:23 <AnMaster> tusho, I get "Error loading stylesheet: (null)" in firefox
20:39:25 <ais523> in Konq I get all the text
20:39:30 <ais523> no formatting or anything
20:39:35 <ais523> just a list of text, one after another
20:40:04 <AnMaster> <?xml-stylesheet href="foaf.xsl" type="text/xsl"?>
20:40:10 <AnMaster> tusho, does that file exist? work?
20:40:45 <ais523> hmm... are you using XSLT to convert FOAF into XHTML?
20:40:57 <AnMaster> tusho, validating the style sheet gives "no doctype found, will only check well-formedness"
20:41:03 <AnMaster> and then "document is well formed"
20:46:26 <tusho> ais523: hmm... are you using XSLT to convert FOAF into XHTML?
20:46:29 <tusho> AnMaster: no it's an xsl problem
20:47:12 <tusho> uh i already knew all that
20:47:12 <AnMaster> tusho, I just tried to check the basics
20:53:21 <fizzie> xsltproc gives rather better error messages than browsers, IIRC; at least that (or some other command line tool from libxslt or something) was what I used when writing a relatively complicated html-with-xsl thing few years back.
20:54:17 <tusho> XPath error : Undefined namespace prefix
20:54:17 <tusho> xmlXPathCompiledEval: evaluation failed
20:54:17 <tusho> runtime error: file foaf.xsl line 47 element value-of
20:54:17 <tusho> XPath evaluation returned no result.
20:54:32 <tusho> WHICH NAMESPACE PREFIX YOU FUCKER
20:54:40 <tusho> <dt><xsl:value-of select="foaf:OnlineChatAccount/@rdfs:label"/></dt>
20:56:14 <tusho> ^ possibly the most semantically advanced URI for a personal homepage evar
20:56:44 <tusho> A FOAF document processed via XSLT into XHTML and in the headers a Pavatar header and a YADIS pointer (also gives YADIS via Accept header)
20:56:48 <tusho> the YADIS contains a link to my openid.
20:57:35 <oerjan> what's that picture in the corner?
20:57:39 <AnMaster> tusho, what is the olb tag for?
20:58:00 <tusho> AnMaster: Protip - when faced with unknown schemas, go to the namespace URI
20:58:12 <tusho> http://purl.org/vocab/bio/0.1/olb
20:58:26 <AnMaster> tusho, you like answering oerjan but why not try to act decent towards me too?
20:58:43 <tusho> AnMaster: oerjan's question wasn't answerable by trivial research on his own
20:59:11 <AnMaster> tusho, in this case sure, but I noticed that if ais523 asks something googable you respond
20:59:20 <AnMaster> while that isn't true if I ask
20:59:22 * oerjan is still wondering what the picture _is_
20:59:27 -!- KingOfKarlsruhe has quit (Remote closed the connection).
20:59:34 <tusho> oerjan: i can show you the entire picture if that might help
20:59:45 <ais523> I ask Googlable stuff on IRC all the time, I'm better at using IRC than I am at using Google
20:59:58 <AnMaster> tusho, yes I wonder what it is too
20:59:59 <tusho> oerjan: http://tentaclerapture.com/uploads/AccountPictures/a0c941ca5f6d27efb1008e61f473813f.jpg
21:00:03 <tusho> Figure it out from that.
21:00:04 <ais523> asking someone who knows is more reliable than asking a search engine that's being manipulated by thousands of people across the world
21:00:17 <oerjan> i hope it's not related to the site name...
21:01:55 <AnMaster> tusho, hm what is that site you linked about, the main page is just an anonymous login page heh
21:02:23 <tusho> AnMaster: All the information that is wished to be exposed about that site is on the front page.
21:02:29 <tusho> You will note the footer includes a reference to some forum software.
21:02:39 <tusho> oerjan: can you tell what it is yet
21:02:46 <tusho> AnMaster: That's all you're getting, then
21:03:01 <tusho> oerjan: that black hole is my chair
21:03:07 <tusho> should give some context
21:03:10 <AnMaster> tusho, but if it was a link to phpBB would it be related to the site using it?
21:03:24 <ais523> <a href="phpbb.com">C-INTERCAL</a>
21:03:25 <tusho> it just tells you that it is a forum
21:03:34 <ais523> *<a href="http://phpbb.com">C-INTERCAL</a>
21:03:37 <tusho> ais523: googlebomber!
21:03:41 <AnMaster> tusho, doesn't say what the forum is about
21:03:46 <oerjan> i see something that _might_ be part of a guitar
21:03:47 <tusho> AnMaster: no, it doesn't
21:03:53 <tusho> as I said, the front page is all the info you're getting
21:04:00 <tusho> if there was more info wished to be exposed, the front page would include it
21:04:18 <tusho> the bump in the top-left is a television
21:04:29 <tusho> AnMaster: I never used the word 'we'.
21:04:42 <tusho> oerjan: and the bottom-right thing, which my av comes from, is an imac box
21:04:48 <tusho> AnMaster: But 'we' would be the members.
21:05:20 <AnMaster> tusho, and what is that site above since you are a member. Really bad style not even saying "invite only forum about <foo>" on the main page or such
21:05:41 <tusho> AnMaster: Can't you grasp the concept of "We don't want to tell you anything, go away"?
21:05:46 <tusho> That is the concept that the main page conveys.
21:05:53 <AnMaster> yes and that is bad style IMO :P
21:06:11 <tusho> Yes, but your world of good style conflicts with human nature.
21:06:21 <tusho> In this case, the want of secrecy.
21:06:34 <AnMaster> oerjan, the guitar I can see, but apart from that nothing
21:06:39 -!- Slereah_ has joined.
21:06:43 <tusho> I believe ais523 knows what it is due to discussions I've had with him of the development of a custom piece of software for it.
21:06:44 -!- Slereah2 has quit (Read error: 104 (Connection reset by peer)).
21:06:54 <tusho> AnMaster: top left bulb is a tv
21:06:58 <tusho> bottom right thing is an imac box with stuff on it
21:07:11 <AnMaster> tusho, what sort of lights did you use?
21:07:13 <tusho> that part of it, blown up, looks neat.
21:07:20 <tusho> AnMaster: photo booth glow effect
21:07:24 <tusho> it was dark at the time
21:07:27 <ais523> tusho: you should have copied the photo elsewhere first
21:07:31 <ais523> to avoid this discussion
21:07:40 <tusho> ais523: it was spur of the moment
21:07:43 <ais523> if you have a http://secretwebsite.invalid, why link it?
21:07:48 <tusho> I was messing with photo booth
21:07:59 <tusho> and because that's where the image was conveniently
21:08:10 <tusho> i linked to a file incidentally on the secret site
21:08:19 <tusho> besides, i'm all for it being linked to
21:08:34 <ais523> tusho: shall we move this conversation over to ##a-cow?
21:13:34 <ais523> it must be, as I linked it here
21:14:37 * oerjan wonders if there is any language where word duplication is the usual means of negating
21:15:50 -!- Slereah2 has joined.
21:16:24 <tusho> oerjan: That's a really really good idea.
21:17:08 <ais523> alt.barney.dinosaur.die.die.die and so on
21:17:11 <ais523> you know they love him really
21:17:26 -!- Slereah_ has quit (Read error: 104 (Connection reset by peer)).
21:17:50 <tusho> alt.lucas-arts.star-wars.jar-jar-binks.die.die.die
21:18:13 <tusho> alt.fan.starwars.jar-jar-binks.die.die.die
21:18:17 <oerjan> i'd say the evidence is - inconclusive :D
21:19:09 <oerjan> i recall there is alt.fan.swedish-chef.bork.bork.bork or something close
21:19:29 <oerjan> which may be the trope starter
21:19:50 <tusho> context.noun.verb.verb.verb is pretty common, iirc
21:27:52 -!- kar8ng1 has left (?).
21:32:43 -!- oerjan has quit ("Good night").
21:37:12 <ais523> I'm sitting next to a computer with IE installed
21:37:25 <ais523> if you don't mind waiting 10 mins or so for it to log in, I'll try
21:37:29 <tusho> ais523: how badly does tusho.net fail in it?
21:37:51 <tusho> i think a friend can check
21:37:58 <tusho> unsurprisingly it fails in lynx
21:38:01 <ais523> I've started logging in now, thoughh
21:38:05 <tusho> could you check in w3m, perhaps?
21:38:06 <ais523> it workes in Epiphany over here
21:38:07 <tusho> don't have it here
21:38:15 <tusho> ah, the friend can't do it with IE anyway
21:38:58 <tusho> i mean, w3m is unlikely to contain an xslt implementation :P
21:39:27 <tusho> ais523: esplain :O
21:39:39 <ais523> OK, first, unexpectedly, it works perfectly in IE7
21:39:51 <ais523> just as well as in Epiphany
21:40:06 <tusho> well I have a lot more respect for IE7 than I did a minute ago...
21:40:13 <ais523> second, they've set up the computers here so now there are no roaming profiles
21:40:21 <ais523> all computers have temporary settings which are lost on logout
21:40:27 <ais523> finally, they 'upgraded' to Word 2007
21:40:36 <tusho> what's wrong with word 2007?
21:40:41 <ais523> probably Microsoft forced it on them though, there's far too close an agreement
21:40:45 <tusho> i've heard a lot of good things about it and the negative things have been very vague
21:40:55 <tusho> e.g. i like the idea of the ribbon
21:40:57 <ais523> tusho: it encourages people to save in formats they can't use, for one thing
21:41:01 <ais523> as for the ribbon, it's missing the point
21:41:11 <ais523> the ribbon replaces menus adequately
21:41:19 <ais523> but then you don't have quick access to the things you need
21:41:29 <ais523> one line of toolbar is enough, you can find everything else on the menus
21:41:38 <tusho> Nielson had the ribbon as one of the best UI innovations of recent years
21:41:40 <ais523> with ribbons you just get giant menus you can't hide and no toolbars
21:42:04 <ais523> but I'm most annoyed at the compatibility breakage, really
21:42:44 <ais523> Word keeps getting "cleverer" and "cleverer"
21:42:53 <ais523> to the point it becomes almost unusable due to all the things it's doing automatically
21:42:54 <tusho> docx sucks, indeed
21:43:09 <ais523> things like Smart Tags
21:48:10 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | I am.
21:52:28 <ais523> the profiles thing is annoying because it means I have to live with XP's interface
21:52:37 <ais523> or else reset it to Windows 95's every time I log in
21:52:47 <ais523> probably I'll just stick to Linux, it's less hassle
22:24:24 -!- jix has quit (Nick collision from services.).
22:24:34 -!- jix has joined.
22:27:04 -!- Slereah_ has joined.
22:28:02 -!- Slereah2 has quit (Read error: 113 (No route to host)).
22:42:18 -!- Slereah_ has quit (Read error: 113 (No route to host)).
23:08:19 -!- jix has quit ("CommandQ").
23:17:56 <fizzie> I don't suppose you're here any longer, but does http://zem.fi/g2map/index.cgi/Travel/2007/Lieksa happen to work in IE? I'm a bit surprised if it does, since I didn't really test it with that at all. (Wrote that thing last summer for some GPS-synchronized photos, haven't bothered to take more.)
23:18:15 <ais523> it would look suspicious logging on this late to test one website in IE
23:18:23 <ais523> I'll try some time later for you if you like
23:18:50 <fizzie> Well, if you feel like it. It's nothing important, though.
23:19:06 <ais523> in that cases I'll try iff I remember
23:20:24 <GregorR> fizzie: I only have IE6 on Wine, but suffice to say that it doesn't seem to work.
23:21:17 <fizzie> Working in IE6 would be even more surprising than IE7.
23:22:31 <GregorR> I assume that if it does happen to work in IE7 you'll fix that?
23:40:59 -!- Sgeo has joined.
23:57:00 <tusho> fizzie: lolwut@those directions
23:58:16 -!- ais523 has quit ("9").