00:05 (quit) dnolen: Quit: dnolen
00:58 (join) jonrafkind
01:11 (join) samth
01:12 rrenaud: i want to make a list with a bunch of random elements, and i am tempted to do something like (map generate-random (make-list n)), but it seems wasteful/wrong to generate a dummy list when all i need is the size, and to totally ignore the parameters in in generate-random
01:14 jonrafkind: (for/list ([n (in-range 1 n)]) (generate-random)))
01:27 (quit) samth: Ping timeout: 250 seconds
01:43 (quit) em: Ping timeout: 260 seconds
01:51 (join) em
01:58 rrenaud: i am trying to randomly shuffle a list
01:58 rrenaud: convinently, it is mentioned in the docs
01:58 rrenaud: (sort '(1 2 3) #:key  (lambda (x) (random))  #:cache-keys? #t)
01:58 rrenaud: . . sort: no case matching 1 non-keyword argument; arguments were: '(1 2 3) #:cache-keys? #t #:key #<procedure:temp20>
01:58 rrenaud: but i still can't manage to get it right
01:59 rrenaud: i see
01:59 rrenaud: i omitted the compare function
02:00 (quit) jonrafkind: Ping timeout: 252 seconds
02:04 (join) acon
03:03 (quit) acon: Quit: acon
03:30 (quit) qha: Remote host closed the connection
03:50 (join) qha
04:29 (join) samth
04:39 (quit) samth: Ping timeout: 240 seconds
05:23 (join) RaceCondition
05:55 (join) masm
08:12 (join) dnolen
08:25 (join) lucian
09:19 (join) MayDaniel
09:33 (quit) MayDaniel: Read error: Connection reset by peer
09:45 (join) carleastlund
10:18 (quit) dnolen: Quit: dnolen
10:27 (join) tauntaun
11:10 (join) anRch
11:14 (quit) lucian: Remote host closed the connection
11:18 (join) ckrailo
11:34 (join) lucian
11:41 (join) samth
11:50 (quit) RaceCondition: Quit: RaceCondition
11:52 samth: rrenaud, you can also use `build-list'
11:52 samth: rudybot, doc build-list
11:52 rudybot: samth: your sandbox is ready
11:52 rudybot: samth: error: help: no documentation found for: build-list provided by: "/usr/local/racket-textual/collects/scheme/main.rkt"
11:52 samth: rudybot, doc build-list racket
11:52 rudybot: samth: expecting: doc <id>
11:53 samth: rudybot, init racket
11:53 rudybot: samth: your racket sandbox is ready
11:53 samth: rudybot, doc build-list
11:53 rudybot: samth: error: help: no documentation found for: build-list provided by: "/usr/local/racket-textual/collects/racket/main.rkt"
11:53 samth: grrr, rudybot
12:00 tauntaun: Does PLT have any plans to implement a faster immutable hashtable?
12:01 (join) RaceCondition
12:01 tauntaun: Let me phrase it more honestly: would PLT consider it?
12:05 (quit) samth: Ping timeout: 250 seconds
12:08 bremner_: can haz code?
12:18 (join) mceier
12:24 (quit) tauntaun: Quit: Ex-Chat
12:31 (quit) anRch: Quit: anRch
12:32 (join) MayDaniel
12:47 (join) anRch
13:03 (quit) rudybot: Ping timeout: 252 seconds
13:06 (join) tauntaun
13:09 carleastlund: tauntaun: are our immutable hashtables excessively slow?
13:10 (quit) anRch: Quit: anRch
13:19 tauntaun: carleastlund: For my present purposes, the immutable hash is fine.  But I'm trying to keep a small community interested in Racket, and I sense (maybe wrongly) that they would prefer something asymptotically better.
13:27 (join) samth
13:28 samth: tauntaun, definitely
13:33 tauntaun: samth: I'm actually not sure what the "definitely" refers to.  :-)   You saying PLT would definitely be interested in improving the hash?
13:33 samth: tauntaun, yes
13:34 samth: irc as async medium :)
13:35 (join) dnolen
13:35 tauntaun: So I'm curious: is there any particular data structure you had in mind?
13:36 dnolen: is it possible to select the language you want to use from the REPL?
13:53 eli: tauntaun: What would you use other than a balanced tree?
13:53 eli: dnolen: You can use the `-I' flag, if your question is about the racket command line.
13:54 samth: eli, there are a bunch of better data structures
13:54 samth: Clojure has a good one, for example
13:54 eli: OK, but that doesn't answer my question...
13:55 samth: no, i mean a better immutable hash table data strucutre
13:55 dnolen: eli: like, 'racket -I r5rs' ?
13:55 eli: dnolen: yes.
13:55 dnolen: eli: thx
13:56 eli: samth: My question is *what* structure?
13:56 eli: (I'm not arguing, I just am pretty out of touch with that stuff.)
13:57 eli: s/just am/am just/
13:57 samth: eli, Hari's thesis describes Clojure's data structure (hash lists)
13:57 samth: that's what I'd try first, if I were tauntaun
13:58 eli: samth: Is there a quick description of that?
13:59 eli: (There is a wikipedia page, which is on something else.)
14:00 rrenaud: wow, log(N) for hash lookup seems pretty slow/scary/lame to me
14:02 samth: eli, P. Bagwell. Fast Functional Lists, Hash-Lists, Deques and Variable Length Ar-
14:02 samth: rays
14:02 eli: rrenaud: log time is "practically constant", I suspect that the advantages would be in creating the hash.
14:03 rrenaud: an in other news, quicksort is O(N)?
14:03 tauntaun: eli: a patricia trie? bitmapped vector trie?  I really don't know, as I'm (quite purposely) not a data structures buff.
14:04 samth: rrenaud, have you benchmarked you use of racket functional hash tables?
14:04 eli: rrenaud: Huh?  Obviously not.
14:05 rrenaud: i've barely used racket, it just surprised me that someone would be happy with a library implementation of hash tables that are as slow as O(log(n)) for lookup
14:05 tauntaun: One of those has an access time of log_32(n), which is pretty close to constant, IMO.
14:06 eli: rrenaud: That's the *immutable* version, mutable hash tables are what you'd expect.
14:07 rrenaud: in my mind, an immutable version should be at least as fast as a mutable version
14:07 rrenaud: take a data structure, remove some constraints from it, it should be no worse
14:08 (quit) tauntaun: Quit: Ex-Chat
14:09 eli: rrenaud: That's certainly possible: you're talking about implemnting the immutable hash table as a mutable one with a bit set that forbids modifying it.
14:10 eli: But the problem with that is adding an item to the hash, or "modifying" it (in the sense of creating a new one with a different mapping for one value).
14:10 eli: If you implement it as a regular hash with a read-only bit, these operations become an extremely expensive copy operation for the whole table.
14:11 eli: samth: The abstract of that thing is talking about vlists...
14:13 eli: (Which is a good point in itself, BTW -- I never thought about replacing conses with something like that as the fundamental type for lists, and it sounds like they did it.)
14:13 rrenaud: i mean, i can see how to make a mutable hash backed by an immutable hash such that those things are still fast
14:14 rrenaud: i guess i am mostly in from the procedural world, where immutable -> mostly static
14:15 eli: rrenaud: The problem is implementing *immutable* hash tables, not mutable ones.
14:15 eli: I can see how coming from an imperative world would make things confusing...
14:15 eli: Think about creating a hash table in steps of adding one item at a time.
14:16 eli: You start with an empty table H0, then you add one item, and the result is H1 that has that item, etc etc.
14:16 rrenaud: right
14:16 eli: You end up with Hn, which is the table of everything you wanted to put in.
14:16 rrenaud: i mean, i can see vaguely see how to do this, but seems like the structure is much more like a tree than a hash
14:16 eli: But the whole point is doing that in a "functional" way, which means that you should still have access to H1, H2, H3, ..., Hn-1
14:17 rrenaud: a hash is a big flat list with some collision resolution mechanism
14:17 eli: And if the implementation is a conventional hash table, then that means that you now have n hash tables in memory -- and that is definitely not a good idea.
14:22 (quit) martinhex: Read error: Connection reset by peer
14:25 (join) martinhex
14:30 (quit) martinhex: Write error: Connection reset by peer
14:33 (join) LarryNY
14:33 (part) LarryNY
14:36 (join) martinhex
14:43 (quit) MayDaniel: Read error: Connection reset by peer
14:46 (quit) martinhex: Read error: Connection reset by peer
14:47 (join) martinhex
14:51 (join) saint_cypher
14:52 (quit) saint_cypher: Remote host closed the connection
14:54 (join) saint_cypher
15:23 (quit) martinhex: Read error: Connection reset by peer
15:24 (join) martinhex
16:13 (join) mceier_
16:14 (quit) mceier: Ping timeout: 250 seconds
16:25 samth: eli, ping
16:26 samth: eli, please update svn permissions
16:36 samth: clklein, one other thing
16:37 samth: here's a hack to get more space @nocopyrightspace
16:37 (quit) ckrailo: Quit: Computer has gone to sleep.
16:38 eli: samth: I just saw it now, and did it.
16:39 clklein: samth: sneaky :) Thanks!
16:39 samth: thanks, eli -- works now
16:39 samth: clklein, makes your life harder if you get in, but that the good kind of problem :)
16:40 clklein: heh
17:05 samth: clklein, another space saver -- cut the gui elements from the screenshot (the font size etc)
17:14 (join) ckrailo
17:26 (join) anRch
17:26 (join) tauntaun
18:04 (join) PLT_Notify
18:04 PLT_Notify: racket: master Robby Findler * 99e9bd9 (3 files in 2 dirs): exported arrow->pict
18:04 PLT_Notify: racket: master Robby Findler * 85b846f (1 files in 1 dirs): added one more accented char to scribble
18:04 PLT_Notify: racket: master commits 8793475...85b846f - http://bit.ly/fkxQbl
18:04 (part) PLT_Notify
18:14 (quit) lucian: Ping timeout: 248 seconds
18:15 (join) lucian
18:17 (quit) ckrailo: Quit: Computer has gone to sleep.
18:22 (join) PLT_Notify
18:22 PLT_Notify: racket: master Vincent St-Amour * 2e81106 (1 files in 1 dirs): Add the capital version of an accented character to Scribble. - http://bit.ly/gRY4T8
18:22 (part) PLT_Notify
18:24 (quit) anRch: Quit: anRch
18:25 clklein: stamourv: Trying to one-up Robby, eh?
18:26 (quit) tauntaun: Quit: Ex-Chat
18:26 DT``: how can I get a byte string from a signed byte, like with integer->integer-bytes?
18:27 stamourv: clklein: my knowledge of odd characters is my pride
18:27 stamourv: did you know that unicode has a character for the zero-thirds fraction?
18:27 stamourv: it's used in japanese baseball to count the number of outs, so of course, it had to go into unicode
18:28 clklein: obviously
18:28 clklein: What else are going going to do with all those bits?
18:29 stamourv: well, they _are_ being somewhat conservative
18:29 stamourv: for instance, I believe they're still adamantly opposed to including the klingon alphabet
18:29 stamourv: which is, as you would expect, the source of much chagrin in some corners of the internet
18:30 clklein: :)
18:30 (join) ckrailo
18:33 (part) rrenaud
18:59 (join) acon
19:18 (quit) mceier_: Quit: leaving
19:26 (quit) dnolen: Ping timeout: 252 seconds
19:26 (quit) acon: Quit: acon
19:46 (quit) ckrailo: Quit: Computer has gone to sleep.
19:50 (quit) saint_cypher: Ping timeout: 240 seconds
20:08 (quit) masm: Quit: Leaving.
20:12 (quit) samth: Ping timeout: 252 seconds
20:47 (join) tauntaun
20:56 (join) dnolen
21:00 tauntaun: Two questions.  First, sorry if I'm being a nag about this, but is no one else lamenting the loss of four days of IRC logs?  Second, would PLT be interested in my presenting my burgeoning text-based RPG at Racket Days, or is that a bit passe?  (Question 2b: is there a deadline for Racket Days?)  (Question 2c: Can I attend Racket Days just as a spectator, without contributing/presenting?)
21:14 (join) saint_cypher
21:16 carleastlund: tauntaun: I can answer 2a and 2c; the answer to both is yes.  Feel free to present any Racket program if you think other Racketeers might find it interesting.  On the other hand, if you don't want to present, you can still absolutely attend.
21:16 (quit) carleastlund: Quit: carleastlund
21:57 (quit) DT``: Ping timeout: 248 seconds
22:00 (quit) tauntaun: Quit: Ex-Chat
22:09 (join) DT``
22:20 (join) tauntaun
22:28 (quit) RaceCondition: Quit: RaceCondition
23:04 (join) samth
23:08 (join) ckrailo
23:09 (quit) ckrailo: Remote host closed the connection
23:10 (join) ckrailo
23:20 (quit) tauntaun: Quit: Ex-Chat
23:23 (quit) qha: Ping timeout: 240 seconds
23:25 (join) qha
23:25 samth: clklein, you still working?
23:33 (quit) DT``: Ping timeout: 255 seconds
23:46 (join) DT``