00:02 (join) Lajla
00:08 (quit) mheld: Quit: mheld
00:21 (join) mheld
00:38 (quit) mheld: Quit: mheld
01:15 danking: This is gonna sound really stupid but, if I was using racket/class how would I make (send object (if option foo bar)) work the way I'd like it to work?
01:15 danking: That didn't really come out right.  How would I write (send object (if option foo bar)) if I didn't want to write a macro or otherwise change the normal way this library works.
01:19 mattmight: list-no-order in match doesn't allow duplicate variables. :(  Why not?
01:27 jeapostrophe: danking: with-method is one option and send-generic is the other
01:28 jeapostrophe: mattmight: you would like it to compile st it appears multiple times?
01:28 jeapostrophe: ie (list-no-order x x 4) would match.... (list 4 3 3) (list 4 4 4) (list 'a 4 'a) etc?
01:29 jeapostrophe: and x ... isn't what you want either?
01:35 danking: jeapostrophe: Meh. It's all so... uck.  Glad I don't normally deal with OO.
01:35 jeapostrophe: i don't like oo much either
01:36 jeapostrophe: it is so imperative, it rubs me the wrong way
01:36 jeapostrophe: (even though I agree with matthias' argument re oo functional programming, the reality is that our class system mimics java and its imperativeness)
01:42 (join) jonrafkind
01:48 mattmight: jeapostrophe: yeah, (list-no-order x x 4) should match a list which contains 4 and the other two elements the same.  (Though, that's not why I need it.)
01:48 jeapostrophe: i imagine that that would be much more expensive than match patterns normally are to compile
01:49 jeapostrophe: but the docs should definitely be updated to mention that if it is not an error
01:49 mattmight: I wanted something like (list-no-order `(last-name ,last-name) `(spouse (app split-name (list spouse-first last-name)))) to match only when the spouse has the same last name.
01:50 mattmight: I checked the docs.  It doesn't mention than list-no-order is treated specially.
01:50 jeapostrophe: ya that would be better
01:50 mattmight: But, I can see that the performance penalties could be severe.
01:52 jeapostrophe: I think my "regular expression" match patterns would be able to do that:
01:52 jeapostrophe: https://github.com/jeapostrophe/exp/tree/master/temporal-ctcs
01:52 jeapostrophe: and i compile them pretty well
01:52 jeapostrophe: there are docs
01:52 jeapostrophe: https://github.com/jeapostrophe/exp/blob/master/temporal-ctcs/automata/scribblings/re.scrbl
01:52 rudybot: http://tinyurl.com/3ae23fo
01:52 jeapostrophe: and tests
01:52 jeapostrophe: https://github.com/jeapostrophe/exp/blob/master/temporal-ctcs/tests/automata/re-test.rkt
01:52 rudybot: http://tinyurl.com/2wrc4z8
01:52 jeapostrophe: but maybe not
01:53 jeapostrophe: either way, something you might find fun
01:53 mattmight: Yeah, looks neat.
01:55 Lajla: jeapostrophe, what is matthias' argument/
01:55 Lajla: That message sends are basically functin calls?
01:57 jeapostrophe: Lajla: http://www.ccs.neu.edu/home/matthias/Presentations/ecoop2004.pdf
01:59 Lajla: jeapostrophe, this is I guess why I think slides and conferences are a bit nonsensical. =(
01:59 Lajla: I find that I have no idea what it is trying to say without Matthias' melodic voice explaining it
02:01 jeapostrophe: the short:
02:01 jeapostrophe: oo was designed to eliminate the need for state
02:01 jeapostrophe: all the original philosophy sounds like functional programming
02:01 jeapostrophe: all the good design techniques people talk about are fp
02:01 jeapostrophe: but if you try to follow them, all the implementations (ie java) are horribly broken
02:01 jeapostrophe: matthias makes a particular case against tail calls
02:02 jeapostrophe: but he gives other examples
02:02 jeapostrophe: this is no surprise given that scheme was invented to study actors (ie oo)
02:02 jeapostrophe: so, the thesis of the talk is that good oo = fp but most oo practitioners don't know that and thus are not good
02:04 Lajla: Hmm
02:04 Lajla: didn't Alan Kay require objects to hold state?
02:05 Lajla: But yeah, when objects always give the same response on the same message,s you have your basic functional programming.
02:05 Lajla: But what I think the difference is is that in Kay OO, object.message is more like object(message); whereas in Java-OO, it's more message(object)
02:14 jeapostrophe: you should try to read matthias' thing
02:14 jeapostrophe: peace out
02:14 (quit) jeapostrophe: Quit: jeapostrophe
02:31 jonrafkind: i think you could write that broken tail-recursive in java using iterators
02:31 jonrafkind: and it would still be OO
02:35 (quit) jonrafkind: Ping timeout: 250 seconds
02:48 danking: My trouble with OO is that it binds my data and my methods.  I'd rather have funcitons that do useful things on all sorts of objects without having to create this whole class hierarchy :/.  Of course, my only experience with OOP is Java and thus not exactly a top notch experience.
02:49 Lajla: danking, I think that it could be more of a thing of dynamic versus static though.
03:14 (join) zenspider
03:27 zenspider: I'm totally stumped by what should be a very straightforward sicp problem
03:28 zenspider: so far, I'm guessing I'm still missing something and that my tools or something else is wrong... not the book
03:28 zenspider: sicp seems to be hinting that * does vector maths
03:28 zenspider: with this: (define (dot-product v w) (accumulate + 0 (map * v w)))
03:28 zenspider: exercise 2.37
03:29 zenspider: mit's errata mentions nothing wrong about this problem
03:31 Lajla: More that map works on vectors, no?
03:31 Lajla: (map * v w) is your basic zipwidth
03:33 Lajla: Say you have some abstract sequence v := {v1, v2 ...} and some other abstract sequence {w2, w2, ...} then (map f v w) would be a sequence {(f v1 w1), (f v2 w2), ...}
03:33 zenspider: Lajla: right
03:33 Lajla: accumulate seems to be your basic fold here
03:33 zenspider: but that means:
03:33 zenspider: (map * '(1 2 3) '((2 -1 1) (0 -2 1) (1 -2 0)))
03:33 zenspider: would be (* 1 '(2 -1 1)) no?
03:33 zenspider: yeah, it is
03:34 Lajla: That would be an error I believe.
03:34 Lajla: It will try to (* 1 '(2 -1 1)) on you
03:34 Lajla: And say 'Oh nox, * is not defined in lists, what will we do?'
03:34 zenspider: right
03:34 Lajla: And that is what excersie 2.37 tries to do?
03:35 zenspider: so * won't double dispatch to a vector *
03:35 Lajla: Nahh, and this is a list by th way
03:35 Lajla: Not a vector.
03:35 zenspider: well... they define the above dot-product as-is... it is a given
03:35 zenspider: I mean a math vector, not a scheme/lisp vector. sorry.
03:35 Lajla: Then they encode vectors (the mathematical ones) in lists instead of in vectors (the data structure)
03:36 Lajla: Whaddayouknow, great minds think alike
03:36 zenspider: so I'm defining: (define v '(1 2 3)) and (define m '((2 -1 1) (0 -2 1) (1 -2 0)))
03:36 zenspider: and trying: (assert-equal '(3 -1 -3) (dot-product v m))
03:36 Lajla: Well, then you have to map a dot product on it.
03:36 Lajla: Not a *
03:36 Lajla: Like (map dot-product v m)
03:37 zenspider: Lajla: yes. but. that code is exactly as given in sicp. I doubt it is wrong.
03:37 Lajla: Hmm
03:37 Lajla: I would concur at this juncture.
03:37 Lajla: Strange
03:37 zenspider: that makes a lot more sense... I just don't get why the book says this
03:37 Lajla: I always had the idea that books about lisp were free of error and typos.
03:37 zenspider: http://mitpress.mit.edu/sicp/full-text/book/book-Z-H-15.html#%_sec_2.2.3 exercise 2.37
03:37 zenspider: note the footnote
03:38 Lajla: zenspider, it's a liberal conspiracy, trying to pave the way to socialism.
03:38 Lajla: Let me see.
03:38 zenspider: about using the N-way map
03:38 zenspider: haha
03:38 zenspider: I wish
03:40 Lajla: Okay, I'm confuzzled here, but this is because I'm a total noob at any of this.
03:40 Lajla: zenspider, you got my backing though.
03:40 Lajla: THis doesn't seem to make sense.
03:51 zenspider: crap. ok. I found someone's online homework and they're doubling up the vector to fit the matrix. argh
04:03 Lajla: zenspider, what do you mean/
04:04 zenspider: instead of v being '(1 2 3) it is '((1 1 1) (2 2 2) (3 3 3))
04:04 zenspider: totally lame
04:05 Lajla: Ahhh
04:05 Lajla: Well,
04:05 Lajla: basically you can't just dot product two vectors of different dimensions as far as I know.
04:06 Lajla: But what do I know, I was a good boy and realized that calculus was the opium of the vulgate a long time ago.
04:06 Lajla: And with that I mean infinitesimal calculus, surely.
04:18 rapacity: you were a boy?
04:19 Lajla: rapacity, a good one even.
04:50 (part) zenspider: "ERC Version 5.3 (IRC client for Emacs)"
17:24 (topic) -: Racket -- http://racket-lang.org/ (logs at http://racket-lang.org/irc-logs/ )
17:24 (names) -: gabot eli tv|z MayDaniel parcs evhan jonrafkind mceier carleastlund cataska tewk mattmight mario-goulart @ChanServ askhader Quetzalcoatl_ sid3k em martinhex rudybot abbe tonyg fmu clklein danking rotty alexsuraci sethalves rapacity spacebat certainty Guest74073 Demosthenes Lajla bremner_
18:10 (quit) mceier: Quit: leaving
18:44 (quit) MayDaniel: Read error: Connection reset by peer
19:05 (quit) carleastlund: Quit: carleastlund
19:06 (join) masm
19:19 (join) mheld
19:21 (quit) mheld: Remote host closed the connection
19:22 (join) mheld
20:49 (quit) jonrafkind: Read error: Operation timed out
21:22 (quit) bremner_: Ping timeout: 272 seconds
21:25 (join) bremner
21:40 (quit) parcs: Ping timeout: 255 seconds
21:41 (join) parcs
22:44 (join) corruptmemory
22:44 (quit) corruptmemory: Remote host closed the connection
22:51 (join) pattern
22:54 (join) corruptmemory
22:59 (quit) corruptmemory: Client Quit
22:59 (quit) masm: Quit: Leaving.
22:59 (join) corruptmemory
23:19 (join) offby1
23:45 (quit) offby1: Quit: Page closed
23:52 (join) offby1