00:31 (quit) misterm: Quit: leaving
00:53 (quit) masm: Quit: Leaving.
00:58 (join) jonrafkind
01:07 (join) dnolen
01:09 (quit) winxordie: Ping timeout: 240 seconds
01:14 (join) ghiu
01:22 (join) winxordie
01:32 ghiu: hey guys
01:36 jonrafkind: sup bro
01:46 ghiu: hey, fine, i just discovered racket and i'm following the tutorial
01:49 ghiu: i'd really love to use a functional language for my next website and i'm making my mind
01:51 jonrafkind: if you use racket's webserver you can even use continuations instead of the normal Restful api thing every other language in existence uses
01:51 jonrafkind: continuations make it easier to write programs without breaking up the logic as dictated by the http protocol
01:52 jonrafkind: although to be fair ive never written a webapp myself, just seen others
01:52 ghiu: where can i read more about continuations?
01:52 ghiu: the thing that bothers me, at the moment, is what i'm seeing about html
01:53 ghiu: i'd rather not spit out html from the code
01:53 ghiu: it's fine for a demo
01:53 jonrafkind: you don't have to, im pretty sure
01:53 ghiu: but for a real world app, i want to write html by hand and use some sort of templating
01:54 jonrafkind: http://docs.racket-lang.org/guide/Continuations.html?q=continuations
01:54 ghiu: thanks for the link :)
01:58 ghiu: so what do you think are viable alternatives for some sort of templating?
01:58 ghiu: also, what is the most mature database interface available for racket?
01:59 jonrafkind: well you can use s-expressions and quasi-quoting as a convenient way to build up data. `(a (href ,my-link) "Click here!")
01:59 jonrafkind: im not sure about maturity, but there are interfaces for mysql, postgres, and mongodb
01:59 ghiu: oh, ok :)
01:59 ghiu: tnx
02:00 jonrafkind: http://planet.racket-lang.org/
02:00 jonrafkind: you can look here for user contributed packages, including some database interfaces
02:02 ghiu: i'm looking at it right now
02:02 ghiu: thanks
02:02 ghiu: is there anything like "eval" in racket?
02:02 ghiu: can i load a file
02:03 ghiu: and execute it like racket code
02:03 ghiu: ?
02:03 ghiu: nevermind
02:03 ghiu: http://docs.racket-lang.org/guide/reflection.html?q=continuations
02:03 jonrafkind: yea eval is there
02:04 jonrafkind: often times eval is not the right tool, but its handy occasionally
02:10 Demosthenes: hmmm, racket/load...
02:11 ghiu: yes, i've seen it
02:11 ghiu: i'm thinking if it would be the right approach to templating
02:13 Demosthenes: i missed Template::Toolkit
02:13 Demosthenes: but there's a string-template lib on planet
02:14 ghiu: i've seen it
02:14 ghiu: but where is the documentation for it?
02:23 Demosthenes: plant link
02:23 Demosthenes: er, planet
02:48 (quit) dnolen: Quit: dnolen
03:38 (quit) jonrafkind: Read error: Operation timed out
03:39 (quit) ghiu: Quit: ghiu
04:38 (join) tfb
05:24 (join) lucian_
05:26 (quit) lucian: Ping timeout: 272 seconds
05:44 (join) lucian
05:45 (quit) lucian_: Ping timeout: 240 seconds
05:46 (quit) lucian: Read error: Operation timed out
05:50 (quit) tfb: Quit: gone
05:50 (join) jesusito
05:51 (join) tfb
05:53 (join) lucian
05:58 (join) tfb_
05:59 (quit) tfb: Ping timeout: 276 seconds
06:21 (nick) tfb_ -> tfb
06:37 (quit) jesusito: Quit: ERC Version 5.3 (IRC client for Emacs)
07:00 (quit) lucian: Ping timeout: 240 seconds
07:02 (join) lucian
07:08 (join) mye
07:18 (join) masm
07:46 (quit) mye: Read error: Connection reset by peer
07:59 (join) lucian_
08:00 (quit) lucian: Ping timeout: 240 seconds
08:01 (join) lucian
08:03 (quit) lucian_: Ping timeout: 240 seconds
08:09 (join) clklein_
08:10 (join) tauntaun
08:13 (quit) clklein: Ping timeout: 246 seconds
08:30 (quit) masm: Quit: Leaving.
08:40 (join) PLT_Notify
08:40 PLT_Notify: racket: master Matthew Flatt * ce82c74 (1 files in 1 dirs): win32: fix ALT-space to pop up system menu ... - http://bit.ly/egsolY
08:40 (part) PLT_Notify
08:53 (join) dnolen
09:17 (join) mceier
09:17 (quit) tauntaun: Ping timeout: 240 seconds
09:31 (join) anRch
10:51 (join) PLT_Notify
10:51 PLT_Notify: racket: master Matthew Flatt * 9cc7333 (3 files in 1 dirs): fix `in-vector' error reporting ...
10:51 PLT_Notify: racket: master Matthew Flatt * 9325fe0 (1 files in 1 dirs): fix `peek-string-evt' and related docs ...
10:51 PLT_Notify: racket: master Matthew Flatt * b405637 (4 files in 4 dirs): generalizations to `subprocess' & company ...
10:51 PLT_Notify: racket: master Matthew Flatt * 5754269 (1 files in 1 dirs): win32: fix some key events ...
10:51 PLT_Notify: racket: master Matthew Flatt * 010e6fc (1 files in 1 dirs): racket/system: fix argument checking with 'exact
10:51 PLT_Notify: racket: master Matthew Flatt * c8acebc (1 files in 1 dirs): fix race in subprocess test
10:51 PLT_Notify: racket: master commits ce82c74...c8acebc - http://bit.ly/gWvKZx
10:51 (part) PLT_Notify
11:15 (join) molbdnilo
11:22 (quit) anRch: Quit: anRch
12:01 danking: Has anyone submitted a patch to pygments to get Racket syntax highlighting into github?
12:25 (join) jonrafkind
12:42 (join) MayDaniel
12:53 (join) masm
12:56 (quit) molbdnilo: Quit: tubby bye-bye!
13:01 (quit) coyo: Quit: KVIrc 4.0.2 Insomnia http://www.kvirc.net/
13:07 (join) Agari
13:12 (quit) Agari: Disconnected by services
13:12 (join) Agari
13:14 (join) scourtier
13:16 scourtier: How do I get a racket REPL working in emacs under windows? Pointing scheme-programming-name to my Racket executable doesn't seem to work; I get a "Spawning child process: Invalid argument" error.
13:17 scourtier: Oops. scheme-program-name, rather.
13:18 bremner: scourtier: I've no idea if geiser works in windows, but that is what i use for a racket repl in emacs
13:19 scourtier: Hmm... Maybe I'll give that a try. I was trying to use Quack. Thanks.
13:22 (quit) tfb: Quit: sleeping
13:37 scourtier: bremner: geiser plays nice! Thanks! I've used linux for so long I forgot what it was like for every stupid little thing to just not work consistently.
13:37 bremner: heh. Funny people coming from windows say the same thing ;)
13:38 scourtier: Ha! I'm sure.
13:38 (quit) noddy: Remote host closed the connection
13:39 scourtier: I guess if you were used to using Visual Studio for everything, unix tools would feel similarly awkward.
13:39 bremner: yup
13:40 (join) rolando
13:41 scourtier: Still, stuff is just wonky. The other day, Windows gave me this freaky virus notification. When I looked into it, it thought I had a virus because a.out crashed!
13:42 scourtier: I feel like Windows doesn't want me coding things (outside of Visual Basic)
13:42 scourtier: Anyway, I'm off. Thanks again!
13:43 (part) scourtier: "ERC Version 5.3 (IRC client for Emacs)"
13:44 rolando: make install has been running for about 19 hours
13:44 rolando: I'm assuming that it isn't normal
13:44 rolando: can somebody please tell me how long it usually takes to run make install?
13:45 rolando: also it's always doing IO to the hard drive
13:45 rolando: usually just read from the disk
13:46 rolando: now it's at "raco setup: running: web-server/scribblings/web-server.scrbl"
13:47 bremner: rolando: on my quad core i7, it takes about 15 minutes
13:47 bremner: it sounds like you are swapping badly
13:47 rolando: ok thanks
13:48 rolando: yeah, it's possible
13:48 rolando: I have an extremely low amount of ram (for todays standards that is)
13:48 rolando: 256 mb
13:48 rolando: but racket is the first program that seems to have problems dealing with that
13:48 bremner: yeah, that is pretty small, even to run racket.
13:49 rolando: ok thanks
13:49 rolando: I'll probably cancel the command
13:50 rolando: it sound like its going to ruin my hard drive
13:50 bremner: what OS? maybe somebody has binaries.
13:50 rolando: ubuntu 7. something
13:50 rolando: I think I saw some binaries on the site
13:50 rolando: I just like having the source
13:50 rolando: sometimes it's useful
13:50 bremner: sure.
13:55 rolando: that's odd, neither "racket" nor "plt-scheme" shows up in the repository
13:55 bremner: yeah, we're working on it.
13:55 rolando: oh ok
13:56 bremner: jonrafkind has an Ubuntu PPA, and there are packages on racket-lang.org
13:56 rolando: I'll just try racket out on one of the college computer's
13:56 rolando: they have about 4 gb of ram, so it shouldn't be a problem
13:56 jonrafkind: http://www.mail-archive.com/users@racket-lang.org/msg03954.html
13:57 bremner: but those most likely won't work with ubuntu 7.x
13:57 jonrafkind: oh yea they definately wont
13:57 rolando: thanks jonrafkind
13:57 jonrafkind: you need at least 9.04 I think
13:57 rolando: but I'll just try it out on another computer
13:57 jonrafkind: your school is running ubuntu 7??
13:57 rolando: nah, I'm the one running it
13:58 rolando: I really need to upgrade it
13:58 rolando: maybe to debian
13:58 jonrafkind: im still on 9.10 and i get a crap load of flak from ubuntu nerds about it
13:58 bremner: rolando: another 256M of memory would really help ;)
13:58 rolando: bremner: it's on my todo list :)
13:58 rolando: see if I get 1 gb of ram or so
13:59 bremner: yeah, racket should work pretty nicely in 1G
13:59 rolando: I used to have 512mb but one ram broke or something
14:00 (join) noddy
14:06 (quit) rolando: Quit: leaving
14:20 (quit) lucian: Ping timeout: 240 seconds
14:22 (quit) offby1`: Read error: Connection reset by peer
14:23 (join) offby1
14:27 (join) lucian
14:46 (quit) dnolen: Ping timeout: 276 seconds
16:02 (join) mithos28
16:17 (join) snaffu
16:26 (join) dnolen
16:32 (quit) snaffu: Ping timeout: 240 seconds
16:45 (join) snaffu
16:47 (join) jesusito
17:15 (join) PLT_Notify
17:15 PLT_Notify: racket: master Robby Findler * 8ae72bc (3 files in 1 dirs): added preference for the test coverage colors in the module language - http://bit.ly/dQkCQb
17:15 (part) PLT_Notify
17:28 (join) Hydrant_
17:31 Hydrant_: hey all
17:32 Hydrant_: I'm starting to get the hang of racket, I don't yet see this promised elegance... but I'm starting to get hold of the basic ideas I think
17:35 (quit) snaffu: Ping timeout: 276 seconds
17:40 Lajla: Hydrant_, define 'elegance'
17:40 Hydrant_: is there a way for me to examine the type of something, so that I can get a feel for how racket sees variables
17:40 Lajla: When they say 'elegant', they usually mean 'simple' as in 'a few simple rules can explain the entire language', 'no exceptions' et cetera.
17:41 Lajla: rudybot, (string? "string")
17:41 rudybot: Lajla: ; Value: #t
17:41 Hydrant_: right, I saw that I can do queries
17:41 Lajla: rudybot, (char? #\l)
17:41 rudybot: Lajla: ; Value: #t
17:41 Hydrant_: but is there something like (type x) that would say 'String or something
17:41 Lajla: Hydrant_, maybe you can in some limited sense.
17:41 Lajla: But an element can be part of multiple types
17:41 Lajla: I mean 1, it's a number, but also an integer, all integers are numbers, and reals.
17:42 Lajla: rudybot, (and (pair? '(1 2 3)) (list? '(1 2 3 4)))
17:42 rudybot: Lajla: ; Value: #t
17:42 Lajla: All lists except the empty list are also pairs.
17:42 Lajla: Point you have to ask yourself is 'what is a type'?
17:42 Hydrant_: okay, what I wanted to understand is how racket sees: '(define x 5)
17:43 Hydrant_: and other expressions I'm building up, I hoped to see myself what they are
17:44 Lajla: rudybot, (define (list-of-characters? l) (or (null? l) (and (pair? l) (char? (car l)) (list-of-characters? (cdr l)))))
17:44 rudybot: Lajla: Done.
17:44 Lajla: Have I just defined a type or not?
17:45 Lajla: Hydrant_, define forms are pretty unique expressions in that they can't just appear at any point in the source code
17:45 Lajla: They are more 'declarations'
17:46 Hydrant_: yes, I get that
17:46 Lajla: But basically 'types', are a pretty vague concept, in the end, they are properties of objects in the same way that the fact that a string starts with an uppercase letter is a property of it.
17:46 Lajla: I could easily define a function which checks for this, have I then defined a type?
17:47 Hydrant_: my current working definition is that a define is a somewhat global binding of a new thing, and let is introducing something new into limited scope
17:47 Lajla: That's the idea yeah.
17:47 Lajla: Let introduces a local block
17:47 Lajla: the block is more important than the binding
17:47 Lajla: define declares a global binding.
17:47 Hydrant_: I don't understand how types are so vague to you... almost all stuff I do in CS is reasoning of types
17:47 Hydrant_: this is a tree... this is a node... this is an edge, and so on
17:48 Lajla: You can also do (define foo) in R6RS, which just establishes the binding, kind of like unitizialized variables.
17:48 Hydrant_: it is odd that I don't have types here, because it feels very unnatural to me
17:48 Lajla: You have types just like you have it in any other place.
17:48 Hydrant_: I can understand type only being valid in runtime
17:48 Hydrant_: what I'd like to do... is practice building up expressions, and investigate what each piece is
17:48 Lajla: Types are a set of objects that all satisfy a common list of properties
17:48 Hydrant_: so... when I see something that is odd to me... like '(1 2 . 3) I can investigate what that is
17:49 Hydrant_: similar to Python, when I get some result I want to see what's in there
17:49 Lajla: Well, it is amongst others a pair, an improper list, a datum, a non-callable value
17:49 Lajla: The system has no way of guessing what you want.
17:50 Lajla: It's generally impossible to find out what the 'type' of a datum is in any language, you can only once you have a type check if a datum is in that type.
17:50 Lajla: Because each datum basically is a member of countably infinite types.
17:50 Hydrant_: this is going pretty far into the abstrat
17:50 Hydrant_: *abstract
17:52 Lajla: Why?
17:52 Lajla: I mean, you understand that each object is a member of an infinite amount of different types, yes?
17:52 Lajla: If I have this "hello"
17:53 Lajla: this is amongst others: a string, a sequence, an immutable object, something which eq?  cannot reliably compare and so on.
17:53 Lajla: THese are all types.
17:53 Lajla: I could go on forever basically finding a new type to place it in.
17:54 Hydrant_: yes, this is true on a blackboard but not how languages generally work
17:54 Hydrant_: at a low-level, yes, typing doesn't exist... it's just globs of bits
17:55 Hydrant_: that could be anything that can be input to a turing machine based on the encoding
17:55 Lajla: One of the fine parts of the scheme spec is that it usually doesn't define a low level representation implementations have to use.
17:56 Lajla: As long as the described behaviour is accomplished it is fine.
17:56 Lajla: Hydrant_, well, how do you query a type in python?
17:56 Hydrant_: type()
17:57 Lajla: Okay, and say you do type("John") what does it return?
17:58 Hydrant_: <type 'str'>
17:58 Hydrant_: because Python is going to implicitly take "..." and convert to a string object, with defined characteristics
17:58 Lajla: Why doesn't it return <type 'str-which-starts-with-uppercase-letter'>
17:58 Lajla: or <type 'sequence'>
17:58 Lajla: ?
17:58 (join) cderici
17:59 Lajla: Or <type 'uncallable object'> ?
17:59 Hydrant_: if I decide I want a list of chars, I can do that by changing the  type myself via list("John")
17:59 Hydrant_: because there is a single mapping of "..." defined... it must be a single type
17:59 Hydrant_: if I want to change the type, I must move it via a function call to another single mapping myself
17:59 Lajla: Well
17:59 Lajla: then types in python are a label
17:59 Lajla: That each object carries
17:59 Lajla: right?
18:00 Hydrant_: yes, I would agree with that because this is also even true in the realm of turing machines
18:00 Lajla: You can basically just give each object a .type variable and house a string in it which names the type.
18:00 Hydrant_: types are just things interpreted by humans
18:00 Lajla: I agree
18:00 Lajla: Well, in Scheme, this is not always true.
18:00 Lajla: Especially the list? "type"
18:01 Lajla: A 'type' is a property an item satisfies.
18:01 Lajla: There are subtypes and supertypes, union types and so on.
18:01 Lajla: So necessarily, each object is a member of multiple types.
18:01 Lajla: Hydrant_, ever heard op 'top type'?
18:02 Hydrant_: I don't agree that type is a property satisfied by an item
18:02 Hydrant_: no, haven't heard of it
18:02 Hydrant_: but this is drifting, I think, to a bizarre philosophical realm
18:02 Hydrant_: are you saying that racket has zero typing internally
18:02 Lajla: Nope, top type is pretty fundamental in type theory.
18:02 Lajla: Not at all
18:03 Lajla: some types are indeed managed internally by labels.
18:03 Lajla: But the top type
18:03 Lajla: is the type that contains every element
18:03 Lajla: THe bottom type is the type that contains no element
18:03 Lajla: THere are supertypes and subtypes in type theory.
18:03 Lajla: union types, intersection types.
18:03 Lajla: Types are basically in static type theory sets in which expressions can be placed.
18:04 Lajla: It's like asking a random object 'in what set are you?', you can't, there are infinite possible sets it is in.
18:04 Lajla: The point is that python, I assume, does not support subtypes and supertypes and union types and top tpyes and bottom types.
18:04 Lajla: And racket does.
18:04 Lajla: In racket, an object can (and always is, basically) in multiple types.
18:05 Hydrant_: okay
18:05 Lajla: Or 'an expression' I should say
18:05 Hydrant_: by type theory do you mean category thoery ?
18:06 Lajla: Nah, that's a different thing.
18:07 (quit) cderici: Quit: WeeChat 0.3.2
18:07 Lajla: Also, dynamic types and static types are basically a completely different and fundamentally unrelated thing in the end.
18:07 Hydrant_: well, I would imagine static types can be optimized out
18:07 Hydrant_: in the end they are just bits
18:07 Lajla: dynamic types are properties of objects, static types are categorizations of expressions.
18:07 Hydrant_: it's just clumsy humans that need types
18:07 Lajla: Yeah
18:08 (join) tauntaun
18:08 Lajla: in the end its practical asset is that it prevents you from screwing things up.
18:08 Hydrant_: static types are going to require properties, because a human put themselves in the middle
18:08 Hydrant_: *dynamic
18:08 Lajla: But Hydrant_ take this "3 + [0,1]"
18:08 Lajla: I added a number to an interval, a mathematical expression
18:08 Lajla: surely this is ill-typed, right?
18:08 Hydrant_: depends on how it was defined by the human
18:08 Hydrant_: I wouldn't say that without more information
18:08 Lajla: Good
18:08 Hydrant_: if + is not a valid operator, then it's not defined
18:09 Lajla: It basically depends on how you see it, usually in foundational mathematics, 3 is defined as some set
18:09 Lajla: so is [0,1], and + is then also defined as an operation on sets.
18:09 Lajla: Basically, if you work it out on paper
18:09 Hydrant_: yah.... representing numbers as sets is funny ;-)
18:09 Lajla: 3 + [0,1] will yield some set, if you work it out mechanically.
18:09 Lajla: It's just that it doesn't really make sense to human beings.
18:12 Hydrant_: Can you show me a way of defining +, so that it is defined on integers, vectors / matrices, complex numbers, and other members of a set without enumerating every valid possibility?
18:12 Hydrant_: as you said, each type belongs to several sets, so there must be a way of specifying which translations are legal
18:13 (join) PLT_Notify
18:13 PLT_Notify: racket: master Matthew Flatt * 85951bf (5 files in 3 dirs): gtk: smoother editor-canvas resize ... - http://bit.ly/ejPCAC
18:13 (part) PLT_Notify
18:14 Lajla: Hydrant_, well, you mean + for adding natural numbers? :')
18:14 Lajla: Let's just recursively define it like:
18:15 Hydrant_: I didn't say specify a field
18:15 Lajla: x + {} := {x} union {y : y in x}
18:15 Lajla: x + S(y) := S(y+x)
18:16 (quit) mithos28: Quit: mithos28
18:16 Lajla: ehhh
18:16 Lajla: No
18:16 Lajla: x + {} := {}
18:16 Lajla: of course
18:16 Lajla: I was confused
18:16 Lajla: S(x) := {x} union {y : y in x}
18:16 Hydrant_: my point is that without typing and treating each type as one value at a time, things are rather hopeless
18:16 Lajla: Well, you can still go right
18:16 Lajla: there's just nothing protecting you from going wrong
18:16 Lajla: x + {} := x
18:16 Lajla: goddamnit
18:17 Hydrant_: yes, you might have each type interpretable in different manners (as you say, many possible types for any object) but the definitions are still going to be one-to-one
18:18 Lajla: It's not about interpretable at all.
18:18 Lajla: For instance
18:18 Lajla: Say you have (if cond then else)
18:18 Lajla: Basically, it evaluates to then if cond is true, else otherwise.
18:18 Lajla: Now, you might have been told that then and else need to have the same type for this to be well-typed
18:18 Hydrant_: nope, I wasn't
18:18 Lajla: And that is also the type of the entire expression, right?
18:18 Hydrant_: :-)
18:18 Lajla: Ah, okay.
18:18 Lajla: Well, this is how it works in C++
18:19 Hydrant_: well, there has to be a common ancestor in an inheritance graph
18:19 Lajla: In reality, this is not how it works in theory, for it to be well typed all that needs to happen is that the type of the entire expression is a supertype of the union of the type of then and else.
18:19 Hydrant_: and what is to be returned must be that common ancestor
18:19 Lajla: Buut
18:19 Lajla: since C++ does not support union types and supertypes and sub types
18:19 Lajla: It basically comes down to that
18:20 Hydrant_: it supports inheritance
18:20 Lajla: Basically, in C++ or python, stuff is a lot simpler because there are no supertypes, union types and all that bla.
18:20 Hydrant_: and that is what you describe
18:20 Lajla: Well, tha's not a supertype.
18:20 Lajla: Nononono.
18:20 Hydrant_: or... more precisely perhaps if you don't want a common ancestor you must only define casting rules from each to each
18:21 Hydrant_: so... you could do T x = cond ? then : else;... so long as then and else are castable to T
18:21 Hydrant_: I'm not sure if the compiler will follow a path to get there, but it might
18:21 Lajla: There should be a death penalty on languages which have implicit casting.
18:21 Lajla: Yeah
18:21 Lajla: but that's basically just syntactic sugar
18:21 Lajla: It automatically fills in the cast
18:21 Lajla: A cast is basically a function
18:22 Lajla: Say you cast a number to a string.
18:22 Lajla: That's just what number->string does in racket
18:22 Hydrant_: right, but this is valid... because I see no difference in having a "set of types" where you can cast some member to any other member, from what C++ can do with implicit casts
18:22 Lajla: Sure
18:23 (join) mithos28
18:23 Lajla: But I'm just saying that it' not entirely cond ? then : else ;
18:23 Lajla: It implicitly adds the cast
18:23 Lajla: it's syntactic sugar
18:23 Lajla: So the actual arguments of the ? : operators are still going to be type T
18:23 Hydrant_: yes, it does add the cast... although I'm not sure how this changes if we consider this expression: auto x = cond ? then : else;
18:24 Hydrant_: auto has been added, and I'm not sure what would happen here
18:24 Hydrant_: because auto stands for 'generic type for compiler to figure out'
18:25 Hydrant_: also... the arguments to ?: will be cast to type T... but certainly are not that type in the beginning... there is simply a valid path to get to T
18:27 Lajla: Well, they aren't the arguments, technially.
18:27 Lajla: Say that (cast) is the casting function
18:27 Lajla: The argument is (cast) then
18:27 Lajla: Not then
18:27 Lajla: And the type of (cast) then is indeed T
18:28 Hydrant_: you mean that there is a single initial type
18:28 Lajla: I'm not quite sure of what you speak.
18:38 Hydrant_: do you know how in drracket to make the levels of () have different colors?  I've seen others do this but can't find the option
18:39 Hydrant_: nm, found it
18:47 Lajla: Hydrant_, where is it?
18:47 Lajla: I'd love to have it too.
18:48 Hydrant_: edit->preferences->colors->parenthesis color scheme
18:49 Lajla: kewl
18:49 Lajla: thank j00
18:50 Hydrant_: np
18:50 Hydrant_: glad I could teach you something about racket
18:50 Lajla: I am completely and utterly ignorant about racket.
18:51 Lajla: I just know some stuff abou type theory.
18:56 tauntaun: We would be honored, if you would join us.
19:04 (join) snaffu
19:05 Lajla: tauntaun, that against me?
19:08 tauntaun: Now who would hold anything against you, Lajla.
19:09 (quit) snaffu: Quit: Lost terminal
19:11 (quit) MayDaniel: Read error: Connection reset by peer
19:12 (quit) Agari: Ping timeout: 240 seconds
19:12 (quit) mceier: Quit: leaving
19:32 Lajla: tauntaun, no, I mean addressed to me.
19:34 (join) Agari
19:44 (join) askhader_
19:46 (quit) tauntaun: Quit: Ex-Chat
19:47 (quit) askhader: Ping timeout: 240 seconds
19:52 (join) gearheadza
19:56 (quit) mithos28: Quit: mithos28
19:57 (join) mithos28
20:01 (quit) Agari: Quit: Quit
20:04 Hydrant_: everytime I search on google for something about racket I get tennis
20:07 danking: Hydrant_: http://www.google.com/search?sourceid=chrome&ie=UTF-8&q=racket ?
20:10 Lajla: danking, google personalizes result.
20:10 Lajla: s
20:10 Lajla: Like, when I google 'girl', you don't want to know what I get.
20:11 askhader_: Lajla: Well I do.
20:11 danking: Google'ing through a proxy and when signed out I get the same top result
20:11 danking: In fact racket appears twice.
20:12 danking: First `racket-lang.org' and then the docs page.
20:12 (join) tauntaun
20:12 Lajla: Maybe Hydrant_ likes Tennis
20:12 Lajla: askhader_, I was joking
20:12 Lajla: you see
20:12 Lajla: I am a very witty man
20:12 tauntaun: I need a hug.
20:12 danking: Perhaps
20:13 danking: Either way, I find `Racket' gets me better results than `Scheme' or `PLT Scheme' ever did.
20:14 Lajla: Why do people always `quote stuff like this'
20:14 Lajla: Yeah, I know where it came from.
20:14 Lajla: But I don't know, it's just awkward, it was a hack in those days and it still is.
20:14 Lajla: tauntaun, it's like oxygen, isn't it?
20:14 tauntaun: ?
20:15 Lajla: tauntaun, hugs
20:16 danking: It's a cultural thing I guess.  I started using it when writing the names of identifiers and then it bled into anything I would ordinarily quote.
20:16 danking: I still use the double quote character " outside of IRC.
20:16 Lajla: Jumping on the bandwagon eh
20:16 Lajla: However, I shall remain defiant.
20:16 danking: Common syntax is handy for communication.
20:16 Lajla: Conformists
20:16 Hydrant_: google: racket match one or more
20:17 Lajla: I think 'quoting like this' is okay enough.
20:17 Lajla: French quotes make a lot more sense
20:17 Lajla: Because they «open and thereafter close».
20:17 danking: I've been chastized `for this` style.
20:17 Lajla: Well, so do English quotes, but not in Ascci
20:17 Lajla: Hmm
20:17 Lajla: that always reminds me of Haskell.
20:17 tauntaun: Lajla: the Haskell compiler?
20:18 Lajla: `div`
20:18 tauntaun: (+)
20:18 Lajla: You can infix stuff in Haskell by putting `...` around it.
20:18 Lajla: Yeah, the reverse of that.
20:18 danking: If I had my way '(quotes would be awesome)....
20:18 tauntaun: `elem`
20:18 Lajla: yeah
20:18 tauntaun: If Haskell were huggable, I would hug it and it would hug me back.
20:18 Lajla: I would just use € for element really
20:18 Lajla: I loathe haskell.
20:18 tauntaun: That's "euro", genius.
20:19 danking: lol
20:19 tauntaun: Haskell loathes you back, you shameless ingrate.
20:19 tauntaun: Do you need a hug, Lajla?
20:19 Lajla: Simon Peyton-Jones has a fat mother.
20:19 Lajla: Nahhh
20:19 Lajla: I am cool
20:19 Lajla: becaaause
20:19 tauntaun: That's a low blow.
20:19 Lajla: I just won a long macro game in SC2.
20:19 tauntaun: You make "fat" sound like a bad thing.
20:19 Lajla: Versus a zerg.
20:19 danking: Peyton-Jones is wicked cool.
20:20 Lajla: Isn't his wife like a member of the church of England
20:20 Lajla: I do not approve of state religions
20:20 Lajla: ळागूNeither did the founding fathers
20:20 tauntaun: You Dutch are always so friggin tolerant.
20:20 danking: He talked here at NU about the parallelizing stuff they're doing with GHC.  I hope to convert my physics professors into haskell guys given that parallelizing is all the rage over there now.
20:21 tauntaun: danking: Good luck.
20:21 danking: :P
20:21 Lajla: It amazes me how many physics people actually use C++
20:21 danking: "WHAT?!  No pointer access!?  It must be slower than UNIVAC"
20:21 Lajla: Oh hai there.
20:21 tauntaun: My dissertation was in physics, involved "parallel" simulations, and there was no way to talk to that community about principled PL.
20:21 Lajla: C++ doesn't have a type system that does dimensional analysis for you.
20:22 Lajla: My dessert was chocolate pudding.
20:22 tauntaun: Mine was whole milk with 2 tbsps malt, 2 tbsps sugar, and several drops of vanilla.
20:22 tauntaun: Fresh, fresh, baby.
20:23 Lajla: I don't know
20:23 danking: tauntaun: I was working with a physics prof and ended up at some seminar.  After the seminar one of the EE/CE professors was deriding Fortran for lacking pointers with a terribly smug demeanor.
20:23 Lajla: I just think that the thing about Haskell
20:23 Lajla: Like, pattern destructuring
20:23 Lajla: what you can and cannot destructure is so arbitrary
20:23 Lajla: I'd rather just destructure manually
20:24 tauntaun: danking: there are many reasons to deride Fortran, but its lack of pointers is probably not one of them.
20:24 Lajla: let x = (head argument), xs = (tail argument) in ....
20:24 tauntaun: Lajla: destruct *this*.
20:24 danking: What struck me was his compelte intolerance of anything higher-level than C.
20:24 Lajla: tauntaun, destructure
20:24 Lajla: you mean
20:25 Lajla: A lot of physicists think they can program, little do they know they aren't as good as I.
20:25 Lajla: I had a lot of debates about programming with them.
20:25 tauntaun: Lajla: no, actually "destruct" is sometimes used in the context of pattern matching.
20:25 Hydrant_: I don't know of any language that can beat FORTRAN for matrix ops...
20:25 Lajla: Also often people that think they care about performance, but don't.
20:25 Lajla: Like, the kind of people that think they optimize, but don't bother to actually empirically test out if it's faster.
20:25 Lajla: And don't realize that every fucking compiler on the planet optimizes that for them.
20:25 danking: Lajla: In my experience, the Physics profs/grads are intimately aware of how little the know about programming.  That's how I got hired :D
20:25 Lajla: Or the people that think that writing assembly produces faster stuff.
20:25 tauntaun: Hydrant_: a language is only as good as its compiler, and I wouldn't be surprised if C/C++ has become faster than Fortran by now.
20:26 Hydrant_: tauntaun: no, it never will be
20:26 Lajla: danking, a thing I have not yet encountered.
20:26 tauntaun: Hydrant_: and why, praytell?
20:26 Hydrant_: afaik fortran can optimize matrix expressions, whereas C/C++ have to treat them as temporary values... and cannot optimize them
20:26 danking: Lajla: Stalin brutually optimizes, eh?
20:26 Hydrant_: linear algebra has nice optimizations that C/C++ cannot ever know
20:26 Lajla: Stalin is an awesome compiler with an awesome name.
20:27 Lajla: Yeah, that's also true
20:27 Lajla: Fortran is very good for matrix stuff
20:27 tauntaun: Hydrant_: I know not of what you speak, and in any case it surprises me.
20:27 tauntaun: My dissertation code was all in Fortran, and I detested it daily.
20:27 tauntaun: I don't want to believe that C/C++ would have been appreciably slower.
20:28 danking: Shoulda wrote it in Scheme, obviously.
20:28 tauntaun: danking: I started that, but it took too long.
20:28 tauntaun: danking: may I ask where and for whom you work as programmer?
20:29 Hydrant_: it is, C++ cannot beat fortran in benchmarks fo matrix ops
20:29 tauntaun: Hydrant_: forget C++; compare C instead.
20:29 Hydrant_: why not compare assembly then
20:30 Hydrant_: C doesn't allow you to do M*b + v
20:30 Hydrant_: and it will be worse in C, actually, because the temporary hell is even worse
20:30 tauntaun: ok, I begin to see your point, which seems to be that Fortran supports array expressions (like M*b + v), yes?
20:30 Hydrant_: vector_add(matrix_vector_mult(M,b), v)... each call requires a temporary
20:30 Hydrant_: you can't pipeline the operations
20:31 Hydrant_: yes, so far as I know matrix ops can be optimized directly
20:31 Hydrant_: so that a single stream is done, whereas in C or C++ temporaries are introduced to evaluate the expressions
20:31 Hydrant_: even with no temporaries, C/C++ will get cache-locality wrong if it doesn't optimize the operation into a nice single-pass form
20:32 tauntaun: Hydrant_: well that's nice.  In that case I would recommend writing the code in C/C++/Scheme/whatever, and then using an FFI to Fortran for the actual matrix operations.
20:33 tauntaun: What I can't stand is writing the *entire* system in god-awful Fortran.
20:34 Lajla: I would think scheme is good.
20:34 Lajla: I think the teaching of C++ cripples the mind, and should therefore be regarded a criminal offence.
20:34 tauntaun: Thank you for understating your position.
20:34 Lajla: It is impossible to teach good programming practices to epople who have any experience with C++, their mind is mutilated beyond hope.
20:35 tauntaun: Lajla: your point is not coming across.  More rhetoric please.
20:35 Lajla: I'm just impersonating dijkstra
20:35 Lajla: swapping basic with C++
20:36 Hydrant_: ugh, so much zealotness in scheme
20:36 Hydrant_: it is a real turn off
20:37 lucian: Hydrant_: if it makes you feel any better i also hate C++ and i'm primarily a python programmer :)
20:37 Hydrant_: I'm a C++ expert
20:37 Hydrant_: C++ is what I use for almost everything, and most of what I do is HPC
20:37 Hydrant_: with physicists / engineers
20:37 tauntaun: Hydrant_: it's "zealotry."
20:38 (quit) mithos28: Quit: mithos28
20:38 tauntaun: This brings up an interesting question: What are the chances that Racket would provide a Fortran FFI?
20:39 Lajla: I my life for aiur.
20:39 Lajla: I refuse to believe that anyone on the planet is a C++ expert, not even Bjarne, the language is too large for that.
20:39 Lajla: Same with racket really.
20:40 Lajla: Too large to know it all.
20:41 lucian: Lajla: racket is much smaller
20:41 tauntaun: jonrafkind: you available for a question?
20:41 danking: tauntaun: I'm a student at Northeastern University.  I was working for Professor Karma here on some GPU stuff he needed.
20:41 jonrafkind: sure
20:42 tauntaun: danking: thanks, sounds great.
20:42 tauntaun: jonrafkind: how hard would it be to provide a Fortran FFI in Racket.
20:43 jonrafkind: in theory not very hard, i mean i think fortran just produces symbols with some simple calling convention (probably C like)
20:43 danking: tauntaun: It's pretty cool.  I feel the heat building now though.  Gotta shape up so I can get into a decent grad school.
20:43 Lajla: You never answer my quaestions
20:53 (quit) gearheadza: Ping timeout: 240 seconds
20:54 (join) mithos28
20:56 (quit) mithos28: Client Quit
20:57 (join) mithos28
20:59 (quit) mithos28: Client Quit
21:03 (quit) Hydrant_: Remote host closed the connection
21:12 (join) mithos28
21:17 (quit) tauntaun: Quit: Ex-Chat
21:21 (quit) mithos28: Quit: mithos28
21:21 (join) jfalcon
21:22 jfalcon: can i make a set which determines equality based on an arbitrary function?
21:27 Lajla: jfalcon, I find your lack of clarity disturbing.
21:27 Lajla: Ohhh
21:27 Lajla: you mean in that way
21:27 (part) jesusito: "ERC Version 5.3 (IRC client for Emacs)"
21:27 jfalcon: ...
21:27 jfalcon: 5.5 Structure Comparisons. That should achieve what I'm trying to do.
21:28 jfalcon: thanks.
21:28 jfalcon: http://docs.racket-lang.org/guide/define-struct.html#(part._struct-equal)
21:28 (join) MelanomaSky
21:32 (join) tauntaun
21:40 (join) snaffu
21:43 (quit) askhader_: Remote host closed the connection
21:46 (join) PLT_Notify
21:46 PLT_Notify: racket: master Matthew Flatt * 9787028 (1 files in 1 dirs): gtk: keep popup menus on the screen ... - http://bit.ly/g8RCHG
21:46 (part) PLT_Notify
22:01 (join) mithos28
22:03 (quit) snaffu: Quit: Lost terminal
22:05 (join) corruptmemory
22:16 (quit) mithos28: Quit: mithos28
22:18 (quit) winxordie: Quit: Leaving.
22:28 (nick) tauntaun -> anthonywest
22:28 (nick) anthonywest -> awest
22:29 (nick) awest -> anthonywest
22:30 (part) anthonywest: "Ex-Chat"
22:32 (join) tauntaun
22:32 (quit) tauntaun: Quit: Ex-Chat
22:32 (quit) jfalcon: Quit: jfalcon
22:52 (join) dherman
22:52 dherman: got a scribble question if anyone has a sec
22:57 jonrafkind: irc etiquette says to ask your questino, not to ask to ask
22:58 dherman: right then
22:59 dherman: curious if there's any scribble forms that render in html as <dl> / <dt> / <dd>
22:59 dherman: otherwise I'll just have to use table hacks
23:00 jonrafkind: can't say ive ever seen those being rendered
23:19 dherman: does quack work well for racket?
23:26 dherman answers self in the affirmative
23:46 offby1: for interactive use, geiser works very well
23:48 (quit) evhan: Ping timeout: 240 seconds
23:50 (join) evhan
23:57 dherman: offby1: nah, just want a mode for editing
23:57 dherman: quack's doing fine
23:57 (quit) jonrafkind: Ping timeout: 240 seconds
23:58 (quit) dherman: Quit: dherman