00:41 (join) neilv
00:41 (quit) neilv: Changing host
00:41 (join) neilv
01:17 jonrafkind: how do you insert a line break in scribble?
01:18 Lajla: You should insert it before or after it.
01:18 Lajla: Not in between
01:18 Lajla: that would be infix.
01:22 jonrafkind: ah linebreak
01:28 (quit) parcs: Read error: Connection reset by peer
01:29 (join) parcs
01:29 jonrafkind: well that didnt work
01:29 jonrafkind: another wonderful error!
01:30 (quit) parcs: Read error: Connection reset by peer
01:31 (join) parcs
01:59 (quit) mheld: Quit: mheld
03:35 (quit) Yann1: Read error: Operation timed out
03:36 (join) Yann1
03:40 (quit) jonrafkind: Ping timeout: 250 seconds
04:27 (quit) coldhead: Read error: Operation timed out
04:28 (join) coldhead
05:06 (quit) parcs: Ping timeout: 240 seconds
05:07 (join) parcs
05:39 (quit) neilv: Quit: Leaving
05:55 (join) tv|z
05:56 (part) tv|z
07:21 (join) neilv
07:21 (quit) neilv: Changing host
07:21 (join) neilv
07:22 neilv: looks like, in the scribble emacs mode, the eldoc help can't always fit practically: http://i.imgur.com/tSydj.png
07:23 (quit) _danb_: Ping timeout: 260 seconds
08:13 (join) mceier
08:24 neilv: i think, for eldoc, i'll have to do it something like this screenshot. http://i.imgur.com/stNpF.png
08:24 neilv: there will be an explicit command for showing quick-reference help that includes the name, type, and default value of each argument
08:39 (quit) certainty: Read error: Connection reset by peer
08:46 (quit) parcs: Read error: Connection reset by peer
08:50 (join) MayDaniel
08:50 (join) parcs
09:11 neilv: and the final version is the least flashy-looking. :) http://i.imgur.com/hgwEt.png
09:46 (quit) MayDaniel: Read error: Connection reset by peer
10:04 (join) jonrafkind
10:10 Tasser: what's out there as reference? One where you lookup a function name based on what it should do...
10:13 jonrafkind: well the docs are at docs.racket-lang.org
10:13 jonrafkind: if you want to google search it use the site: thing that google provides
10:36 Tasser: is there a kind of drracket plugin for vim/emacs/eclipse ?
10:38 jonrafkind: there is a mzscheme plugin for vim, and theres quack for emacs
10:38 jonrafkind: mzscheme is the old version of racket
10:39 Tasser: sounds like it's generally recommended to use drracket?
10:40 jonrafkind: yea
10:41 Tasser: well, the editor sucks ;-)
10:41 jonrafkind: you get some nice tools with drracket too, like a debugger and 'check syntax'
10:41 jonrafkind: i wrote a vim plugin for drracket if you like that sort of thing
10:41 Tasser: go on
10:41 jonrafkind: http://planet.racket-lang.org/display.ss?package=vi.plt&owner=kazzmir
10:42 jonrafkind: its not 100% compatible but I use it occasionally
10:42 Tasser: best way to install it?
10:44 jonrafkind: type (require (planet kazzmir/vi:1:11/tool)) into the drracket interactions window
10:44 jonrafkind: then restart drracket
10:45 Tasser: require: not a valid planet path; should be: (require (planet STRING (STRING STRING NUMBER NUMBER)))
10:46 jonrafkind: what version of drracket do you have
10:46 Tasser: drracket: unknown switch: --version
10:46 Tasser: hmmm
10:46 jonrafkind: anyway, try this line (require (planet "tool.ss" ("kazzmir" "vi.plt" 1 11)))
10:46 jonrafkind: thats the "old" planet format
10:46 Tasser: 5.0.2
10:46 jonrafkind: i think thats new enough
10:47 jonrafkind: very strange
10:47 Tasser: that's the newest stable
10:48 Tasser: read: illegal use of "."
10:48 Tasser: setup-plt: error: during making for <planet>/kazzmir/vi.plt/1/11 (Vi mode)
10:48 Tasser: setup-plt:   read: illegal use of "."
10:49 jonrafkind: bleh
10:51 Tasser: but seems to work
10:53 Tasser: what am I doing wrong with those 'let' ? http://pastie.org/1440280
10:54 rapacity: that doesn't look like scheme
10:54 rapacity: or racket
10:54 jonrafkind: change let to define
10:54 jonrafkind: also operators are prefix, not infix
10:54 jonrafkind: (< height width)
10:54 jonrafkind: instead of (height < width)
10:54 Tasser: duh
10:55 Tasser: Module Language: there can only be one expression in the definitions window
10:56 jonrafkind: wtf
10:56 Tasser: (fixed all the above stuff)
10:56 neilv: add a "#lang" line at the top
10:56 jonrafkind: #lang racket
10:56 Tasser: jonrafkind, your vi mode doesn't support O
10:56 jonrafkind: yea its missing a bunch of stuff
10:58 (join) MayDaniel
10:58 (quit) MayDaniel: Changing host
10:58 (join) MayDaniel
10:58 Tasser: what's the difference of (require 2htdp/image) and (require "
10:58 Tasser: 2htdp/image")
10:59 (join) mheld
10:59 jonrafkind: with quotes it means try to load a local filename
10:59 jonrafkind: without the quotes it refers to an internal library
11:00 Tasser: what's the difference of the argument data type?
11:00 jonrafkind: 2htdp/image is a symbol, "2htdp/image" is a string
11:11 Tasser: what's a symbol?
11:13 jonrafkind: its sort of like a string but its immutable
11:13 Tasser: same as a
11:13 Tasser: Symbol in Ruby?
11:13 jonrafkind: I think so
11:13 Tasser: stupid enter key
11:14 jonrafkind: usually you have to write 'f in racket to get a symbol
11:14 jonrafkind: but `require' is a macro so it can get the symbol without forcing you to write the '
11:15 Tasser: can I change a Symbol into a function?
11:15 Tasser: aka ('+ 2 3)
11:16 jonrafkind: to go from a symbol to values you need to use eval usually
11:16 jonrafkind: but generally you shouldn't do such things
11:16 Tasser: Values?
11:17 jonrafkind: like from the symbol '+ to the actual procedure +
11:17 jonrafkind: rudybot: eval '+
11:17 rudybot: jonrafkind: your sandbox is ready
11:17 rudybot: jonrafkind: ; Value: +
11:17 Tasser: so a value is somewhat of a named lambda?
11:18 jonrafkind: i guess 'value' is the wrong name here, a symbol is a value too
11:18 jonrafkind: a value is anything that can be stored in a variable, basically
11:18 Tasser: oke
11:18 Tasser: basically an object in an OO language?
11:19 jonrafkind: yea
11:20 neilv: working through a book is the best way to start to learn racket
11:20 Tasser: neilv, yeah, and spice it up with some IRC
11:24 neilv: ok, that works
11:27 Tasser: ich nehm die stadt sonst
11:28 bremner: jaja aber bitte Englisch hier
11:29 Tasser: uhm, duh. Wrong tab
11:31 offby1: rudybot: t8 de en  ich nehm die stadt sonst
11:31 rudybot: *offby1: I'll take the city otherwise
11:31 offby1: rudybot: now please explain that
11:31 rudybot: *offby1: I think bremnder is right though: "When a non-free program or system is well known, you can mention it in passing—that is harmless, since users who might want to use it probably already know about it. For instance, it is fine to explain how to build your package on top of some widely used non-free operating system, or how to use it together with some widely used non-free program."
11:32 bremner: good grief. my mispelled nick follows me around.
11:33 offby1 laughs cruelly
11:57 (quit) jonrafkind: Read error: Operation timed out
12:00 (join) Yann3
12:00 (quit) Yann1: Read error: Operation timed out
12:15 (join) anRch
12:46 (quit) coldhead: Remote host closed the connection
12:49 (join) coldhead
12:49 (quit) anRch: Quit: anRch
12:59 (join) MayDaniel_
13:02 (quit) MayDaniel: Ping timeout: 250 seconds
13:03 (join) anRch
13:10 (join) epochwolf|2
13:21 (quit) MayDaniel_: Read error: Connection reset by peer
13:46 (quit) epochwolf|2: Read error: Connection reset by peer
13:47 (join) epochwolf|2
14:03 (quit) anRch: Quit: anRch
14:03 (join) stis
14:26 (join) MayDaniel
14:52 (join) mwolfe
14:59 (quit) MayDaniel: Read error: Connection reset by peer
15:30 (quit) epochwolf|2: Ping timeout: 255 seconds
15:43 (join) masm
15:48 (join) samth
15:48 samth: clklein, ping
16:01 clklein: pong
16:03 samth: a couple redex questions
16:04 samth: (a) lack of support for consing in patterns/term constructors is v. annoying
16:05 samth: (b) you should look at this paper: http://www4.in.tum.de/~urbanc/Publications/esop-11.pdf
16:05 samth: clklein, as above
16:08 samth: i guess those weren't really questions
16:11 clklein: samth: (b) looks interesting from the abstract. One of the things that stopped me from trying to improve Redex's support for binders is that the usually solutions don't seem all that convenient for anything except lambda-like binders.
16:13 samth: their system is apparaently similar to what Ott does
16:13 samth: it passes my test, which is that it works for `letwreck'
16:13 samth: which i suppose we've now actually added to the core in the form of `->'
16:14 samth: rather `->i'
16:14 samth: also, i want to specify multiple #:source arguments to redex-check
16:14 clklein: I don't know letwreck. My test has been whether it can express something like `define', where the scope is "sideways"
16:17 clklein: I don't see any obstacles to extending #:source. If your sources are reduction relations, maybe you can union them
16:17 clklein: in the meantime
16:17 samth: letwreck looks like ->i, where you specify which of the variables are in scope for each clause
16:17 Tasser: someone cares to explain why (null? '()) is true?
16:17 Tasser: oh
16:17 Tasser: nevermind.
16:18 samth: clklein, it's from two metafunctions, so i can't union them
16:18 samth: what about consing?
16:20 clklein: I always encode pairs as lists of length two.
16:20 clklein: I take it you don't like that because they print kind of ugly
16:20 samth: this isn't really pairs
16:21 samth: say you have a continuation which is a list of frames
16:21 samth: and you want to push a new one on
16:21 samth: on paper, the rule would look like:
16:21 samth: <c,e,k> -> <c',e', f::k>
16:22 samth: in redex, you have to rewrite it as:
16:22 samth: <c,e,(frame ...)> -> <c',e',(f frame ...)>
16:23 clklein: I would do this: (--> (c e k) (c' e' (f k)))
16:23 clklein: that's what I meant by replacing pairs with length-two lists
16:24 samth: ah
16:24 samth: so that's ok for continuations, but sucks for environments
16:24 clklein: The downside is that you end up with lots more parens in the term
16:24 samth: b/c you can use the neat lookup-in-the-middle pattern matching tricks
16:25 clklein: oh right, then you can't do ((x_0 v_0) ... (x_i v_i) (x_i+1 v_i+1) ...)
16:25 clklein: right
16:25 clklein: hrm
16:25 samth: it would be nice if you just supported (a . b) for this
16:26 clklein: How would it typeset?
16:26 clklein: a :: b maybe?
16:26 samth: yeah, that's how it is in this paper
16:27 clklein: I guess we could make it configurable too.
16:28 clklein: Should (p . q) be a pattern too?
16:30 clklein: I don't immediately see the harm.
16:31 clklein: I like this idea. I'll give it a try.
16:37 samth: yeah, (p . q) should be a pattern
16:40 (join) martin_hex
16:40 (quit) martinhex: Disconnected by services
16:40 (nick) martin_hex -> martinhex
16:45 clklein: Wait a minute. Can't you already do this now?
16:48 samth: no, it rejects dotted pairs
16:48 clklein: in patterns, but it seems to work in term templates
16:49 samth: yes, it does
16:50 clklein: Probably doesn't typeset right, though
16:52 samth: yeah
16:53 clklein: I'll try to fix that, at the very least
17:01 samth: i personally care more about the patterns
17:02 clklein: really?
17:03 clklein: They don't help in your example, do they?
17:10 clklein: oh, nevermind. I see.
17:31 eli: samth: You pinged me a few days ago?
17:31 samth: yes, can you move the commit notification to this channel?
17:32 (join) chturne
17:33 eli: samth: I don't care much about it, but do other people want that?
17:33 eli: (Also for bug notifications, right?)
17:33 samth: we could take a survey, but that doesn't seem that productive
17:33 samth: i was just thinking the commits notes
17:34 samth: the bug notification is (at least currently) limited to developers
17:34 samth: so it makes sense to split it that way
17:36 Tasser: what's the difference between a hash table and a dictionary?
17:36 eli: samth: Splitting will require some work, since the code just reports everything in a single channel.
17:37 samth: eli, the commits are done by github
17:37 eli: Ah, right.
17:37 samth: so it should be really easy
17:37 eli: I shouldn't even have looked in the bot's code.
17:37 (quit) Yann3: Read error: Connection reset by peer
17:39 (join) Yann1
17:41 (join) PLT_Notify
17:41 PLT_Notify: racket: master Robby Findler * 66f6f04 (1 files in 1 dirs): improve the drawing of the menu popup when it is yellow and the mouse is over it
17:41 PLT_Notify: racket: master Robby Findler * da9c34c (1 files in 1 dirs): use the new refresh-now method
17:41 PLT_Notify: racket: master Robby Findler * 49952e5 (1 files in 1 dirs): adjust scribble's in-drracket html generator button to use the sandbox so user code cannot corrupt drracket. ...
17:41 PLT_Notify: racket: master commits cdd099f...49952e5 - http://bit.ly/guwKcz
17:41 (part) PLT_Notify
17:41 eli: samth: There.
17:42 samth: eli, thanks!
17:42 samth: also, i have a much bigger favor to ask you, eli
17:43 samth: which is to respond to my email about PLT files
17:43 eli: What's the subject?
17:43 eli: (My mailbox is a mess.)
17:44 samth: "Setting up scribble documentation in a .plt archive"
17:44 samth: you're the only person i know with experience w/ plt files
17:45 (quit) masm: Quit: Leaving.
17:46 eli: samth: The issue about the lack of indexing is probably due to running setup only on your collection, and not on scribblings/main, which is what does the indexing.
17:46 eli: So you should probably add a ++setup flag for that too.
17:47 samth: but it doesn't build the documentation at all
17:47 eli: As for the fact that no documentation is built when the plt file is installed, I have no idea why that doesn't happen.
17:48 samth: the lack of indexing is only an extra problem when i do something that I can't ask the students to do
17:48 eli: The setup code was messy, and it's generally getting worse.
17:48 samth: tewk, ping
17:49 eli: I'm pretty sure that the ++setup thing would resolve your second problem; the first one is obviously more important anyway.
17:49 eli: samth: BTW, did you try to install the plt file from both raco and drr?
17:50 eli: (...and got the same problem in both?)
17:50 samth: at some point i tried it w/ raco, but maybe not will the full set of options to build the plt file
17:50 samth: so i will try that again when i get a chance
17:51 eli: It'll help to try both, since they might be invoking things differently.
17:56 samth: clklein, it would also be nice to have built-in support for sets
17:58 clklein: samth: Yes, finite maps too.
18:05 (join) somnium
18:06 (join) MayDaniel
18:11 offby1: Tasser: did you get an answer?
18:11 Tasser: offby1, more or less
18:12 offby1: good.  (I didn't see one in the scrollback)
18:12 offby1: anyway, my answer would be: hash tables are one particular (very popular) implementation of dictionaries.
18:12 Tasser: offby1, read it
18:13 samth: clklein, there's something pretty great about writing down a theorem about galois connections and having redex check it
18:13 Tasser: offby1, yep, read that in the docs
18:37 (quit) mwolfe: Remote host closed the connection
18:43 (quit) mheld: Quit: mheld
18:53 samth: clklein, ping
19:03 clklein: samth: pong
19:03 (join) epochwolf|2
19:03 (quit) epochwolf|2: Changing host
19:03 (join) epochwolf|2
19:04 samth: clklein, i have a transition system that's defined as a fixed point of a transition function
19:04 samth: which should be what redex is good at
19:05 samth: the problem is that the components are tuples of sets rather than sets of tuples
19:05 samth: so i'm unsure how to model it
19:07 clklein: I don't understand.
19:08 samth: ok, i'll try again
19:08 samth: imagine we have a cek machine
19:08 clklein: k
19:08 samth: we can think of a collecting semantics for this as being the lfp of a function that takes a set of configurations, and steps each element
19:09 samth: redex is great at modeling this sort of thing -- finding that lfp is built in
19:10 samth: the function, we'll call it F, has the signature F : P(<C,E,K>) -> P(<C,E,K>)
19:10 samth: where P is the powerset function
19:10 samth: so far so good?
19:11 samth: clklein, ^
19:11 clklein: Still thinking.
19:12 samth: the function F is just (lambda (l) (apply append (map apply-reduction-relation l)))
19:13 clklein: I'm trying to figure out what the lfp is
19:13 samth: least fixed point
19:13 clklein: of this particular function i mean
19:14 clklein: (what it represents)
19:14 samth: if the inital configuration is <program, empty, []>, then the lfp is the set of all reachable configurations
19:14 samth: basically every element of the `traces' graph
19:15 clklein: so this is all wrt a particular seed configuration?
19:16 samth: the function F is independent of the initial configuration, it's just the transfer function
19:17 clklein: I don't see why F even has a lfp
19:18 clklein: Wait, wouldn't the empty set be the lfp?
19:18 samth: we work around that by adding back the initial state every time
19:19 clklein: so F is parameterized by a seed state
19:19 samth: yeah
19:19 samth: i guess that's the best way to think about it
19:20 clklein: K, got it now. Sorry it took so long :)
19:20 samth: ok, so that's the setup
19:20 clklein: so sets of tuples is easy. But you want tuples of sets?
19:20 samth: right
19:21 samth: so imagine we have F' : <P(C),P(E),P(K)> -> <P(C),P(E),P(K)>
19:22 samth: where a \in F(B) if some choice from each component of B gives you a
19:23 clklein: OK, I see how this would come with abstract interpretation.
19:23 samth: if you want a more fully worked out example, you can look at figure 5 here: http://www.irisa.fr/celtique/jensen/ICFP09-ANFCFA.pdf
19:24 samth: so, it's easy in redex to map  <P(C),P(E),P(K)> ->  P(<P(C),P(E),P(K)>)
19:24 samth: where each component in the result is a singleton set
19:24 samth: but then you want to flatten it all back together
19:26 clklein: Do you map <P(C),P(E),P(K)> -> P(<P(C),P(E),P(K)>) with a rule that picks any element of each component and reduces the choice?
19:27 clklein: Err, no, that would be <P(C),P(E),P(K)> -> P(<C, E, K>) I think
19:27 samth: yes, and you just wrap a list around the result to get the bigger signature
19:31 clklein: a list around each component in the result?
19:31 samth: yeah
19:31 samth: and what i want to to flatten everything
19:32 samth: i suppose i could define it with one relation, and then another one that calls `apply-reduction-relation'
19:36 clklein: One relation takes a <P(C),P(E),P(K)> to a set of tuples, where each component is a singleton set, and another one that defines a tuple of sets by pairwise unions?
19:38 clklein: err, maybe that second one isn't right.
19:39 samth: the second one isn't right
19:40 samth: wait, no, it is right
19:40 clklein: heh
19:40 clklein: I'm starting to think it depends on how you understand what I meant by the second one. I'm not longer sure :)
19:47 clklein: Is the second one actually a function?
19:49 clklein: Oh, even if it is, you still want to define it as reduction relation so you can do something like traces to compute the lfp-like thing. I think.
19:52 (quit) stis: Remote host closed the connection
19:55 clklein: Anyway, not really sure if I helped, but that was interesting. I'll try to read the paper soon.
20:04 samth: hmm
20:05 samth: yes, i want to define it as a reduction relation for exactly that reason
20:12 (join) offby1`
20:14 (quit) offby1: Ping timeout: 240 seconds
20:37 (quit) MayDaniel: Read error: Connection reset by peer
20:56 (quit) mceier: Quit: leaving
21:07 (quit) parcs: Ping timeout: 240 seconds
21:08 (join) parcs
21:21 (quit) martinhex: Ping timeout: 250 seconds
21:27 (join) mheld
21:50 (join) martinhex
21:51 (join) mwolfe
22:33 (quit) epochwolf|2: Quit: Linkinus - http://linkinus.com
23:28 (quit) samth: Ping timeout: 255 seconds