00:00:00 --- log: started forth/11.04.17
00:29:41 --- quit: saper (Remote host closed the connection)
01:01:05 --- join: qFox (~C00K13S@5356B263.cm-6-7c.dynamic.ziggo.nl) joined #forth
01:03:43 --- part: DocPlatypus left #forth
01:31:17 --- join: saper (saper@wikipedia/saper) joined #forth
01:53:17 --- join: nHarzilei (harzi@camelot.germanistik.fu-berlin.de) joined #forth
01:53:22 <nHarzilei> hi
01:53:56 <nHarzilei> uhm, i only looked at forth a long time ago, but now i was reading this here: http://en.literateprograms.org/Fisher-Yates_shuffle_%28Forth%29
01:54:04 <nHarzilei> what's with the curly brackets?
01:54:34 <nHarzilei> where are they defined? do they actually assign variables, popping the inputs from the stack?
02:01:28 <ASau> Nowhere, it is non-standard thing.
02:03:51 <nHarzilei> ASau: so any idea what kind of implementation the author used?
02:04:26 <ASau> Yes, but the code doesn't comply style.
02:04:49 <ASau> It is in Fortran style only written in Forth.
02:05:20 <ASau> And literate programming doesn't help its clarity.
02:07:05 <ASau> What do you want?
02:07:26 <ASau> Just FYS or code in good style?
02:07:43 <nHarzilei> "FYS"?
02:09:47 <nHarzilei> actually i want to implement it in another remotely forth-like language
02:10:19 <nHarzilei> and searched for "fisher yates" "forth" and it came up with this and i was puzzled by the curly brackets
02:11:51 <ASau> Then why don't you just take FYS code in language you understand?
02:13:54 <nHarzilei> i wanted to look if there's a good implementation. i agree picking from the array again and again looks a bit less optimal
02:14:31 <nHarzilei> and of course now that i looked i'm curious in which forth variantes the curly brackets thing occurs 
02:14:52 <ASau> They occur in many variants.
02:15:00 <ASau> Only in different meanings.
02:15:43 <ASau> E.g. John Hayes' tester uses them.
02:17:26 <ASau> I'm not sure how you want to avoid array fetch.
02:21:45 <ASau> And you want to improve algorithm so that it doesn't use "choose random between" operation.
02:23:32 <ASau> With random ( u -- u' ) where  0 <= u' < u, you can write it this way:
02:24:01 <ASau> : shuffle ( ptr len ) times dup r random th  over r th  @swap repeat drop ;
02:33:57 <nHarzilei> heh, i guess your conciseness goes with the seasoned forth programmer. i just found another implementation and it also defines a randRange helper 
02:34:23 <ASau> What languages do you understand besides Forth?
02:35:15 <nHarzilei> perl, c, python, php, javascript, 6502 assembly ;)
02:35:23 <nHarzilei> java ;)
02:35:24 <ASau> Alright, let it be C then.
02:35:46 <nHarzilei> nah, it's ok, i don't need an explanation of fisher yates
02:36:17 <ASau> Consider you have "unsigned random(unsigned n)", not the standard one.
02:36:23 <ASau> Well...
02:36:26 <ASau> Let's rename it.
02:36:36 <ASau> "unsigned rnd(unsigned)", alright?
02:36:53 <ASau> Let it return random number in below its argument,
02:37:00 <ASau> and zero for zero argument.
02:37:19 <ASau> Then FYS is this:
02:38:22 <ASau> unsigned i = len; while (i--) swap(&a[i], &a[rnd(i)]);
02:39:14 <ASau> Do you understand the code?
02:39:19 <nHarzilei> i think you need to make sure you only swap with those that haven't already been swapped
02:39:40 <ASau> Why?
02:40:15 <ASau> Do you understand how the code above relates to description you reference?
02:44:30 <ASau> Alright, minor correction :)
02:44:45 <ASau> unsigned i = len; while (i) {swap(&a[i], &a[rnd(i)]); i--;}
02:45:29 <ASau> or you can check for (i>=2).
02:45:58 <ASau> Shuffling single-element array is rather stupid.
02:46:52 <ASau> Whatever, there's space for nitpicking.
02:47:54 <nHarzilei> yeah, i'm just looking at the other implementation and look why its author needed the range
02:48:37 --- join: MayDaniel (~MayDaniel@unaffiliated/maydaniel) joined #forth
02:49:03 <ASau> I think that this is how original algorithm was written,
02:49:26 <ASau> and because of going from first element to last seems more natural.
02:49:52 <ASau> But this isn't natural for some languages, thus I go from last to first.
02:57:54 --- quit: MayDaniel (Read error: Connection reset by peer)
03:08:52 <nHarzilei> hmm...
03:09:53 --- quit: nighty__ (Ping timeout: 240 seconds)
03:09:54 <ASau> Alright, do you understand how this algorith works on sets?
03:10:09 <nHarzilei> : randomSubset ( ptr len num ) times dup r random the over r th @swap repeat ; should work, right?
03:10:47 <ASau> Why do you need "num" parameter?
03:11:16 <nHarzilei> only get last num elements shufflen for less work when i only need num random elements out of the array
03:11:20 <nHarzilei> shuffled*
03:11:52 <ASau> Well...
03:12:07 <ASau> First, the definition you write is inconsistent with stack comment.
03:12:39 <ASau> Second, if you use "random ( n -- n' )", it is broken.
03:13:20 <ASau> ...when given three parameters, as stated in stack effect.
03:16:07 <ASau> Do you understand that Forth doesn't check all those things for you?
03:17:52 <nHarzilei> yes
03:18:24 <nHarzilei> i'll see if i can find the sample (i think that's how the word should be called) solution myself
04:02:22 <nHarzilei> is times .. repeat actually standard forth?
04:08:25 <ASau> No, but it is implementable.
04:11:51 <ASau> Do you need definition?
04:19:06 <nHarzilei> that'd be nice
04:23:21 <ASau> : `  postpone postpone ; immediate
04:23:21 <ASau> : times  ` >r ` begin ` r> ` ?dup ` while ` 1- ` >r ; immediate
05:03:24 --- join: MartianW (~Martian@unaffiliated/martianw) joined #forth
06:02:18 --- join: MayDaniel (~MayDaniel@unaffiliated/maydaniel) joined #forth
06:11:42 --- quit: MayDaniel (Read error: Connection reset by peer)
06:33:48 <nHarzilei> : th cells + ;
06:33:51 <nHarzilei> is that the definition?
06:39:54 <ASau> Yes.
06:42:31 <nHarzilei> so this is actually 4p you are using there?
06:43:44 <ASau> No.
06:43:51 <ASau> I've forked from pforth.
06:44:17 <Free-man> i fork()ed from spoon.zip
06:48:51 --- join: kernalogs (~jackokrin@nat67.mia.three.co.uk) joined #forth
06:59:29 --- quit: xpololz (Remote host closed the connection)
07:12:03 --- part: kernalogs left #forth
07:39:07 --- join: MayDaniel (~MayDaniel@unaffiliated/maydaniel) joined #forth
07:45:40 --- quit: MayDaniel (Read error: Connection reset by peer)
07:57:05 --- join: foocraft (~dsc@89.211.214.110) joined #forth
07:59:43 --- join: nixness (~dsc@78.100.177.99) joined #forth
08:02:19 --- quit: foocraft (Ping timeout: 250 seconds)
09:09:31 --- quit: MartianW (Quit: Bye all.)
09:41:34 --- quit: nixness (Ping timeout: 260 seconds)
10:16:02 <nHarzilei> is it hard to make something like {{ 1 2 3 4 5 }} that returns an addr with those and the number of items?
10:30:55 <ASau> What do you want exactly?
10:31:04 <ASau> Create an array in heap?
10:36:47 <nHarzilei> i need to create it on the heap if i want an "anonymous" array, right? as regular , will only work with a name? some way to define a name would help as well, with addr and number of items ending up on the stack when i run it (i think that's done with does>)
10:37:21 <nHarzilei> i'm not picky, i just don't want to count my items
10:37:22 <nHarzilei> ;)
10:38:05 <ASau> Well... Then make Forth count them.
10:38:10 <ASau> variable cnt
10:38:16 <ASau> : {{  0 cnt ! ;
10:38:29 <ASau> Hrm.
10:38:36 <ASau> : {{ depth cnt ! ;
10:38:49 <ASau> : }} depth cnt @ - ;
10:39:48 <ASau> I hope you'll figure how to move those items from stack into dynamic memory yourself.
12:48:04 <saper> or use PostScript :)
13:12:01 --- quit: martinhex (Read error: Connection reset by peer)
13:12:01 --- join: martinhex (~mjc@93-97-29-243.zone5.bethere.co.uk) joined #forth
13:35:10 --- quit: qFox (Quit: Time for cookies!)
13:49:50 <nHarzilei> ASau: hmm... times only really works good because it counts to 0
13:51:04 --- join: grai (~grai@38.70.70.115.static.exetel.com.au) joined #forth
13:55:05 <nHarzilei> ASau: is there a reason it's not
13:55:36 <nHarzilei> nvm
14:01:42 <ASau> Yes, that's the point.
14:01:59 <ASau> It counts to zero because it is easier to and more effective.
14:04:14 <nHarzilei> ASau: yeah, but for just getting K random elements out of the array i need to count from len down to k
14:04:22 <ASau> In particular, you don't need complex logic within your primitives that implement "do ... loop"
14:04:35 <ASau> Well... You asked for FYS first.
14:04:48 <nHarzilei> yes, i'm not complaining
14:05:03 <ASau> With incomplete FYS you do need more general looping constructs.
14:06:15 <ASau> If you don't use FIG-style strict controls, you can do it like
14:07:21 <ASau> n times last-n r < while ... repeap rdrop then
14:08:19 <nHarzilei> i was actually considering doing it with gforth u+do, but of course i'm curious if i can do it with begin .. while .. repeat instead of loop
14:08:21 <ASau> (I think that this "search" idiom/pattern should be known.)
14:11:03 <nHarzilei> "strict controls"? because times already sets up while?
14:11:45 <ASau> With FIG-style controls you can't use another "while", thus
14:11:45 <ASau> you can't write common search idiom like
14:12:27 <ASau> "begin more? while not-found? while next repeat found else not-found then"
14:16:35 <ASau> Nothing stops you from writing FYS in its original form.
14:16:57 --- quit: ygrek (Ping timeout: 246 seconds)
14:19:49 <ASau> I don't quite understand what your problem is.
14:20:23 <ASau> Obviously, what you want goes beyond FYS.
14:20:32 <ASau> But you don't tell it.
14:26:15 <nHarzilei> over in ircnet we have a channel where the bot can be programmed with oneliners in a forth-like dynamically typed toy language. this morning someone entered a word 7 rand 0 do userrandom ", SPAMM SAPAPAPAMMM SPAMMM" . say loop (. is concatenation there and say is type to irc channel)... despite the program being annoying anyway i thought that the possibility of people getting spammed twice is awfully inelegant as well.
14:27:37 <ASau> What can I say?
14:28:24 <ASau> Don't expose Turing-complete languages to possibly unfriendly users.
14:29:02 <ASau> I'm going off-line now.
14:29:11 <ASau> Gute Nacht!
14:29:14 <nHarzilei> thanks for your input :)
14:29:16 <nHarzilei> gn
14:31:59 <crc> regarding {{ ... }}, in retro this works:
14:32:00 <crc> : {{ ( "-an ) here 0 [ getToken dup "}}" compare [ drop 0 ] [ toNumber , 1+ -1 ] if ] while ;
15:22:31 <nHarzilei> so this is what i'll try now : ishuffle ( ptr len num ) over swap - to last-n times last-n r < while dup r@ random th over r@ th @swap repeat rdrop then
18:48:57 --- join: Fox78 (~fox@123.118.155.138) joined #forth
19:01:38 --- join: DocPlatypus (~skquinn@dsl253-084-031.hou1.dsl.speakeasy.net) joined #forth
19:57:41 --- join: foocraft (~dsc@178.152.70.10) joined #forth
20:11:13 --- join: cataska (~cataska@210.64.6.233) joined #forth
22:06:26 --- quit: foocraft (Quit: Leaving)
23:36:24 --- join: ygrek (debian-tor@gateway/tor-sasl/ygrek) joined #forth
23:59:59 --- log: ended forth/11.04.17