00:00:00 --- log: started forth/04.07.01 00:29:18 --- join: lalalim_ (~lalalim@p508ABA41.dip.t-dialin.net) joined #forth 00:47:14 --- quit: lalalim (Read error: 110 (Connection timed out)) 01:11:10 --- nick: lalalim_ -> lalalim 01:26:02 --- quit: Herkamire ("night") 01:36:56 --- join: SolarFire (SolarFire@pD9545912.dip.t-dialin.net) joined #forth 01:37:35 * SolarFire greets 01:47:42 --- quit: warpzero ("Tried to warn you about Chino and Daddy Gee, but I can't seem to get to you through the U.S. Mail.") 01:50:53 --- quit: proteusguy ("Leaving") 01:54:32 --- quit: aum () 03:08:11 --- quit: SDO (Read error: 110 (Connection timed out)) 04:07:24 * Tomasu is back (gone 05:43:52) 04:07:33 Hi Tomasu 04:07:41 yo 04:53:37 --- quit: I440r ("Leaving") 04:56:14 --- join: I440r (~mark4@216-110-82-1.gen.twtelecom.net) joined #forth 05:26:11 --- join: jdrake (irc_user@CPE00045afdd0e8-CM014410113717.cpe.net.cable.rogers.com) joined #forth 07:15:13 ok, is it possible to have a direct threaded forth CPU ? 07:15:44 direct? 07:15:48 yes 07:15:52 not subroutine threaded 07:16:11 can you give a small description that I might understand 07:16:34 erm its a little difficult to explain on short notice hehe 07:17:04 its to do with compiler design - the threading mechanisms are direct, indirect, subroutine and token 07:17:19 there may be more threading models but those are the most common 07:17:38 in direct threaded each token is a pointer to an address to be executed 07:17:52 in indirect each token is a pointer to a pointer to an address to be executed 07:18:15 in token threading each token is an index into an array of addresses. 07:18:23 i am only familiar with how one does it in a higher level language, example for cocoa where you tell NSThread to send a message to an object in another thread 07:18:25 in subroutine threading you dont have tokens. you have calls 07:18:28 --- quit: madgarden (Read error: 104 (Connection reset by peer)) 07:18:43 this has nothign to do with threads of that sort 07:19:03 forth threads are not the same as say "linux threads" 07:19:42 in fact i would say that threads as described by multi tasking operating systems are NOT threads 07:19:48 they are merely execution streams 07:19:53 assembler is NOT threaded 07:20:54 hmm, i am probably not the best person to know then :-) I am not familiar at all with what you describe 07:21:49 :) 07:22:05 threading in forth is not a complex subject 07:22:27 i am using ficl and am new to forth in general 07:22:32 but its not easy to describe in detail until you know more about forth :) 07:27:42 i got it :) 07:32:45 ei I4_wrk 07:32:47 Hey* 07:32:50 And jdrake 07:33:00 hey rob 07:33:27 sorry, that is bad form: "ROB HEY" 07:33:34 Indirect threaded Forth CPU... I don't see how it would be different from a subroutine threaded. 07:33:37 jdrake: ;) 07:34:21 table of calls, vs a table of pointers to code executed by an inner interpeter (?) 07:34:32 s/table/series/ 07:35:29 Aren't there IDT forths running on MachineForth CPUs? 07:36:21 I don't know. I don't know as much as I would like about the various MF CPUs 08:44:45 --- join: SDO (~SDO@68.170.20.201) joined #forth 08:47:17 the pain, the agony. 08:47:25 started to paint our new house 08:47:32 its been a while since i've been this broken 08:48:21 haha 08:48:34 physical exertion is good for you 08:49:56 ugh tell that to my legs 08:50:04 and arms 08:56:13 --- join: crc (crc@0-1pool176-42.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 09:02:42 --- part: crc left #forth 09:21:21 --- join: Serg (~z@193.201.231.126) joined #forth 09:21:28 hi ! 09:24:42 --- quit: jdrake (orwell.freenode.net irc.freenode.net) 09:24:42 --- quit: jc (orwell.freenode.net irc.freenode.net) 09:24:42 --- quit: cmeme (orwell.freenode.net irc.freenode.net) 09:28:09 --- join: jdrake (irc_user@CPE00045afdd0e8-CM014410113717.cpe.net.cable.rogers.com) joined #forth 09:28:09 --- join: jc (~jcw@65.3.39.49) joined #forth 09:28:09 --- join: cmeme (~cmeme@216.184.11.30.unused.swcp.com) joined #forth 09:50:19 do any forth chips do instruction pipelining ? 09:53:46 Not sure. And I don't see kc5tja/solar_angel around 09:53:53 right 09:53:56 dammit! 09:53:57 :P 09:54:13 i miss solar_angel too ;) 09:54:30 i rally like this kind of girls ;)) 09:54:33 you guys do of corse realise she has a girlfriend right ??? :) 09:54:38 *snicker* 09:54:55 who is also linux savy 09:55:13 yeas 10:10:22 --- join: wossname (wossname@HSE-MTL-ppp61432.qc.sympatico.ca) joined #forth 10:13:07 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 10:24:43 --- join: randolm (wossname@HSE-MTL-ppp66739.qc.sympatico.ca) joined #forth 10:29:51 * Tomasu is away: bbl 10:31:05 hi hi hi 10:31:15 --- quit: wossname (Read error: 110 (Connection timed out)) 10:31:30 solar_angel has a g/f? :) 10:32:39 I4_wrk: hmmm........ 10:32:42 Yes 10:32:52 Robert: :) 10:32:56 Robert: anything new? 10:33:01 I always feel so sucessful when I meet girls who've had more girlfriends than I have 10:33:06 Not really. 10:33:47 hey arke, i figured out the problem I had with understanding ficl 10:33:58 ficlWord *ficlDictionarySetPrimitive(ficlDictionary *dictionary, char *name, ficlCode code, ficlUnsigned8 flags) 10:34:05 ficlCode should be ficlPrimitive 10:34:35 :) 10:35:07 brb 10:42:03 back 10:42:10 * arke is in Linux now, yay 11:00:22 --- quit: Serg () 11:03:42 arke, you know what would be cool to see? Objective C wrapper for gtk :-) 11:23:33 jdrake: they have that :) 11:24:18 one that isn't gtk1 old? 11:24:55 think so 11:25:41 --- join: SolarFire-| (SolarFire@pD9EE19F3.dip.t-dialin.net) joined #forth 11:29:21 hi So 11:29:33 oops 11:29:36 hi SolarFire-| 11:31:11 --- quit: SDO (orwell.freenode.net irc.freenode.net) 11:31:11 --- quit: lalalim (orwell.freenode.net irc.freenode.net) 11:31:15 http://www.informatik.uni-osnabrueck.de/elmar/projects/gtoolkit/ 11:31:25 jdrake: thats one 11:32:07 although its GTK 1.2 11:33:44 --- join: SDO (~SDO@68.170.20.201) joined #forth 11:33:44 --- join: lalalim (~lalalim@p508ABA41.dip.t-dialin.net) joined #forth 11:33:48 exactly 11:33:48 gtk 1.2 sucks compared to 2.x 11:33:51 one that uses glib and gtk for base support would be nice 11:33:59 then I might actually bother to learn gtk 11:33:59 ::) 11:34:02 make your own 11:34:04 its OSS 11:34:05 ;) 11:34:10 too busy with this other project 11:34:15 ok 11:34:31 and I will have an api for you to check out in a moment 11:34:53 c00l 11:38:00 http://paste.lisp.org/display/1557 11:38:41 the basic thing to note about this api is that it should be taken as something to implement, those typedefs up top are meant specifically to be implemented by another datatype, but where a void pointer will work effectively 11:41:37 any comments? 11:42:35 --- quit: SolarFire (Read error: 110 (Connection timed out)) 11:47:21 lisp? 11:47:33 lisp what 11:47:43 very long evil names! 11:47:53 oh no 11:47:56 with intercaps 11:47:59 >:( 11:48:14 this thing is a C api that is supposed to go language independant 11:48:24 it is where forth comes in at a layer below this one 11:48:26 i'm gonna start me an anti-intercaps-campain some day 11:48:38 it just looks better this way 11:48:56 (yeah dont let me put you off :) 11:49:16 me go have shower now :-) 11:49:16 hot 11:56:58 underscores mixed with camelCase... yum 11:57:16 that should be a PITA to remember how to type in properly 12:03:25 _ 12:03:25 STACKEFFECT: --- x 12:03:25 DESCRIPTION: 12:03:25 Leave an undefined value x. Presumably it is to be dropped at some time, or it is a place holder. 12:03:27 lol 12:04:00 not random though 12:11:35 --- quit: randolm (Read error: 104 (Connection reset by peer)) 12:12:23 Herkamire, there is a reason to the naming 12:12:37 as an example: void Display_SetCaption(DisplayRef d, const char *string); 12:13:45 in the second implementation that I designed this was a Display:SetCaption(string) in lua, now that I am doing an independant api that is sort of object oriented (in C) it makes sense to know what primary set these functions belong to 12:15:19 i wish C were more typesafe though 12:18:07 sorry, make that 'i wish C was typesafe' as it is by no means typesafe :-) 12:21:52 Define 'typesafe'. No, it's not, not like Ada or other languages. But if you use typedefs and prototypes and compile with -W -Wall, it becomes much more so. At least, at compile time. Run time, you're on your own. 12:24:24 In Ada, Ocaml, et al. you can have distinct types that are unique. In Ada specifically you can create a type that is an Integer, but won't work where an Integer will. 12:24:38 In C, a typedef merely aliases types 12:24:54 which causes some nice confusion when you need to know what type something really is 12:25:26 --- join: warpzero (~warpzero@mi016.dn178.umontana.edu) joined #forth 12:29:32 typesafety = bad 12:29:34 :) 12:29:52 arke, do you believe that or just being anti-smerdy 12:30:03 both 12:30:19 i think in most cases typesafety is preferably 12:30:35 but if you notice I do have an interest in forth and lua, so not draconian like smerdy 12:30:43 :) 12:31:01 forth has one way of being type safe though 12:31:09 you can distinguish between integers and floats 12:31:38 one shouldn't use floats in most cases 12:31:49 they're quite easy to circumvent most of the time 12:32:04 well, when I do my opengl binding, i am using floats 12:32:25 but being opengl has integer functions I will bind those too 12:48:10 I don't like typesafety either 13:04:57 --- quit: warpzero ("Tried to warn you about Chino and Daddy Gee, but I can't seem to get to you through the U.S. Mail.") 13:08:13 --- quit: Herkamire ("out") 13:18:05 * Tomasu is back (gone 02:48:13) 13:29:26 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 13:49:38 --- join: trap1010 (~trapxvi@CPE0030654f85d8-CM014400102747.cpe.net.cable.rogers.com) joined #forth 13:51:51 --- quit: tathi ("leaving") 13:55:55 --- join: MorgothDc (~morgoth@gw-bdi-3.ticom.pl) joined #forth 13:56:23 Hell[o] 13:56:33 trap1010: I see you are here too, great 13:56:39 trap1010: Have you take a look ? 13:57:56 I can but my NAT is apparently buggering the DCC 13:58:31 trap1010: You didn't noticed the link -> http://www.tbs-software.com/morgoth/files/menu 13:59:05 trap1010: I upload it to my home page. it's only a little modified version of script that I found on the net 13:59:53 trap1010: Anyway take a look. Generaly it works fine, I need only a way to abort it (For example when user press diffrent key) and back to OF prompt 14:00:04 trap1010: Forth is a bit exotic language for me 14:00:38 Just get used to the concept of tracing it on a VM in your head and its eerily elegant 14:01:27 trap1010: Yeah, but for that I need to read a tutorial or manual first ;) I do not plann to do anything in forth in the future 14:01:48 trap1010: Only finish my bootmenu generator, this is a reason of my question 14:02:10 trap1010: BTW Maybe you see something to improve ? 14:03:18 what is wrong with apple's boot menu 14:03:56 jdrake: I can't use it 14:03:58 Judging by the fact that this has MorphOS in the header 14:04:01 he's not on an apple 14:04:22 trap1010: you have right 14:04:33 can I have right too? 14:04:43 are you on a pegasus? 14:04:52 jdrake: Pegasos. Yes 14:05:12 how much did that cost you all told? 14:05:26 Are you using an uncounted loop? 14:05:33 it always looked bit expensive to me 14:05:37 Because I would use begin again for that 14:06:33 jdrake: In compare to Apple it's not so expensive 14:06:46 jdrake: it cost ~450 euros now 14:06:54 That's not bad 14:07:04 is that complete or just the board/chip 14:07:31 trap1010: begin ? Hmmm, can you paste a example ? Or better modify this script for me (I know, I know, I am lazy bastad) 14:07:38 jdrake: Only board + G4 14:07:51 quite a bit much then 14:08:03 maybe those will get lower price after time 14:09:03 jdrake: Who knows. Generaly it's realy nice piece of hardware 14:09:21 I'm not sure exactly how OF works, but I woudl think that if you simply end your program without issuing a boot command 14:09:27 it would fall back to the OF prompts 14:09:45 i have been kind of stuck in sub-ghz computers for quite a few years here, I am waiting to upgrade to something fast in a year or two 14:10:41 this one is good: http://www.reallifecomics.com/ 14:12:50 jdrake: the 1Ghz Pegasos II G4 is 1015 canadian from ultraspec-canada.com 14:13:02 jdrake: Have you seen new Peggy HP ? (www.pegasosppc.com) 14:13:06 It runs linux + MOL but not Mac OS X 14:13:43 trap1010: Yes, MacOSX won't work. 14:13:58 trap1010: And you forgot about MorphOS :) (www.morphos.net) 14:14:17 Well, I suppose if you did enough work in Darwin, you could run OS X on top of that 14:14:29 What software is availible for MorphOS? 14:14:48 It just seems like it a niche OS 14:15:10 I mean, if there's a tool for everything, good, that's like Mac OS X 14:15:27 trap1010: No, this is "work in progress" stuff 14:15:41 trap1010: But quite impresing now (It boots in ~4s) 14:15:46 trap1010: For example 14:15:51 Not bad 14:17:21 trap1010: Hmmm, propably yes. But you need someone to do it first 14:18:14 I'm just wondering why I'd spend 1015 on a 1Ghz G4 + Mobo when I can Athlon 64 2800 and a mobo for $500 14:18:43 For the same price I could probably get at least an Athlon 64 3400 14:18:49 because its a ppc 14:18:55 and it can run mac os 14:19:03 But it can't run Mac OS 14:19:07 oh 14:19:16 screw it then 14:19:20 :P 14:19:22 It can run Mac OS inside a virtualized environment inside Linux 14:19:33 er what G4 system can't run mac os? 14:19:34 trap1010: Because the CPU eats only 7.5-9W of power 14:19:39 that seems oxymoronic 14:19:49 trap1010: because it's extremaly silent (Passve cooler) 14:20:11 MorgothDc the x86 chips take a lot more energy to run, right? 14:20:22 Oh hell yess 14:20:45 x86 is a scorcher, the Prescotts have a higher thermal flux per unit area than an electric iron 14:20:53 PPCs are cold 14:21:14 mine comes up from sleep within about 5 to 6 seconds :-) 14:21:26 what does coming up from sleep have to do with anything? your x86? 14:21:28 jdrake: :) 14:21:31 It would make a great mobo for a car PC 14:21:43 my powerbook wakes up in like <=1second 14:21:48 it's fucking nice 14:21:59 Yeah 14:22:03 ianp: I realy like so silent computer (I have zero collers in my tower case now) 14:22:06 ive never really owned another laptop 14:22:08 mine boot 10.3.4 in < 20 seconds 14:22:16 ianp: Apple laptops ROX !!! 14:22:20 yeah 14:22:22 I love my powerbook 14:22:25 1.5ghz! 14:22:30 Same 14:22:34 ianp: IMHO state of the art 14:22:35 makes OSX run sorta fast 14:22:35 1.5 Ghz 15" Latest Rev 14:22:40 yeah same trap 14:22:45 shes a beaut 14:22:45 ianp: But you can't run MorphOS :) 14:22:56 MorgothDc, not yet anyways... 14:22:57 not even in virtualpc? 14:23:10 ianp: As far I know no :( 14:23:15 oh its a ppc os 14:23:18 * ianp knows noting about it 14:23:21 whyt can't they run oit 14:23:22 ? 14:23:32 it hasn't been ported yet 14:23:35 it probably will though 14:23:37 ianp: Yes, it's PPC OS 14:23:39 to what 14:23:47 mac whatelse 14:23:53 mac is a ppc 14:23:58 you mean to run under mac os? 14:23:58 I would diffenitiely consider using this board for the car PC I've been planning 14:23:58 jdrake: I think no. MorphOS cordinator do not like Apple :( 14:24:08 I was going to use mini-ITX thing 14:24:09 it isn't the same ian 14:24:14 but this would be a lot better 14:24:27 MorgothDc, I think yes - because wider audience = more profits 14:24:35 jdrake: For now it works on Pegasos (1&&2) and on Amiga with powerup cards 14:24:45 jdrake: propably you have right 14:24:55 trap1010: Board is realy tiny 14:25:09 MorgothDc, 'are' not 'have' 14:25:15 Yeah, 23.6 x 17.2 cm 14:25:21 different verbs 14:25:49 jdrake: Yes, yes, I know I suck 14:25:59 jdrake: let it alone, we're not on #english-grammar 14:26:01 i think kc5's project will be sweet 14:27:45 The project I have in mind is to build a PC for my car, audio, GPS, media, and Internet 14:28:10 I didn't realize how small this pegasos board was 14:28:34 trap1010: :) 14:36:32 --- join: proteusguy (~proteusgu@69.79.24.31) joined #forth 14:36:49 OK, ThX for help for all 14:36:53 Time to test it 14:36:58 --- quit: MorgothDc ("the dance is held to the end ... when the shades of darkness transcend ...") 14:40:13 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 14:40:21 OH FUCK THIS 14:40:28 $#@%#@((#&@%()&*@($*# 14:40:31 thank you. 14:40:56 Who here has implemented FORTH on a PowerPC? 14:41:11 The organization of the stack is giving me trouble 14:41:19 oooh me! I am 14:41:44 its a load/store orientated machine, its not meant to have a lot of data on a stack 14:41:47 the bootable version of ciforth is a much older version then the ciforth running under an os (dos) i was working with. the older version doesnt even have goddamn definitions for the basic words to work with its multiple vocabulary support 14:42:24 if you look closely, you can just see the anger flowing of me 14:42:45 If this was x86 it would be straigh forward how to organize this: just use the stack normally, but if you it on PowerPC you're going to cripple yourself 14:42:56 Herkamire: how did you get around it? 14:43:02 trap1010: get around what? 14:43:09 I don't know x86 14:43:15 The load/store orientated nature of PPC 14:43:18 I don't have any problem on ppc 14:43:27 what are you trying to do? 14:43:44 --- join: MorgothDc (~morgoth@gw-bdi-3.ticom.pl) joined #forth 14:43:54 you keep TOS in a register, and the rest in memory 14:44:03 TOS? 14:44:07 You mean the stack pointer? 14:44:09 Top Of Stack 14:44:15 Yeah 14:44:19 That seems like a terrible idea 14:44:25 it's not 14:44:30 because you're memory-bottleknecked 14:44:38 what? 14:44:48 the inferiority of that approach is the reason RISC exists :P 14:45:00 how else colud you do it? 14:45:07 Herk: the x86 has 8 GPRs, the PowerPC has 32 14:45:29 you can keep the top two stack items in registers if you like. but that's a bit more complex 14:45:42 I would try to keep the top of the stack in the GPRs, but that's what's driving me nuts trying to schedule 14:46:00 More like the top 20 words 14:46:01 there are a couple nifty aproaches to trying to optomize out stack use 14:46:27 I don't see why you're getting mad at PPC because you're having a hard time doing something you can't do at all on x86 14:46:50 I mean, realistically, your stack is in L1 all the time 14:47:07 the trouble is when you compile a word, it has to know where it's operands are 14:47:08 herk: I can do it the x86 way, but that's not exploiting the PPC arch 14:47:13 yeah 14:47:24 That's the other half that's driving me nuts 14:47:40 it is tough to exploit PPC in forth 14:47:51 Its not meant for stack-orientated programmign 14:48:00 x86 and PPC are designed to run optomized C code 14:48:29 Well, x86 can use a stack without much penalty because it has so few registers that most C code ends up doing a lot of memory access 14:48:47 Most of those GPR's are taken up with pointers anyways 14:48:59 the way I intend to exploit PPC is to have 5 or 6 processes sharing the 32 registers (each process only neads 5 or 6 registers) so task switching can be lightning fast 14:49:05 --- quit: MorgothDc ("the dance is held to the end ... when the shades of darkness transcend ...") 14:49:13 so realistically you have 4 GPR 14:49:43 Herk: yeah, but then you limit a compiled word to only running on one thread 14:50:02 yeah 14:50:12 I figured it'd be that way anyway 14:50:25 I've never done multi-threaded apps 14:51:26 Much as I'd like to use Forth for serious programming, its not going to work out on PPC for me 14:52:27 In fact, I'd go so far as to say that RISC and Stack-orientated paradigms at fundamentally at odds with eachother 14:52:32 I to have been frustrated that it is hard to get forth to use all the strength of PPC. but I've mostly gotten myself to stop worrying about that and pay attention to making a simple system that's easy to program. 14:53:04 they certainly aren't all that well suited. 14:53:11 You haven't exploited PPC until you've written a routine to calculate the exp of a vector of floats without any function call overhead 14:53:15 but I think PPC does forth about as well as x86 14:53:23 except for the branch to and branch from 14:53:36 Well, on x86, you have to do a lot of interacting with the top of the stack 14:53:43 so really you're not losing anything 14:54:03 forth is suited to MISC processors 14:54:04 time to write the author me thinks 14:54:30 When you carry that over to PowerPC, its a major bottleneck 14:54:47 I don't see the bottleneck you're talking about 14:55:02 on PPC you keep the top of stack in a register 14:55:02 it binds you to memory 14:55:14 example, in x86, you call a function with parameters in a stack 14:55:19 on x86 you have he stacks in memory too 14:55:27 in PPC, most function calls, the parameters are in registers r3-r10 14:55:45 right. but not in forth 14:55:48 Unless you have a huge argument list 14:55:48 yep. x86 only has one hardware stack, the other stack is in mem 14:55:49 right 14:55:58 (at least i believe one is hardware 14:56:25 So with Forth, you're tying yourself to a bottleneck (slow memory) that the load/store arch is meant to address 14:56:40 Of course, as I said before, in practice the stack is in L1 cache, so its not too terrible 14:56:48 but its not as good as having it all in registers 14:57:12 yeah. 14:57:24 I have all kinds of ideas about fancy optomizers 14:57:48 but optomization is not a current concern 14:58:13 optomization is a complex issue 14:58:23 sometimes it's bettor to have small code than fast code. 14:58:37 Well, when you write functions in PPC assembler, if you don't make any function calls, and you only use the parameter registers for calculation, you can really scream 14:58:42 eg if you can make your entire program fit in the cache :) 14:58:57 yep :) 14:59:15 No, not quite 14:59:36 I mean, I never make an actual function call perse, just a branch-to and a branch-back op 14:59:58 bl 14:59:59 blr 15:00:02 When you make this a function like say exp that's heavily used in the kind of programming I tend to do 15:00:17 --- quit: lalalim (Connection timed out) 15:00:29 Its practically free inlining wihtout any of the disadvantages 15:02:07 it's not _that_ hard to write an optomizer in forth that gets rid of stack use within one function. 15:02:24 but things get harry (or just not very optomized) when you have to call functions 15:02:35 exactly 15:03:13 hi arke 15:03:26 hello everyone 15:03:35 what would you name a word that in lisp you might call it close-display or c++ Display::Close 15:03:57 Display Close 15:03:58 two words 15:04:06 two words? 15:04:08 --- nick: SolarFire-| -> SolarFire 15:04:09 first, put display on the stack, and then call close :P 15:04:13 (jk) 15:04:25 well it makes sense actually, because Display is a separate object 15:04:37 i plan on supporting multiple displays eventually 15:04:41 but only one for now 15:04:41 s" Display " close might be more correct 15:04:51 heh 15:04:53 that is just nasty 15:04:59 Well, if display is your object, my idea was a type of polymorphism 15:05:05 display-close works 15:05:26 i wasn't planning on doing any specific forth oop 15:06:52 : exp over swap for over * next nip ; 15:07:35 ~~ afk 15:07:48 ok, i don't know those stack operations et 15:08:06 : over ( a b -- a b a ) 15:08:12 : swap ( a b -- b a ) 15:08:21 : nip ( a b -- b ) 15:08:31 Herk: by exp its usually meant: e^x 15:09:03 oh 15:09:09 I did a^b 15:09:27 except I messed up, and that code there is a^(b+1) 15:09:29 wouldn't fexp be better e^x? 15:09:44 well, yeah 15:09:59 but exp() is the standard math shorthand for ASCII 15:10:14 right. I did pow() 15:10:29 Plus that's an inefficient way of doing it 15:10:33 are there ppc instructions for either of those? 15:10:36 no 15:10:37 surely 15:10:43 i would hope anyways 15:10:50 Not that I know of 15:10:52 there is some exp / ln combination that can do it 15:11:01 : exp ( x e -- x^e ) over swap 0 do over * loop nip ; i think, quite the same as herk looking back 15:11:11 that can do pow yeah 15:11:29 not entirely proper, but it'll do 15:11:45 qFox: we should call that "pow" as exp means power of e 15:11:57 : exp e swap pow ; 15:11:57 --- quit: SolarFire ("bye") 15:12:02 hey he said exp means e^x 15:12:16 i dont care what ya call it :p 15:12:19 qFox: yes. e=2.71828 15:12:28 oh really e 15:12:31