00:00:00 --- log: started forth/04.01.18 00:00:59 * kc5tja is back (gone 05:24:52) 00:01:25 kc5tja, hi 00:01:30 dude 00:01:30 kc5tja, greetings from windows 2k 00:01:30 Oi. 00:01:32 tis kc5tja 00:01:41 kc5tja, do me a favor? 00:01:43 * warpzero grabs kc5tja and steals him 00:01:48 What started out as simple house chores turned into playing host for a combined birthday and going away party. 00:02:40 :) 00:02:49 kc5tja, please help me decide between the following: 00:03:04 (a) IRC, Browse the Web, and download random stuff until I fall asleep. 00:03:32 (b) Go outside, attempt to hook up the cable connection to my room, lay in my bed, and wank to cheap cable pr0n 00:03:36 arke: what the fuck is your problem get the fucking porn cable now 00:03:47 (c) Set up SuSE to install overnight, and sleep 00:03:54 :) 00:04:16 arke - why don't you just access internet pr0n? 00:04:26 I will abstain from this decision. It sounds more or less like a personal problem, and I'm not inclined to assist with such matters. 00:04:34 because my outgoing x0ns are being logged 00:04:42 arke - those *bastards* 00:04:48 arke - get freenet. 00:05:03 kc5tja, well, ignore the pr0n part, I'd just end up watching a comedy.. 00:05:09 ayrnieu, er, parents, not ISP :) 00:05:18 arke - yes -- get freenet. 00:05:39 freenet as my parent provider? 00:06:02 MY dear friend, I am quite liberal, but I would prefer Nature(tm) as my Parental Provider. 00:06:23 arke - er, no. see freenet.sf.net 00:08:57 . 00:09:12 * ayrnieu remembers discovering that his father monitored his IRC usage. 00:09:46 heh 00:09:52 my dad just looks where I go. 00:10:08 if its something like bash.org or maddox he won't care. 00:10:09 His monitor truncated channel names, so when I joined some nonsense #suicide_now_is_the_time_blah_blah_blah he saw '#suicide_now' and discussed this with me. 00:10:28 but once I went to goatse on accident, and he asked me the next day if I had some sort of gay fetish 00:10:38 aiee. 00:10:48 :D 00:10:59 and you said "Yes, I may as well come out of the closet.", right? 00:11:24 No, because then this would be a bash quote, and not a normal IRC conversation 00:12:31 please don't confuse me with the sort of person that finds bash.org interesting, or says things like 'bash that. You must bash that.' 00:15:15 Not to say anything bad about those people. 00:15:24 Actually, please just pretend that I didn't say anything at all. 00:16:21 arke - so what do you care if he knows that you like to read the bestiality category of www.mcstories.com ? 00:16:43 whats mcstories.com? 00:17:06 goooodddd dddaaaammmmnnnn iiiiittttssss ccccoooollllddd ooooouuuuttttssiiiiddddeee 00:17:22 not only that, but I need another cable to hook up my room. 00:17:28 So I gotta go find one. 00:17:35 and i probably wont :( 00:17:57 arke - Erotic Mind Control Fiction archive, or something like that. 00:18:38 eek :) 00:18:59 and you should definitely avoid http://trollse.cx/ ! 00:19:09 .... 00:19:12 i wanna see that. 00:19:20 do me a favor, send it to me by email? :) 00:19:23 dammit 00:19:38 er, go ahead. It just parodies goatse.cx 00:19:39 I got 2 cable outs in my room, how do I know which one is the right one? or would they both work? 00:19:49 ayrnieu, whats it look like? 00:19:53 It doesn't actually have anything that your father would find bothersome. 00:20:34 likewise, http://oralse.cx/ has something hat you wouldn't expect =) safe for work and parents and such, and cute -- except for the URL. 00:21:16 haha 00:21:16 cute 00:23:40 ugh. just my luck. just as i was about to head out to find a cable, the heater turns off, rendering even the slighest noise quite loud 00:26:51 Am I to assume you should be in bed right about now? 00:26:56 (as should I, technically) 00:31:20 OK, I'm going to bed. 00:31:28 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 00:45:13 O 00:45:14 MY 00:45:15 GOD 00:45:22 I could be a friggin' spy after this 00:45:45 I woke nobody except the dog, and thats because she got a blast of light in her eyes :) 00:45:56 i hope i did it right... 00:47:28 *sniff* 00:47:31 it doesn't work. 00:47:38 do I have the right to cry? 00:48:53 Permission denied. 00:50:07 :( 00:50:14 I'm gonna have to reverse it tomorrow. 00:50:20 make it seem like I never did it. 00:50:24 :( 00:50:38 Or maybe I should try switching the cable thingies 00:52:50 nope :( 00:52:53 * arke cries 00:57:32 what are you doing? 00:57:38 crying. 00:57:53 before that 00:58:26 trying to get my cable to work in my room so that i can watch TV from my computer 00:58:37 oh boo hoo 00:58:48 thats why i download south park and simpsons and stuff :> 01:05:51 grr! 01:15:10 --- quit: arke ("Leaving") 01:19:06 --- part: Nutssh left #forth 02:14:50 --- join: schihei (~schihei@pD9548FC0.dip.t-dialin.net) joined #forth 02:26:12 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 02:28:00 Dobry den! 02:29:14 Privet :) 02:29:31 What's "den"? 02:30:06 "Day" 02:30:13 Ah.. :) 02:30:17 God dag. 02:31:08 IIRC, you're interested in FPU, aren't you? 02:31:35 I've wrote a set of words to control FPU state. 02:31:35 I've been doing a little work with it, but not very much. 02:32:19 Now, I can change rounding mode etc from console. 02:32:30 :) 02:33:42 Also I can indicate control flags and tags (free/zero/NaN/num) visually. 03:54:23 privet 03:58:18 kak dela? 04:13:31 Dobyjj den'! 04:15:14 :) 04:25:36 --- quit: fridge ("Client exiting") 04:36:01 What would you think about this asm implementation? 04:36:20 1. Tree organized vocabulary. 04:36:37 2. Root vocab. is ASM 04:37:24 3. Each word to define addressing type and such changes vocabulary to inner one. 04:38:24 That means: a) words with more operands lye deeper; b) words with more operands that lye on stack are not visible. 04:41:14 When you write BX )@ AX MOV you make 3 changes of CONTEXT vocabulary: BX and AX are registers and should show themselves as such, )@ should change addressing type from register to indirect. 04:41:39 Ah! 04:41:58 Also command should reset to root ASM voc. 04:42:35 Advantages: simple correctness control 04:43:02 Disadv.: it's relatively hard to define macros. 05:17:30 --- join: fridge (~hovil@CommSecureAustPtyLtd.sb1.optus.net.au) joined #forth 05:29:06 erll 05:29:07 well 05:29:25 as forth is word based langauge it woudl be interesting to see dictionary AI implementation out of it 05:29:47 see, in dictionary the words refer endlessly to other word 05:29:53 one word is described with another word 05:32:34 someone gave me that description once 05:32:40 it confused me greatly 05:34:38 as I couldn't figure out how you'd get the processor to do an add, say, if all forth did was continually reference other words, when did it actually tell the computer to do stuff! 05:35:13 he didn't know enough about forth to tell me that not all forth words are created equal 05:35:26 and some actually flip bits about 05:39:06 well 05:39:13 you need to know semiotics 05:39:32 to understand the thing 05:39:55 and the words are not equal in dictionaty, by no means 05:40:27 that woudl require perhaps own dictionary for the processor 05:40:35 and then references between dictionaries 05:40:36 7win 5 07:59:53 --- join: Nutssh (~Foo@gh-1029.gh.rice.edu) joined #forth 08:39:12 Dobryjj vecher! 08:39:18 Hi :) 08:43:23 --- quit: qFox ("if at first you dont succeed, quit again") 08:46:45 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 08:48:21 --- join: madgarden (~madgarden@Kitchener-HSE-ppp3576567.sympatico.ca) joined #forth 08:49:09 Dobryjj vecher! 09:27:39 --- join: warp0x00 (~warp0x00@216-161-218-25.hlna.qwest.net) joined #forth 09:27:57 What architecture was it that used the zero-page? 09:36:03 68xx? 09:36:26 Apple ][? 09:39:31 ya 09:39:38 68 hundred? 09:39:47 was there such a thing? 09:49:11 Or 6502/03. 09:55:55 6502 and its myriad derivatives. 09:56:39 Okay. 10:53:37 --- quit: schihei (Client Quit) 11:04:35 --- quit: qFox ("if at first you dont succeed, quit again") 11:54:51 --- join: thin (thin@bespin.org) joined #forth 11:56:17 Hi 11:56:27 hey 11:57:13 Dobryjj vecher! 11:57:52 :) 11:59:02 --- quit: thin (Client Quit) 12:13:22 --- quit: warp0x00 ("Leaving") 12:15:09 get you get forth onto a PIC? 12:15:22 can 12:16:14 --- part: Nutssh left #forth 12:55:42 --- quit: fridge ("Leaving") 12:58:09 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 12:58:45 --- quit: tathi (Client Quit) 14:06:18 ? 14:07:37 --- join: fridge (~fridge@dsl-203-33-164-15.NSW.netspace.net.au) joined #forth 14:10:34 --- join: aktnot (ident@233.80-202-65.nextgentel.com) joined #forth 14:10:41 --- quit: aktnot (Client Quit) 14:27:51 --- join: Sonarman (~matt@adsl-64-169-92-25.dsl.snfc21.pacbell.net) joined #forth 14:28:27 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 14:51:47 --- join: TheBlueWizard (TheBlueWiz@ip-207-198-223-116.nyc.ny.FCC.NET) joined #forth 14:51:47 --- mode: ChanServ set +o TheBlueWizard 14:51:50 hiya all 14:52:33 Terve! 14:53:43 did you know that J.R.R. Tolkien loved finnish? 14:54:26 hiya Robert 14:55:01 I heard he also loved Ayn Rand. 14:55:06 Sonarman: um, I'd say he got inspired by Finnish....I am not sure if he *loved* Finnish 14:55:08 Thus Finnish must be Evil! 14:55:43 Ayn Rand also liked capitalism! 14:55:58 she also breathed oxygen! 14:56:38 I haven't heard of that one...but equating X with evil cuz someone who loves X also loves Y and Y is evil is fallacious... 14:58:07 Really? I have to redo my whole life philosophy then.- 14:58:22 * TheBlueWizard reboots Robert :) 14:58:39 Hehe. :) 14:59:23 Tolkien wrote that finding a book on Finnish grammar was like discovering a wine cellar full of wonderful wines 14:59:41 s/finding/discovering 14:59:42 yep 14:59:49 so he did love it :) 15:00:01 at least, i guess so :) 15:00:06 How do you know he didn't hate wine? 15:00:18 I don't know if that *means* he loved Finnish :) 15:00:27 ok, whatever 15:00:47 he could have hated wine 15:00:57 :) 15:00:59 * TheBlueWizard watches TV news now 15:01:04 hey tbw 15:01:54 --- quit: fridge (Remote closed the connection) 15:02:36 terve mur :) 15:04:47 --- quit: ayrnieu ("system") 15:11:03 --- join: ayrnieu (~julian@206.61.132.55) joined #forth 15:17:25 --- quit: qFox ("if at first you dont succeed, quit again") 15:29:18 * TheBlueWizard finishes watching...for now :) 15:34:05 --- join: rO| (rO_@pD9E591FB.dip.t-dialin.net) joined #forth 15:34:42 * rO| greets 15:36:05 hiya rO| 16:28:40 gotta go...bye all 16:28:49 --- part: TheBlueWizard left #forth 16:34:19 --- join: Mark4 (~Mark4@12-160.lctv-a5.cablelynx.com) joined #forth 16:41:23 --- quit: Mark4 ("Leaving") 16:41:23 --- quit: rO| (Read error: 104 (Connection reset by peer)) 16:46:54 --- join: I440r (~mark4@12-160.lctv-a5.cablelynx.com) joined #forth 16:51:44 --- join: Nutssh (~Foo@gh-1177.gh.rice.edu) joined #forth 17:25:33 --- join: kc5tja (~kc5tja@66-91-231-74.san.rr.com) joined #forth 17:25:34 --- mode: ChanServ set +o kc5tja 17:33:24 --- quit: I440r (Read error: 110 (Connection timed out)) 18:04:04 --- quit: chandler (Remote closed the connection) 18:04:06 --- join: chandler (~chandler@d-84-77.dhcp-149-159.indiana.edu) joined #forth 19:51:11 --- join: arke (~arke@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth 19:51:16 WTF!? 19:51:29 XChat works, but IE and ping don't? 19:51:52 .... 19:51:55 and now IE works. 19:51:57 what the hell. 19:52:10 Windows is like that. 19:52:50 Yes, indeed ^_^ 19:52:56 how are you, my friend/ 19:53:09 I am doing well. 19:53:17 :) 19:53:20 I'll be afk cooking spaghetti in a few minutes though. 19:53:29 * kc5tja finally completed the J Primer too. 19:53:33 fun fun fun 19:53:38 We had sushi today 19:53:40 Man, J is a nice language. I really, really like it. 19:53:41 mucho good 19:53:50 whats J (Primer)? 19:54:03 J is, for all intents and purposes, APL3. 19:54:09 kc5 - I prefer K to it, but K has licensing problems far worse than J's. 19:54:24 ok... 19:54:27 It's a plain-ASCII implementation of the APL programming language, with a some sticky issues fixed. 19:55:01 ayrnieu: Yes, I agree. K is a better language. J appears to not have a built-in "where" function, for example. But it's easily synthesized in J, fortunately. 19:56:08 hrm, do you have a link to it?> 19:56:19 http://www.jsoftware.com 19:56:24 and K at www.kx.com 19:56:29 Dobrot utro! 19:56:30 ayrnieu: Also, I find it's easier to learn K once you already know J. 19:56:39 And J has, IMO, better documentation and tutorials. 19:56:57 APL suffers terribly from a scarcity of free implementations. 19:57:09 ayrnieu: Yes, I agree completely. 19:57:21 I only know of A+, but it doesn't seem very healthy. 19:57:26 While learning J, I was actually considering writing my own array-processing language that was GPL or something. 19:57:53 Well, A+ is APL2, basically, complete with custom font pseudo-requirement. 19:58:27 an implementation of K would certainly please me. 19:58:31 Most people won't learn A+ because (a) you need a custom font, or (b) the symbol names are lengthy and non-mnemonic, and people find it too hard to type. 19:58:56 I looked, but I couldn't find any publicly available specifications on the K language. 19:58:58 You don't really need the custom font -- just to read any of the tutorials. 19:59:29 K defines itself by its implementation, pretty much. It also suffers from the scarcity of documentation that you noted. 19:59:35 ayrnieu: The things I did read all assumed I had the custom font, and reading it without it made it next to impossible. 20:00:34 Things like "iota rho foo" looks goofy though, too. It especially becomes a problem if you happened to have a variable in your program called iota or rho. :) 20:08:05 --- quit: ayrnieu ("system") 20:08:09 Hmm...looking at K's 'amend' operation. I have to admit that I prefer J's amend over K's. K's seems to be less flexible overall. 20:09:27 --- join: ayrnieu (~user@206.61.132.55) joined #forth 20:12:15 eeke 20:12:20 what does 'amend' do ? :) 20:17:01 arke: amend changes a specific range of an array. 20:17:19 So instead of saying something like (assuming C-style syntax): 20:17:25 foo[3] = 45 20:17:28 you can say this: 20:18:14 foo = (3 } foo) 45; 20:18:20 On the surface, this looks rather useless. 20:18:33 But it becomes immensely powerful when you start working with multiple values of an array at the same time. 20:19:14 foo = ((foo = 3.14159) } foo) 6.28319; // Replace all instances of 3.14159 with 6.28319. 20:19:29 err, well, you get the idea. 20:20:24 It's hard to explain without going into a lot of other detail. 20:20:34 It's best to just read the tutorial from the beginning, several times if need be. 20:21:13 (should read foo == 3.14159 above) 20:24:42 anyway, I'm going to be right back. 20:24:53 Might take a small nap, and then, it's time for me to cook up some food. 20:27:56 --- quit: Nutssh ("Client exiting") 20:39:07 * kc5tja is away: cuc(c)ina. 20:39:47 --- quit: arke (Read error: 104 (Connection reset by peer)) 20:39:57 --- join: arke (~arke@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth 20:40:22 kc5tja chooses to remain neutral in the Italian-Spanish war :) 20:41:23 --- join: Nutssh (~Foo@gh-1177.gh.rice.edu) joined #forth 20:42:24 Actually, I normally wouldn't be, but I forgot the proper spelling. :) 20:42:32 Especially since I'm an Italian. 20:44:14 hehe, cool 20:44:21 I got bash.org/?random on my desktop 20:55:33 --- quit: ayrnieu (Remote closed the connection) 20:57:42 web on your desktop? ugh :? 20:58:51 :P 21:00:38 --- join: ayrnieu (~julian@206.61.132.55) joined #forth 21:14:33 * kc5tja is back (gone 00:35:25) 21:15:59 :) 21:16:09 I just had a sudden urge 21:16:13 And I think I will 21:16:21 write the smallest ever DOS Forth 21:17:04 STC? :) 21:17:36 btw, a project some people here might be interested in is http://www.s2.org/~chery/projects/lisp500/lisp500.c 21:17:41 it's lisp in 500 lines of C 21:17:46 including garbage collection 21:17:54 wow! 21:18:13 eeek 21:18:16 was that IOCCC? :) 21:18:19 no 21:18:22 he's still hacking on it 21:18:36 I think at this point he wants to get it to the point where it can host a tiny CLOS 21:18:49 eek 21:18:51 gly 21:18:52 :) 21:18:58 Sonarman, Yes, STC. 21:18:59 I didn't say it was readable 21:19:03 I said it was in 500 lines of C 21:19:11 Sonarman, hardware stack being return stack 21:19:13 Chan:) 21:19:14 er 21:19:17 chandler, :) 21:19:26 works pretty well too 21:19:35 very basic 21:20:00 Fascinating. K is supposedly implemented in only 1200 lines of C code. 21:20:41 in how many lines of asm do you think I could create a super-minimal forth? 21:20:50 as in, no high-level words, and only a handful of low level ones 21:21:17 arke: With, or without, the interpreter/compiler/user interface? 21:22:04 with a minimal user-interface 21:22:12 well, no, actually. 21:22:15 no user interface. 21:22:42 just a BOOT variable, and the rest turnkeyed :) 21:23:17 3 machine language instructions. 21:23:21 MOV ESP, 21:23:27 MOV Exx, 21:23:36 JMP [

] 21:23:54 eek 21:23:54 but there's still the interpreter. 21:23:57 :) 21:24:07 but its still the same, none-the-less. 21:24:24 I asked explicitly about the interpreter/compiler/user interface, and you later said no. :) 21:24:25 interpreted words are jumped to none-the-less, but are just compiled different 21:24:33 i said user interface = no 21:24:34 :) 21:25:27 arke: 1 line of asm: 21:25:35 db .... 21:25:59 BTW, look at postForth. 21:26:20 ^_^ ^__^ ^___^ ^__^ ^_^ 21:26:20 postforth, ok. 21:26:20 lemme google 21:26:46 There should be link from: dmoz.org/Computers/Programming/Languages/Forth/Implementations/ 21:27:19 http://jcomeau.freeshell.org/www/domains/risp.org/members/jcomeau/postforth.html 21:27:23 Check spelling, it's from directory: dmoz.org 21:28:10 I'm really glad that languages like K, J, APL, and Forth aren't "dead." 21:28:37 Or, another thing I could do is get myself some C compiler and do a small optimizing forth compiler 21:28:45 Sure, nobody would hire a Forth programmer, but each language community is alive and well, albeit a bit specialized perhaps. 21:28:46 Or I could do it in assembly anyway 21:28:56 :) 21:29:00 arke: I'd just use assembly. 21:29:06 C is useful if you're looking to make a portable Forth environment. 21:29:15 i guess that settles it then :)\ 21:29:17 But assembly is what you want if you're looking for both small size and overall runtime efficiency. 21:29:38 Yep yep yep, I can code assembly better than C in some cases. 21:29:51 I can code better in assembly than in C in all cases. 21:29:58 :) 21:30:02 even x86? 21:30:04 Yes. 21:30:12 x86 is easy but b0rk3d 21:30:13 :) 21:30:39 Most of the optimization rules for x86 also apply to RISC engines. 21:31:15 Yeah. 21:31:20 I have a question. 21:32:03 I have an answer. Whether it will apply to your question, however, remains to be seen. 21:32:19 --- join: arke_ (~arke@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth 21:32:33 Would a 3-step pipeline in which every instruction will always spend only one clock ineach step be faster than say the x86 20-step pipeline? I would think so.... 21:32:50 For run-time throughput of straight-through code, no. 21:32:58 Each instruction will still appear to take only one instruction. 21:33:08 It's only with branching that pipeline depth comes into play. 21:33:44 what amazes me is that nobody uses a register dependent branch prediction. 21:33:53 ?? 21:33:59 --- quit: arke (Client Quit) 21:34:28 basically, you would try to optimize branches the same way as any other --- to avoid register dependencies 21:34:44 that way, it can be predicted and the right code automatically fetched. 21:35:11 Well, it's not possible on x86, because the flags register is the sole, global place all condition codes are located. 21:35:22 even if it takes to insert a nop, its still better than having to wait "until the last minute" 21:35:29 kc5tja, true, but its possible on MIPS 21:35:33 However, both MIPS and DEC Alpha do register-based branching, so I'm sure they are able to do the optimization you speak of. 21:35:34 and probably PPC as well 21:35:41 yep 21:35:48 PPC is a hybrid between the two types of architectures. 21:35:53 hehe, i came up with that on my own :) 21:36:15 While it uses a global flags register, the register consists of subfields, each are independent from each other for the purposes of contention and branch prediction. 21:36:40 ? 21:36:53 Imagine if the x86 had 8 EFLAGS registers. 21:37:09 We'll call them CR0 through CR7. 21:37:14 eek!? 21:37:23 cr0 -cr7 already exist :) 21:37:26 The last comparison or flags setting instruction always sets CR0. 21:37:38 arke_: You have no imagination, you know that? 21:37:49 Fine. 21:37:53 er, why that? 21:37:54 We'll call them FR0-FR7. 21:37:59 :) 21:38:08 just pointing it out :) sorry 21:38:21 Because then you'd know that CR0-CR7 is clearly and patently a name I chose for this discussion, and applies only to this discussion, and has zero to do with x86's real CRx registers. 21:38:25 :) 21:38:37 ok .. :( 21:38:56 There's a reason why I called them CR0-CR7 anyway. 21:39:05 (largely because that's what they're called in PowerPC assembly language) 21:39:20 Anyway, imagine if the x86 had 8 flags registers: fr0-fr7. 21:39:23 Ok. 21:39:29 Ok. 21:39:43 Everytime you execute a CMP instruction, it always affected fr0. 21:40:00 The old fr0 would be placed in fr1, and fr1 would be copied to fr2, etc. 21:40:21 So fr1 has the last comparison results, and fr2 the results before that one. 21:40:22 Etc. 21:40:42 so like a queue 21:40:47 * kc5tja nods 21:41:02 Now imagine that branch instructions took not only an address, but also an FRx register. 21:42:02 You could launch a CMP instruction, do some other stuff, then JGE FR0,SomeLabel 21:42:21 Or CMP, CMP again, then immediately Jxx FR1,SomeLabel. 21:42:40 Hrm, 21:42:47 You could still do that optimization then :) 21:42:52 PowerPC works more or less along those lines. 21:43:11 It's slightly more restricted, but overall, yes. 21:44:49 interesting. 21:44:56 I don't like the idea of flags too much though 21:46:19 I agree. 21:46:22 I think they're senseless. 21:46:42 It also wastes an opcode: CMP and SUB are precisely the same "instruction" -- the only difference is CMP doesn't store the result in the destination register. 21:46:43 --- join: arke (~arke@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth 21:47:02 --- quit: arke_ (Client Quit) 21:47:42 Do you propose to join comparison and branching? 21:48:09 ASau: Only against zero. I see no reason to use anything else. 21:48:38 Hm. 21:48:43 cmp a, 100 21:48:56 bge label 21:49:05 vs. ? 21:49:21 Additional cell required. 21:53:05 --- quit: arke (Read error: 104 (Connection reset by peer)) 21:53:36 ASau: Versus: 21:53:44 ASau: sub a,100 21:54:01 ASau: jpos label ; branches if a-100 >= 0 21:54:13 ASau: jneg label ; branches if a-100 < 0 21:54:27 You see you need one additional cell. 21:54:34 For what? 21:55:24 Whoops, I meant to write: 21:55:26 jpos a,label 21:55:29 jneg a,label 21:55:44 You chang ``a'' hence you should restore it in some cases. 21:55:50 change 21:55:52 Yes. 21:56:02 Not really an issue on machines with plenty of registers. 21:56:09 Or on stack machines. 21:56:24 Interestingly, it becomes an inconvenience only on machines that are register limited. 21:56:45 x86. 21:57:24 I use SUB on x86 for another important reason: the arguments to CMP are swapped from its counterpart (as far as my experiments have shown). 21:57:36 In practice, it's not really an issue for me. 21:57:47 I've not programmed machines with plenty of registers. 21:58:07 I have. 21:58:12 It's truely a dream. 21:58:22 * kc5tja loves the 68000 and MIPS architectures. MMMMMmmmmmm..... 21:58:38 * chandler loves PPC 21:58:53 I suspect the PowerPC to be sufficiently close to the MIPS to be equally applicable. 21:59:01 it is :-) 21:59:22 with all sorts of wonderful instructions to discover 21:59:35 ? 21:59:40