00:00:00 --- log: started forth/05.07.10 00:27:30 --- quit: sproingie ("Konversation terminated!") 01:57:52 --- quit: saon (Read error: 110 (Connection timed out)) 02:06:25 --- quit: saon|smg1 (Read error: 60 (Operation timed out)) 02:31:43 --- quit: kunphuzil ("Download Gaim: http://gaim.sourceforge.net/") 02:49:27 --- join: qFox (~C00K13S@92pc222.sshunet.nl) joined #forth 05:27:12 hi forthers 06:01:43 --- join: Robert (~snofs@c-f778e055.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 06:31:45 howdy virl :) 06:32:14 I've started work on decifering the coloroforrth. 06:32:48 tathi has taken Mark_s cf2html.exe and made improvments. 06:33:53 I took tathi's work and ran it thru DJGPP and made a windows version called cf2html-jg.exe, 06:34:41 It can be used like Mark_s, only the block numbers can be correct if used properly, and the links are better, tho will still need hand hacking. 06:38:18 --- join: saon (1000@c-66-177-224-222.hsd1.fl.comcast.net) joined #forth 06:40:01 --- join: PoppaVic (~pete@0-1pool73-50.nas24.chicago4.il.us.da.qwest.net) joined #forth 06:40:43 Mornin' 06:43:18 --- join: saon|smgl (~saon@c-66-177-224-222.hsd1.fl.comcast.net) joined #forth 06:58:37 Hmm.. I'm a thinnin' that 'immediate' is not necessary. 07:03:12 hmm.. Yeah.. We could live without it. 07:10:29 --- join: JasonWoof (~jason@c-24-218-95-147.hsd1.ma.comcast.net) joined #forth 07:10:30 --- mode: ChanServ set +o JasonWoof 07:10:42 lo, herk 07:11:00 hi 07:14:01 JasonWoof PoppaVic was just dismissing `immediate' as unecessary. :) 07:14:19 I was? 07:14:31 mostly, I think I see an alternative method. 07:15:12 I've figured out an alternative method of acheaving basically the same feature 07:15:49 I'm not entirely pleased with it 07:17:07 eh, wait, immediate not required? i'd like to see you have words like postpone 07:17:45 you do realize that a different naming convention does not mean you have no 'immediate' word, right? ;) 07:17:47 I'm thinking thusly: a preprocessor voc(list?) preceeds lang voc(list) - stick everything immediate and whatever into the preprocessor, ALWAYS check it before the main lib. 07:18:10 There must be an index of forth coversations where this point has been argued before, if just to save time :) 07:18:26 :) 07:18:34 prolly many - forth is more debate than anything else 07:18:43 yes 07:18:56 just think that the language exists for about 45 years, as well as most of the core language 07:19:01 oh, ray was saying that PoppaVic wis dismissing `immediate' 07:19:01 and that there's a reason for it :p 07:19:07 thought he said I was. 07:19:12 thanks JasonWoof for saving me :) 07:19:18 right wrong :_ 07:19:19 nah - poor quoting 07:19:20 :) 07:19:28 absolutely /me ashamed. 07:19:31 up all night 07:19:44 PoppaVic: the seperate vocamularies thing works fine. that's how chuck does it in colorforth 07:19:45 watching movies and making --- um --- ya um. 07:20:26 my eyes are basically glazed over. but I got all this energy to document a colorforth. 07:20:26 Good deal, JasonWoof - too bad we lack a really good site with linkages to arguments by topics from current to past ;-) 07:20:40 eww 07:21:02 hehe 07:21:39 herk: so, basically they use the initial-voc as the lexer/parser/preprocessor 07:22:45 jasonwoof... == herk? 07:22:55 oh 07:22:56 after jumping into protected mode, the dictionaries and the editor is built. 07:22:59 he said it right there :p 07:22:59 qFox: yeah 07:23:12 i actually just remembered that your name is jason :) 07:24:06 the dictionaries are 2 pieces each. the macro and the forth dictionaries both have two pieces. 07:24:17 qFox: I figured some people would just know my name, but most would take a bit. I don't like nick changes... but I've decided to start using my real name for everything 07:24:31 two pcs?? 07:24:32 I only know the real names of about 5 forthers here 07:24:37 hehe 07:24:45 * Raystm2 is not a real forther :( 07:24:53 herk: you can also leave Herk for "official" 07:25:00 but I knew Jasons name early on for me. 07:25:05 official? 07:25:26 yeah, like "God-like chanop", etc 07:25:38 * Raystm2 see's herk as JasonWoof alter-ego. To do all the "heavy lifting" and the like. 07:25:46 heh 07:25:56 chanop followed me to jasonwoof 07:26:03 yeah, jas is Dr. Banner - herk is Hulk - Hulk Smash! ;-) 07:26:12 hehe 07:26:21 I think it's wise to the whole nickserv link thing 07:26:35 bathing. brb 07:30:40 your surname is woof? 07:31:27 Woofenden 07:31:43 of course the world shortens that to Woof. 07:32:10 Apparently, even with-in his family, as he has explained in the past. 07:32:37 hey, I liked his battle-jittney name ;-) 07:32:59 me too. 07:33:22 strong first impression 07:34:59 like Hercules fighting his way thru the moors of England or something. 07:35:25 Strong but slightly bogged down. 07:35:53 I was thinking of the movie, myself.. 07:36:04 movie? please? 07:37:51 "Mystery Men" 07:38:12 havn't seen yet. /me notates for next library visit. 07:38:32 it's sorta' a 'B' movie, but it's cute. 07:38:47 I like some B movies :) 07:39:10 "The Shoveler", "The Blue Rajah", "The Spleen", heh - whacky stuff 07:39:24 :) 07:39:53 it's only B because it was just deliberately campy.. We got a copy, but I can only see it maybe once a year - I mostly listen as it plays, now 07:40:17 hehe ya I understand that. 07:40:34 How did you come up with PoppaVic? 07:40:43 anyway, if you have never seen it, you'll chuckle 07:41:03 I'm old, and it's phonetic-alphabet for me 07:41:04 worth a go. :) 07:42:27 oh oh ok got it. 07:42:37 takes me a while, sometimes 07:42:38 I'm still trying to visualize how to get to C-calls in a nice, easy fashion.. *sigh* I do not want to read my old code and adopt those solutions. 07:43:03 Poppa Foxtrot Victor seemed a bit long 07:43:22 yes, I see this also from time to time. Nice to call from forth, so routine or application that does not really need to be reinvented. 07:43:49 so=some :( 07:43:53 well, my mess will be a lib, written in C - but I want to make it as portable as possible 07:44:41 ..and I hate setjmp/longjmp - but it seems likely I'll need to deploy with it. 07:45:49 For the last couple days I've been obsessing about how to code the mess with C and lists... Contemplating how turning the usual mess on its head might be advantageous 07:46:45 does Minotaur do something like what you looking for. http://www.equi4.com/minotaur/minotaur.html 07:47:08 let me check - I think I visited there long ago 07:47:51 tho I read that its for python, tcl, and perl, I think it also glues 'c' 07:48:08 looks like it's just for cross-scripting 07:48:21 so not your thing then. 07:48:51 I think not.. but the ideas are still decent 07:49:09 * Raystm2 nods* 07:49:13 I'm trying to also consider adjustable-lexing. 07:49:29 lexing is a term I do not know. 07:49:51 Last I spoke with madwork, we considered allowing every voc to instantiate their own lexer. 07:50:09 okay parsing yes? 07:50:24 oh.. lexing is the precursor to parsing.. in forth they are about the same, but elsewhere "lexing" is building up a valid word 07:50:25 or lexical analyzation? 07:50:49 imagine a fixed set of regex - that's the typical lexer 07:51:01 got it, 07:51:20 now, forth generally doesn't have the regex - it's dead simple-minded 07:51:22 a progy to translate colorforth to html fits this description? 07:51:36 but, it seems to me that vocs should control the lexing. 07:51:48 hmm 07:51:54 vocs as in vocabularies? 07:52:01 nright 07:52:28 so, instead of a mess of pad and whatever... 07:52:33 hot swapable vocs, picked with pattern matching? 07:52:58 use the topmost voc to set the lexer, and it gets first-dibs... Work yer way down the vocs 07:53:23 sorta' like, I guess... THe concept is seriously open/vapid at this point. 07:53:23 following... 07:53:33 :) 07:54:13 so, we could demand "foo bar snafu" in A, but since it doesn't fit: B gets to try "foo.bar snafu", etc 07:54:29 order matters? 07:54:52 the vocs in the 'order' and the input-stream would, sure 07:55:03 * Raystm2 nods* 07:55:40 It's a really twisted, weird idea... But it borders right there - on the edge - of making some sense. 07:56:09 It lets folks complicate their "language" almost as much as they want 07:56:40 okay, had to be something good about it :) hehe 07:56:51 ..particularly since we still allow folks to parse the input-stream as usual for things like "variable" and "constant" 07:57:36 would this allow us to say, change those variables and constants depending on some conditions then? 07:57:48 well, by default we'd use the default forthish... But, by allowing this we allow all sortsa' wild ideas to flow 07:58:04 nono - yer missing it. 07:58:15 yes :) i'm sure :) 07:58:27 * Raystm2 sometimes needs hand holding. 07:58:32 I used those two as examples of the word-class that reaches AHEAD to grab the next word and define it 08:00:31 some how this is going to allow you to run applications for other languages? 08:00:41 nope 08:00:44 okay :) 08:00:49 fishing here :) 08:00:56 This itself is for itself.. Interfacing to C is a given. 08:01:10 okay. 08:01:33 All I want to insure is we can get back and forth from C to this TIL and the rest can take care of itself. 08:02:00 C can link the lib; the TIL shell can call on C modules. 08:03:14 however, yes... I think what all the above does for us is let us dynamically define lexers so that, in theory, you could parse and compile the C (if you got seriously mad) 08:04:17 so, a forth based `c' compiler is with in reach of this idea? 08:04:38 My largest bitch remains that getting from the datastack to the real stack to make C-calls is gonna' suck. 08:05:07 I have no idea, really - my target is merely a basic lib, so I can write a shell, so I can proceed with Metabuilder 08:05:34 Metabuilder, yes you mentioned this. What does it build? 08:05:47 My prob is.. I hate writing things over and over - and if writing one lib now saves years later, I'm all for it. 08:05:55 --- join: sproingie (foobar@64-121-15-14.c3-0.sfrn-ubr8.sfrn.ca.cable.rcn.com) joined #forth 08:05:56 :) 08:06:01 sproingie :) 08:06:02 wb 08:06:34 My Metabuilder project is sorta' targeted at replacing autoshit/make/sh - with an extensible, forthish system that ports 08:06:42 PoppaVic; sounds smart to me. 08:07:24 heh - if it was "smart", I'd have thought 2 million monkeys at the university-level woulda' done it already. 08:08:07 however, it's been a gripe for mefor years.. I like C, generally, and I like forth, and I get damned tired of all the nonportability between/within each 08:08:14 they'er still working on Shakespear. 08:08:36 anybody written an FFI for C before? 08:08:51 I wonder if you could loop through the lib and make a forth word for each symbol 08:08:55 there are several, but they all look odd as hell 08:09:19 there is even one IN .fs with gforth, but I merely glanced at it 08:09:22 I think I'd go with the approach of making aforth word myself for each symbol 08:09:47 and specifying the number of parameters there, but I'm curious if you could have it done automatically for all symbols 08:09:56 very, very tedious - also means examining headers or lib-links and doesn't do much for args/returns 08:10:05 you'd just need some way of looping through them and finding out how many operands they take (and if they return anything) 08:10:09 Yeah, me too - I'd love that 08:10:40 I'd THOUGHT of using gcc -E and then working over the prototypes. Ctags and etags do that, I believe 08:11:24 The thing to remember is, there is no shame in several programs being used to do simple transmogrification to feed each other. 08:11:35 you need an online database of previously, married libraryies, and an interface that can grap what you need from the database on the fly ;) 08:12:04 and, yes: you'd want a database file, not store it all in RAM 08:12:39 I'm not imagining that you'd need a whole lot af symbols from a lib 08:13:02 thinking of SDL mostly, where I doubt you'd need more than 10 08:13:07 Personally, I'd also like to see 'vocs' for ansi, posixN.N, etc 08:13:29 oh, you mean a dlopen from the forthish? 08:13:40 or a special format, anyway? 08:15:16 There otter be a way to "gcc -S test.c", and rape that to learn how args are passed, too. BUT, yer back to having a special rapist for every arch. 08:16:12 WHat I've NEVER seen is a nice, standardized datafile for how that stuff is generated by each gcc - or cc for that matter. 08:16:23 it's a mess 08:16:30 it sure as hell is 08:16:35 I say heck with it 08:16:40 I'm not linking 08:16:45 and, this is why I get so disgusted with all sortsa' folks 08:19:32 this is what makes me think that it'd be easier to get the computer to behave the way I want it to by writing my own OS that working with existing stuff 08:19:52 although it'd probably be easier to work of an existing kernel like L4 or something 08:20:12 No, actually.. I still believe we need a vmasm and plugins for archs 08:20:24 vmasm? 08:20:33 c calling convention precludes knowing how many args a function takes if there's any optimization 08:20:34 virtual machine assembler? 08:20:37 virtual-machine assembler.. 08:20:42 like Java? 08:20:49 like C? 08:20:53 no idea, I despise java 08:21:36 PoppaVic: i pointed you at llvm earlier, right? 08:21:43 No, I mean this: universal opcodes and regs and semantics, use a voc and special module for transmogrifying to true mc 08:22:06 righto, sproingie really interesting, but the docs are sorta' beyond me and the C++ is godawful 08:22:15 PoppaVic: what about architectures that don't have that many registers, or who's registers aren't wide enough 08:22:21 sproingie: they have a good idea in there 08:22:22 PoppaVic: there's been some work done on that, i forget the name. something like "universal binaries", that basically do exactly what you're talking about 08:22:35 PoppaVic: there's a compiler for some subset of it 08:22:37 I'm sure they do, sproingie 08:23:01 JasonWoof: it's easy: if there are no available regs, it's a var in the vm - always.. everywhere 08:23:55 the arch-module it uses could optimize as it compiles, at that point we wouldn't care 08:24:16 and if your compiling for a 16-bit architecture? 08:24:18 i dunno, to me it sounds like you're describing any compiler 08:24:30 sounds a hell of a lot like C to me 08:24:34 the arch-specific "modules" being backends 08:24:39 sproingie: sure I am - it's also portable. Which no 'asm' is at this point 08:24:41 except with somewhat different syntax 08:24:46 JasonWoof: C is kinda nasty as portable asm. C-- is probably better 08:24:54 JasonWoof: or LLVM bytecode 08:25:04 JasonWoof: yeah, except C is screwed by assorted std as well 08:25:46 i suspect C could be made to do what you want with "naked" calls and bitfields 08:25:53 but it would be soooo ugly 08:26:09 yep 08:26:15 not even sure gcc supports naked calls on most arches 08:26:25 msvc does, never seen it used tho 08:27:23 "naked calls"? 08:27:35 no stack convention whatsoever 08:27:43 ahh 08:27:44 basically syntax sugar for a jump 08:28:12 you juggle the stack manually 08:28:17 yeah, I can see that 08:28:27 you'd do forthish gymnastics 08:28:29 never seen it used. most people either use inline or asm 08:28:58 yep, and inline-asm is what breaks C worse than their goddamned commies and ansi/posix crap 08:29:22 inline asm can be more portable than twiddling with libraries, actually 08:29:38 fer example, if you want to atomically increment a long 08:29:42 I doubt it: change the arch, the code is dead 08:30:01 the x86 asm will run on every x86. but no OS has a standard function for it 08:30:15 yep, yer stickin' to a single arch 08:30:21 you'd have to #ifdef it for all these "architectures" anyway 08:30:29 nope 08:30:54 as I said, you'd load the apropos module - dlopen 08:31:02 as i just said, you don't have the module 08:31:11 right, then you need one 08:31:16 and if you write the module, magic elves won't take care of implementing it 08:31:28 in fact, C won't help you at all. your module has to be in asm 08:31:42 oh, shit - yer back to C - yer right 08:31:51 and a call into a DLL isn't atomic, so it defeats the purpose 08:32:10 granted that's a special case 08:32:15 'atomic' is another mess It was at least as bad on DOS 08:32:26 but calls into dll's are still expensive 08:33:05 DOS didn't really need many atomic ops, not being multithreaded 08:33:23 Expensive? Sure, from a z80, or a 286. And, multithreading isn't that thrilling. 08:33:47 I don't believe I've written a single forth, threaded, program 08:34:01 i dont believe i've ever seen a forth that supports it 08:34:11 I've done it in C, mostly for tests. ANd Pth is easier. 08:34:19 sproingie: come now, FIG did ;-) 08:34:24 shared state concurrency is overrated anyway, erlang has the better idea 08:34:28 so did f83 and fpc, iirc 08:34:58 PAUSE is cooperative multitasking, not really what i'd call multithreading 08:35:17 let's presume the OS can fork fine.. Forget threads for this mess. 08:35:57 i thought the whole forth philosophy was that portability was overrated 08:36:06 no. 08:36:17 i can see how that philosophy would get started in the embedded world 08:36:25 shit 08:36:26 the repair people broke the car i was supposed to get 08:36:28 so, no car for me 08:36:37 It was that extensiblity and interactivity was really useful, as well as incremental compiling a lexicon 08:37:10 embedded-folks are just that: a minority 08:37:13 sounds identical to the lisp philosophy 08:37:25 prolly, lisp predates forth 08:37:38 I can't read lisp and don't want to, though 08:37:55 i read some paper by this russian guy, saying russian hackers were using forth for the same things american programmers were using lisp 08:38:05 symbolic programming and all that 08:38:08 I wouldn't be suprised 08:38:18 must have been painful without gc. then again, might have been faster 08:38:23 early gc was awful 08:40:09 gc is an interesting idea - and generally overrated WHen you need it, you need it bad - otherwise, mind yer memory 08:41:08 you'd need a special allot/allocate for gc. stuff from create would be damn near impossible to reclaim 08:41:25 yep. Seen that 08:41:49 i have a forth gc sitting around, written for gforth i think 08:42:06 there are several solutions, and because of the way C sees the universe, and forth sees the univer: none are truly interesting 08:42:34 add the asm/os universe and it gets worse yet 08:42:48 what might be better for rf anyway is to just use multiple interpreters. free an interpreter, free its memory 08:43:06 take a very parallel view of problems 08:43:14 define 'interpreter' 08:43:18 use IPC like erlang does to transfer state 08:43:27 forth interpreter 08:43:40 right now it only likes having one interpreter running in an address space 08:43:58 there's no hard reason why there can't be several 08:44:47 lots of languages have to deal with multiple thread states in a single interpreter 08:44:59 rf is so small that it's easier to just copy the entire interpreter 08:45:46 fork 08:46:04 bleh. wastes at least a meg per fork 08:46:15 I dunno erlang, other than to know it exists. Never cared 08:46:40 forks are cheap, fast and easy. 08:46:50 IPC isn't as nice 08:47:15 --- quit: JasonWoof ("brb") 08:48:48 ^^ 08:49:02 lo, ark 08:49:34 hehe - the Great Assembler Wars in ##C ;-) 08:49:39 hehe 08:49:57 fasm wins. 08:50:02 (so far) 08:50:04 bad llama 08:50:16 I can live w/o fasm quite well 08:50:23 or nasm or masm 08:50:33 as makes more sense 08:50:41 fasm is better than the rest, but the macro support is still inadequate. 08:50:49 as is still cpu/arch-specific 08:51:02 my assembler isnt! 08:51:04 except that its vaporware 08:51:13 Suprise, suprise... 08:51:23 as I've said before, we need a universal-asm 08:51:30 you won't get one 08:51:32 I made designs for that 08:51:48 sproingie: as I never write asm, I'm not too worried, either 08:51:50 the whole point of asm is to provide mnemonics for opcode numbers 08:51:59 yep 08:52:07 ..and thus I said a module 08:52:08 EVERYTHING else is higher level 08:52:17 hell, to the chip designer, asm IS high level 08:52:18 sadly, no 08:52:22 addd %0, %1, %2 08:52:30 sproingie: ^^ 08:52:34 ugh, at&t syntax bites 08:52:34 almost everything else dips down and we are supposed to like it 08:53:24 att as syntax adds some ops that yield multiple instructions without being macros 08:53:28 most assemblers do not 08:53:30 ##C? whats that the C# channel? 08:53:32 --- join: JasonWoof (~jason@Herkamire.student.supporter.pdpc) joined #forth 08:53:32 --- mode: ChanServ set +o JasonWoof 08:53:49 yeah, ## means "generic" 08:54:17 so, a portable C? 08:54:23 virl: it's the channel about the #C channel ;) 08:54:25 if you say so ;-) 08:54:37 the ## is a freenode issue 08:54:57 we gave up and made them happy 08:55:43 anyway, some would call forth a universal asm 08:55:55 it should be, but ain't 08:55:56 of course most i've seen make gross assumptions about cell size 08:56:02 yah 08:56:26 not sure that's forth's fault tho 08:56:40 llvm had the same idea of "universal", but C++? Not interested 08:56:53 you will search in vain forever 08:56:57 i guarantee this 08:57:13 you can't blame the language for the committees or lack of vision 08:57:27 nor can you blame them for your personal distaste of their language choices 08:57:38 true 08:57:40 they wrote it, a language is just a means to an end 08:57:46 great 08:57:54 and, the idea idles 08:58:32 PoppaVic: well, how about you and I work together to make a universal asm? 08:58:46 forth lets me play at the low end without the awful static mess of asm, but i dont see myself learning much new theory-wise from it 08:58:50 PoppaVic: i promise I'll actually code this time ^^ 08:59:06 arke: all you need is a vm lib, module form and x86 example. 08:59:26 and the perfect language to write it in of course 08:59:53 write it in C such that thereafter it can write itself - or the same in forth, whatever 08:59:58 PoppaVic: well, there we go 09:00:05 you need a metacompiler 09:00:08 PoppaVic: have you designed your lib? 09:00:27 good lord no.. I got ideas and notes all over 09:00:53 err, i meant vm 09:00:58 can I work with you on this? 09:01:00 requiring ansi or posix C would prolly work fine, they even have cross-compilers 09:01:09 I'm trying to find a meaningful C project to tag along on 09:01:12 you can comment and contrib, sure. 09:01:43 maybe it will end up being a tagteam project. Right now, it's just one of those "damn the torpedoes.." things 09:01:44 beautiful 09:01:44 so basically the only problem with LLVM is that it's written in C++, is that it? 09:02:02 PoppaVic: so, do tell, how far are you with the vM? 09:02:16 C++, stl, templates - and I was having issues finding just plain old file-formatting 09:02:16 might want to stay away from C-- then, its written in ocaml 09:02:27 yeah, I've seen that 09:02:56 ridiculous. not even investigating something because you dont like the language it's in 09:03:02 if it were in shell script, that'd be one thing 09:03:02 arke: at this point, as I said, I've been building up notes and "differences" so I know where to begin 09:03:35 PoppaVic: aah. I have an old .txt file somewhere with some VM ideas.. 09:04:55 PoppaVic: register based or stack based or both? 09:06:03 --- join: PoppaVic_ (~pete@0-2pool238-171.nas24.chicago4.il.us.da.qwest.net) joined #forth 09:06:15 --- quit: PoppaVic (Nick collision from services.) 09:06:17 --- nick: PoppaVic_ -> PoppaVic 09:06:44 PoppaVic: aah. I have an old .txt file somewhere with some VM ideas.. 09:06:45 PoppaVic: register based or stack based or both? 09:07:08 you can't rely on regs, not and suffer w/o asm or C 09:07:26 and, the only stack I so far expect is a datastack 09:07:40 structured, i take it 09:07:50 define 09:08:22 you mean OO? no 09:08:34 no, not OO 09:08:37 but without GOTO blah 09:08:42 if you mean a whole lot of c structs, yes 09:08:55 well, i guess theres gotta be goto 09:09:17 goto is a C keyword 09:09:22 another stack based vm. yay. 09:09:29 09:09:34 ^^ 09:09:46 lua's got a register vm, and it's small enough to learn from it 09:09:58 COMEFROM owns GOTO. 09:09:58 great 09:10:10 at the end it will be just a special forth system for portable installation. 09:10:13 Robert: yer humor is as usual 09:10:20 PoppaVic: Below the bottom? 09:10:20 parrot also has one but it's made by the same people who gave us SVpQgFqTj **pzIvAL type identifiers in perl5 09:10:33 virl: basically, plus the C interface 09:11:04 you could use retro for that 09:11:11 sproingie: !?!?!!1111 09:11:31 perl5's c source is horrid 09:11:31 ehm, bullshit sry. retro is asm. 09:12:01 and where did i mention rf in this? 09:13:17 don't forget FICL, too 09:13:18 not you, I mentioned rf. and that was stupid. 09:14:07 ah. i still don't fathom how you're going to avoid some portability concerns somewhere 09:14:28