00:00:00 --- log: started forth/02.06.01
00:16:12 --- quit: sif (Read error: 110 (Connection timed out))
00:57:33 --- join: davidw (~davidw@adsl-66-101.38-151.net24.it) joined #forth
01:32:34 <rob_ert> Good morning, all!
03:34:13 --- join: mur (ammu@baana-62-165-189-112.phnet.fi) joined #forth
03:36:34 --- quit: mur (Client Quit)
05:29:08 --- quit: Stepan (Read error: 110 (Connection timed out))
05:29:18 --- join: Stepan (~stepan@p5084783E.dip.t-dialin.net) joined #forth
05:42:14 --- join: Fare (fare@samaris.tunes.org) joined #forth
05:44:37 <rob_ert> Hi Fare.
05:44:40 <rob_ert> And wb Stepan.
05:48:05 <davidw> I need a name for this ecos/ficl thing
05:53:57 <davidw> I'm going to call it Zog
05:54:04 <davidw> maybe
06:12:03 <davidw> come on guys
06:16:41 <rob_ert> No idea =)
06:20:41 <Fare> what does ficl do?
06:21:27 <davidw> it's a forth
06:22:06 <Fare> would it run on arm?
06:22:14 <Fare> is it written in (yuck) C ?
06:22:21 <davidw> yes and yes
06:22:33 <rob_ert> Why not Visual BASIC? :(
06:22:40 <davidw> vb isn't portable
06:22:44 <davidw> C is
06:22:55 <Fare> yes, VB is portable to all systems that matter
06:23:07 <rob_ert> Who needs anything else than Windows?
06:23:38 <davidw> the non-iq deprived?
06:23:56 <Fare> what deprived?
06:24:16 * rob_ert is just trolling
06:24:20 <rob_ert> Ignore me for now :)
06:37:34 <Fare> you have a son? how old?
06:40:06 * rob_ert wonders if Fare is in the right channel.
07:10:12 <Fare> indeed, not
07:15:36 --- join: sif (~siforth@ip68-9-58-81.ri.ri.cox.net) joined #forth
07:15:36 <sif> Type sif: <forth code here> (or /msg sif <forth code here> to play in private)
08:29:45 --- join: Etaoin (~david@ljk3-7.sat.net) joined #forth
08:31:43 --- join: Speuler (~l@195.30.184.4) joined #forth
08:31:46 <Speuler> g'day
08:42:23 <Etaoin> hi
08:42:29 <Fare> lo
08:47:26 --- quit: Speuler (SendQ exceeded)
08:49:45 --- join: Speuler_ (~l@195.30.184.4) joined #forth
08:50:08 <Speuler_> hi agn
08:50:16 <Speuler_> brin.openproj is dead
08:50:19 <Speuler_> just fied
08:50:20 <Speuler_> died
08:50:44 <Speuler_> or was there a netsplit a few moments ago ?
08:51:16 --- join: Speuler (~l@195.30.184.4) joined #forth
08:51:17 --- quit: Speuler (Remote closed the connection)
08:51:26 <Speuler_> oh
08:51:46 <Speuler_> weird things going on ..
08:52:07 --- nick: Speuler_ -> Speuler
09:45:45 --- join: bugslayer (~thin@h24-64-175-61.cg.shawcable.net) joined #forth
09:47:40 --- quit: bugslayer (Client Quit)
10:24:33 --- join: I440r (~mark4@1Cust40.tnt2.bloomington.in.da.uu.net) joined #forth
10:25:20 <rob_ert> Hey :)
10:25:24 <I440r> hi rob :)
10:58:59 --- join: thefox (fox@adsl-209-182-168-45.value.net) joined #forth
10:58:59 --- mode: ChanServ set +o thefox
10:59:39 <I440r> thefox!
10:59:43 <I440r> hi :)
11:00:16 <I440r> i need to recode parse-word to delimit on EOL or the specified char
11:00:41 <I440r> word/parse-word and parse in isforth are basically laxen and perry code
11:00:48 <I440r> and its obfuscated
11:01:16 <I440r> ive walked thru that code mentally a gazillion times and i STILL cant figure out how it works
11:01:16 <I440r> heh
11:03:06 --- join: ASau (ASau@158.250.48.197) joined #forth
11:03:28 <I440r> asau!
11:03:29 <I440r> hi
11:03:37 <ASau> Hello!
11:03:46 <I440r> was it you that had that forth printf code or aum ?
11:04:15 <ASau> No
11:04:41 <ASau> I had no need to write printf
11:04:55 <I440r> heh me either
11:05:02 <I440r> but it was interesting code :)
11:05:17 <I440r> but that the time i didnt have a compiler capable of actually COMPILING it heh
11:05:30 <ASau> I haven't seen it
11:05:59 <I440r> it was a ." that allowed you do do things like x y z ." blah blah %n blah blah %s blah blah %d"
11:06:29 <Speuler> I440r: i got an implementation doing something like this.
11:06:47 <Speuler> allows inserting strings, nums, and some other stuff
11:06:53 <I440r> speuler cool. i could code it in 10 minutes lol
11:07:01 <ASau> I see no need to do anything like this
11:07:01 <I440r> yes %s
11:07:12 <ASau> Why I cannot use
11:07:17 <Speuler> I440r: i'm sure you can
11:07:41 <ASau> ." ...." D. ." ..." . ." ...." ( etc)
11:07:59 <ASau> I think this looks better
11:08:19 <Speuler> ASau: won't works if the string is generated instead of stitched together
11:08:30 <ASau> No need to use man page for all this formatting issues
11:08:31 <I440r> asau yes but for every ."   " your comping a (.") and a dot - the %s %m %i mechanism would compile smaller code
11:09:06 <ASau> I440r and code to do all this formatting...
11:09:25 <I440r> asau forth does perfectly without printf formatting
11:09:34 <I440r> but it would be an interesting experiment :)
11:09:41 <Speuler> but t'is basically trivial stuff
11:09:42 <I440r> what i need to be doing now is coding SEE
11:09:46 <ASau> That's what I say
11:09:47 <I440r> so i can write the forth debugger
11:10:03 <ASau> But I have more interesting problems
11:10:32 <I440r> asau me too - "see/debugger/assembler/meta-compiler" and then some
11:10:57 <Speuler> helps also inserting line breaks, form feed and the like into a single string
11:11:09 <I440r> isforth now memorymaps its source files but im still grabing chars one at a time out of the map and copying them into tib
11:11:30 <I440r> will be much faster if i made the memory map itself TIB
11:11:49 <I440r> set #tib to filesize
11:12:04 <I440r> but for that i need parse-word to delimit on EOL and the specified character
11:13:12 <ASau> ... CONSTANT EOL  ( ...) EOL WORD ( ?)
11:13:45 <ASau> Are sitting under UN*X?
11:13:50 <I440r> asau no i mean. when someone does  x word - where x can be a '"' or a bl etc
11:14:14 <I440r> parse-word should parse UP TO that character. but if it sees an EOL before the specified delimiter. it must stop THERE
11:14:47 <ASau> Hmm, WORD stops at end of line...
11:15:03 <ASau> Ahh, QUERY, stops at eoln!
11:15:05 <I440r> asau no because word normally never SEES eol
11:15:19 <I440r> query and expect dont store the EOL in tib
11:15:25 <I440r> they set #tib
11:15:39 <I440r> word will delimit on end of tib reached or the specified character
11:15:43 <ASau> Well, I use Fig-Forth dialect.
11:15:55 <I440r> but when the WHOLE FILE is now in tib you have problems
11:16:04 <ASau> My QUERY stores EOLN in TIB
11:16:32 <I440r> asau and what comes AFTER that eol in tib?
11:16:36 <I440r> blank unused space
11:16:38 <ASau> Rewrite WORD. Or change EOLN to zero
11:16:43 <I440r> or junk data
11:17:08 <I440r> asau i need to recode parse-word
11:17:23 <I440r> 1: memorymap the specified source file
11:17:30 <I440r> 2: point 'tib at the memory map
11:17:38 <I440r> 3: set #tib to the file size
11:17:47 <I440r> 4: revector refill to UNMAP the mapped file
11:17:55 <I440r> then just interpret as normal
11:18:34 <I440r> the inner interpreter will keep doing bl word find state if , else execute then nothing left until
11:18:45 <I440r> for the WHOOLE file
11:20:03 <ASau> Well, I've seen one my FIG-f. version interpretes  command line
11:20:19 <ASau> I just changed EOLN to blanks
11:20:26 <I440r> asau isforth can interpret the command line too...
11:20:34 <I440r> ./isforth - : foo blah blah ;
11:20:48 <I440r> launches isforth with foo defined because isforth interpreted the tail
11:21:11 <Speuler> wow
11:21:11 <ASau> Make the WORD not to distinguish EOLN and BL
11:21:28 <I440r> asau still the same problem, still need to parse up to the specified delimiter OR the blank/0x0a
11:22:02 <ASau> I see no problem: try to change WORD.
11:22:18 <I440r> asau THATS the problem hehe
11:22:20 <ASau> What do you miss in this case?
11:22:37 <I440r> the "word" or rather in this case "parse-word" im using is from FPC
11:22:49 <I440r> which is originally from F83
11:22:54 <I440r> i.e. laxen and perry code
11:23:10 <I440r> ive single stepped through that code in fpc.
11:23:15 <I440r> dry ran it in my head
11:23:19 <ASau> I've not seen L&P code. Have you got it?
11:23:21 <I440r> over and over and over....
11:23:32 <I440r> and i have absolutely NO idea how it works
11:23:42 <I440r> i could cut and paste the code
11:24:31 <ASau> Look at Fig-f ENCLOSE and WORD code.
11:24:50 <ASau> For WORD there're cources form FIG.
11:24:52 <I440r> colon 'parse-word', parseword
11:24:52 <I440r>  dd tor
11:24:52 <I440r>  dd source, tuck
11:24:52 <I440r>  dd toin, fetch, sstring
11:24:52 <I440r>  dd rfetch, skip
11:24:53 <I440r>  dd over, swap
11:24:55 <I440r>  dd rto, scan
11:24:58 <I440r>  dd tor
11:24:59 <I440r>  dd over, minus
11:25:03 <I440r>  dd rot, rto
11:25:05 <I440r>  dd dup, znotequals, plus
11:25:07 <I440r>  dd minus
11:25:09 <I440r>  dd toin, store
11:25:16 <I440r>  dd exit
11:25:16 <I440r> flood :)
11:25:24 <I440r> its that dup 0<> + - >in ! that throws me every time
11:26:12 <I440r> >r source tuck >in @ /string r@ skip over swap r> scan >r over - rot r> dup 0<> + - >in !
11:26:19 <I440r> obfuscated
11:26:40 <Fare> I440r: did you look at ccg?
11:26:51 <I440r> ccg ?
11:27:25 <I440r> i know how i can code the above much simpler, but i would need to modify the return vbalue of scan
11:27:59 <I440r> scan returns the address of the delimited character plus the number of characters LEFT in the parsed string AFTER that character
11:28:02 <ASau> What SOURCE does?
11:28:11 <ASau> rto is R> ?
11:28:21 <I440r> word would be much simpler if scan returned the string plus a lenght
11:28:26 <I440r> rto is >r
11:28:29 --- nick: Speuler -> PhoodPhrenzy
11:28:32 <I440r> and source just returns 'tib and #tib
11:29:17 <ASau> rto is >R ...  tor is >R ?
11:29:47 <I440r> no r-to to-r  rto is r>  tor is >r
11:29:56 <ASau> OK!
11:30:08 <I440r> those are assembler labels placed on the CFA of each word
11:30:21 <I440r> had to think of a name the assembler would like yet was still forth-LIKE
11:30:33 <ASau> : SOURCE 'TIB #TIB ; ( ? )
11:30:34 <I440r> it gets confusing tho when you use rot >r and r> in the same word :)
11:30:58 <I440r> : tib 'tib @ ;
11:31:06 <I440r> : source tib #tib @ ;
11:31:06 <thefox> r> is pronounced r-from
11:31:34 <I440r> to r and r to - i always pronounce '>' as TO
11:31:40 <I440r> so does my father heh
11:32:09 <I440r> except when its all by itself of corse
11:32:15 <ASau> (I used to say FROMR for R>...)
11:32:35 <ASau> : TUCK  ( ??? ) ;
11:32:40 <thefox> I can understand that.  I was suprised to read the pronunciation for a number of Forth words that I had always pronounced a certain way.
11:32:57 <I440r> : tuck swap over ;
11:33:39 <ASau> : SSTRING ( ??? ) ;
11:33:44 <I440r> thefox isforth is already totally NON standard - would it be realy bad if i changed the return values of scan ?
11:34:25 <I440r>  heh /string is easy it takes address and count and n1. it adds n1 to address and subtracts n1 from count
11:34:27 * rob_ert seems to be the one in here not thinking in english.
11:34:54 <I440r> : /stringdup negate under + under + ;  <-- leet l&p code :)
11:35:03 <I440r> erm /string dup  <-- inject a space where i missed it :)
11:35:05 <ASau> : /STRING ( str -- addr count ) ( ??? ) ;
11:35:33 <I440r> : /astring ( a1 n1 n2 --- a1+n2 n1-n2 )
11:35:45 <ASau> : /STRING  DUP CHAR+ SWAP C@ ; ( ??? )
11:35:57 <I440r> argh.. mental note must learn to type
11:36:26 <I440r> advance a1 by n2
11:36:29 <I440r> decrement n1 by n2
11:36:57 <I440r> it clips the front end off a string :)
11:37:47 <ASau> sstring is /STRING ?
11:38:02 <ASau> : /STRING ( str -- addr count ) ( ??? ) ; ( am I right? )
11:38:29 <I440r> if scan would return ( a1 n1 c1 ---- a2 n2 ) where a2 was the address of the scanned up to character and n2 was the number of characters skipped it would be easier
11:38:36 <I440r> asau yes /string is called sstring in the asm code
11:38:40 <I440r> but your stack coment is incorrect
11:38:52 <ASau> ok
11:39:06 <ASau> What is correct?
11:39:15 <I440r> : /string ( a1 n1 n2 --- a1+n2   n1-n2 )
11:39:19 <I440r> add n2 to a1
11:39:26 <I440r> subtract n2 from n1
11:39:30 <ASau> Ahh... Ok!
11:39:39 <I440r> n1 is the length of the string at a1
11:40:02 <ASau> I'm slowed down :)
11:42:31 <thefox> I just looked it up, r> IS pronounced r-from  I always have to look up a lot of stuff in ANS, then I don't always interpret it right, but in this case it clearly says the pronounciation is r-from
11:43:09 <I440r> thefox it might be r-from in ans but how does CHUCK pronounce it!
11:43:37 <ASau> I've read ANS, a little... I don't use ANS.
11:44:09 <ASau> toin is >IN ?
11:44:15 <I440r> yes
11:44:34 <ASau> >IN @ is address of current pos. for INTERPRET ?
11:45:00 <thefox> I remember that my friend Dr. Michael Montvelishsky always said DUP with a short u like up instead of a long u like TO as in DUPlicate.
11:45:02 <I440r> >in is how far into 'tib we have parsed so far
11:45:40 <ASau> i.e. >IN is displacement  (?)
11:45:50 <I440r> how do you pronounce "char" - i know alot of people pronounce it like "car"
11:45:54 <I440r> asau yes
11:46:41 <ASau> I pronounce, as "tscharr", as it's mentioned by MLG
11:47:41 * I440r applies linseed oil to the 1911 grips he is making for his 45....
11:47:46 <I440r> from bubinga!
11:48:15 <thefox> like care.  char is pronounced exactly like care.  as in I care about that
11:48:52 <I440r> i always pronounced it as it is spelled.. like the ch in church
11:49:39 <thefox> it is short for character where the first sound is `care' 'act' 'er'   care like air
11:50:12 <I440r> i know... heh i just always pronounced it the way it was spelled :)
11:50:16 <ASau> i440:  : SKIP  ( a cnt c -- a' cnt' ) ( ... ) ;  ( ? )
11:50:22 <thefox> k-air
11:50:53 <I440r> asau skip is used to skip leading characters
11:51:07 <I440r> skip all leading blanks is what its doing here
11:51:11 <ASau> i440: ok!
11:51:21 <I440r> or rather, all leading DELIMITERS
11:51:26 <I440r> which is USUALLY a blank heh
11:51:30 <ASau> i440: SCAN waits for delimiters?
11:51:49 <I440r> scan scans UP TO a delimiter
11:52:00 <I440r> it scans until it sees the delimiter
11:52:22 <ASau> OK! Make EOLN a DELIMITER.
11:52:45 <I440r> asau the delimiter is currently specified by whoever calls skip or scan
11:53:10 <I440r> i need to make scan stop on TWO delimiters
11:53:20 <I440r> the one specified or an eol
11:53:24 <Fare> ccg is a runtime assembler in C++ by ian piumarta
11:53:25 <I440r> but i cant modify scan
11:53:25 <ASau> Tsch\"ort! ;)
11:53:43 <Fare> it's nice, but what's nicer is what he does with it
11:53:58 <Fare> look for squeak's j3
11:54:03 <I440r> what does he do with it ?
11:54:27 <Fare> generate code
11:54:31 <Fare> optimized
11:54:50 <I440r> i can fix my problem but the fix BREAKS something
11:54:57 <I440r> i either rewrite scan
11:55:04 <I440r> or i rewrite parse-word
11:55:16 <I440r> thats my delema
11:55:27 <I440r> plus the fact that parse-word is totally unreadable :)
11:55:51 <I440r> fare i am diametrically opposed to optimizations done by an assembler or a compiler
11:56:02 <I440r> it destroys the 1:1 co-relation between source and object
11:56:12 <ASau> Rewrite parse-word. It's more simple task.
11:56:23 <ASau> Rewriting it breaks nothing
11:57:02 <thefox> I think a 1:1 correspondence is very powerful because it helps in a powerful feature of forth, treating code as data and data as code
11:57:29 <I440r> asau the rewrite for parse-word has 2 paths. 1: write a special version of scan 2: rewrite scan altogether
11:57:40 <I440r> thefox exactly
11:57:55 <I440r> it also makes debugging the code an order of magnitude easier
11:58:13 <I440r> optimize that forth/assembler code and you get very fast very difficult to debug code
11:58:18 <ASau> i440 follow #1.
11:58:44 <I440r> asau thats what i was thinking
11:59:07 <thefox> That is also something that may be hard to do with someone else's off the shelf Forth where distribution restrictions apply and hence a good place for diy.
11:59:23 <ASau> i440 Or you can change all 0x0a to BL.
11:59:47 <I440r> asau nope
11:59:59 <I440r> asau how would \ work then
12:00:29 <I440r> \ parses up to the next eol
12:00:40 <I440r> and discards whats parsed
12:00:41 <ASau> I don't use \
12:00:48 <Fare> I440r: no optimizations from a compiler???
12:00:59 <I440r> i use it ALOT
12:01:08 <I440r> fare none what so ever.
12:01:09 <Fare> not even in code generation for 2 2 + ?
12:01:12 <I440r> not any compiler I write
12:01:16 <I440r> tho...
12:01:23 <ASau> In this case such changing is not appropiate
12:01:29 <I440r> fare who is smarter. me or the compiler
12:01:29 <Fare> A @ 2 + A ! ?
12:01:51 <I440r> fare i would personally always code 2 a +! there
12:02:00 <Fare> you're not smart if you don't teach your compiler to be less dumb
12:02:05 <I440r> I should be doing the optimization, NOT the compiler
12:02:16 <I440r> fare i disagree.
12:02:37 <thefox> compiler optimization arguments are very much bound to architecture.
12:02:37 <I440r> there was a thread on CLF not long ago that suggested ALL development should be done on an old 8088 slowest machine you can find
12:03:18 <I440r>  a quote from the poster "always distrust coders who insist on using fast machines to develop their code on"
12:03:34 <I440r> fare who is smarter... the compiler? or the coder!
12:03:41 <thefox> in most discussions in c.l.f people always are speaking just about the architectures that they use, but they rarely say for .... such and such is true.  They represent it as universal truth. ;-)
12:03:51 <I440r> the coder might be using some trick that the compiler then optimizes out of existence
12:04:11 <I440r> thefox :)
12:04:21 <onetom> hi
12:04:25 <I440r> thefox my latest "semi arguement" with clf was on recursion
12:05:01 <I440r> deep down inside me i have this gut feeling that ANY recursive algorithm has a much nicer itterative counterpart.
12:05:18 <I440r> that would include things like ackerman. the problem is we just didnt discover them yet
12:07:00 <thefox> one might notice that what Chuck calls recusion isn't what other people think of as recursion.  He uses no smudging so using a word name inside of the word just compiles a call to itself, recursion.  But when combined with ; and tail-recursion it really just becomes iteration equivalent to begin again.
12:07:53 <I440r> thefox i dont smudge either but a word isnt visible to find until its placed at the head of some hashed vocabulary thered
12:08:14 <I440r> actually, i am smudging but it dont test that bit anywhere
12:08:22 <thefox> so you can force it visible and use it recusively?
12:08:41 <I440r> : recurse last @ name> @ , ;
12:08:48 <thefox> ok
12:08:57 <I440r> last always points at nfa...
12:09:00 <thefox> immediate
12:09:03 <I440r> of the word being defined
12:09:09 <I440r> erm yes lol
12:09:27 <I440r> well spotted :P~
12:11:00 <I440r> less than a second to compile 20 source files of about 116k total
12:11:09 <I440r> i think i can double that maybe
12:11:14 <I440r> or maybe even more
12:11:32 <onetom> dont just think, but do it !
12:11:55 <I440r> i did however expect the memory mapping would give ALOT better results that it has so far :P
12:11:59 <thefox> I wrote a number of Forth compilers with various degress of optimization for different architectures but eventually became very fond of what Chuck recommended.  The 1:1 is there in machineForth although it does optimize most stuff by compiling inlined opcode instead of nested calls.  But it is like a really simple token threaded system because the opcodes are like tokens that don't require a software layer to interpret.
12:12:03 <I440r> onetom i AM doing it heh
12:12:27 <I440r> even if it doesnt improve the speed much it DOES simplify things elsewhere
12:12:44 <I440r> fload is basically a replacement for query/expect AND the interpret loop right now...
12:13:18 <I440r> doing this would make it a replacement for query/expect ONLY... floaded files would just be interpreted by the same inner interpreter you use interactivly
12:13:33 <I440r> no need for a "special" version of it for files
12:14:53 <onetom> thefox: "don't require a software layer to interpret" -- what do u mean by that? could u explain a bit more:
12:14:53 <onetom> ?
12:15:40 <onetom> thefox: u were refeing 2 a machine4th chip, werent u?
12:16:11 <I440r> thefox ive never "studied" the way chuck thinks but from the 2 interviews here and from what you say im finding myself in agreement with him on almost everything :)
12:17:05 <thefox> onetom: yes.  But I noticed that Chuck used the same code inside of colorForth on Pentium.
12:17:47 <onetom> thefox: but there *was* a sw layer in that color4th, wasnt there?
12:17:54 <I440r> thefox when chuck codes in assembler within forth does he like the "5 # ax mov" stuff?
12:18:04 <I440r> i cant handle asm like that, its just all WRONG to me...
12:18:19 <I440r> but alot of forth coders perfer it
12:19:05 <onetom> thefox: what forth do u use nowdays?
12:19:21 <thefox> In a token threaded Forth like OF a bytecode may get shifted to produce an index into a table filled with the addresses of the words.  It gets done in software on most machines, but if the 'tokens' that get threaded are opcodes the assembler is a token threaded Forth with a 1:1 correspondence like other threaded forths.  hardware token threaden
12:19:24 --- nick: PhoodPhrenzy -> CaffeineJunkie
12:19:33 <onetom> thefox: or which forth, if its a suitable question
12:21:15 <thefox> I have a couple of windows forths when I have to, and FPC and several machineForths.  I will bootstrap the aha compiler to the aha operating system, but I may port the machineforth that I am using as a cross compiler first to that environment.  There is the bootstrap problem.  How much do you cross develop and when do you go to the target for dev?
12:21:46 <onetom> thefox: woooow, thanx 4 the explanation. i didnt think of token threadin this way earlier
12:22:36 <thefox> it also means that if you compile rule-based systems directly to Forth that the hardware is the inference engine, no software layer needed.
12:23:31 <onetom> thefox: im only developing 4 microchip's pic controllers
12:24:38 <onetom> thefox: but they r so small @ tied 2 hardware, so rarely develop on them
12:25:05 <thefox> chuck mentioned them the other day.   He would use them for many things that other people would think require a PC.
12:25:33 <onetom> thefox: what is a rule-based system, anyway? is it a declarative programming related stuff?
12:25:36 <thefox> Developing on a pic, now there's a challange!
12:25:59 <I440r> thefox some of the newer pic's are very well suited to forth im told
12:26:15 <onetom> I440r: sure, the newer 1s r bigger
12:26:33 <onetom> I440r: but the really small 1s require crosscompilation
12:26:39 <I440r> yes :)
12:27:05 <onetom> thefox: do u know the instruction set of the pics?
12:28:28 <thefox> it can be, in the sense that that is what prolog does, but it just means decision trees.  You give a programmer a set of logical rules and they construct a solution with nested IFs or whatever.  Chains of logic statements.  One of those things where mixing code and data can be very powerful.  Building executable data structures is a nice Forth technique so it is easy to conpile logic rules into executables
12:29:21 <thefox> I don't know much about the pic line, just the basics. (bad pun intended)
12:30:35 <thefox> I saw that Al Mitchel had written a basic compiler on top of his Forths and could duplicate the functionality of some basic stamps with it, but 1000x higher performance according to the presentation write up.
12:31:25 <onetom> hmm, i never heard about it b4
12:31:37 <I440r> who is al mitchel ?
12:31:40 <thefox> chuck seems to like the little pic stuff
12:31:51 <onetom> so, do u know enough about the 2 answer this question? :
12:32:13 <onetom> is it possible 2 implement a machine4th on pic?
12:32:21 <thefox> al mitchel research, former FIG board member, full time embedded Forth shop, lots of neat stuff and good stories.
12:32:39 <onetom> good stories <-- :))
12:32:55 <I440r> any jobs goin ??? :)
12:33:08 <onetom> he must have a homepage then
12:33:39 <thefox> oh, sure I always like hearing about how he replaced 386 motherboards with 8051 running Forth etc.  He has a wide range of target machines, cool software tools, and lots of experience.
12:34:21 <I440r> thefox im gona finish my direct threaded 8051 forth :)
12:34:33 <I440r> built for small size, not speed...
12:34:43 <I440r> well. eventaully ill finish it
12:34:45 <thefox> I am sure that one could write a machineforth for pic.  I have no idea what would and or would not be efficient however since I am just not that familiar with the architecture.  But since Chuck seems to think it is ok... I would just guess that it might not be too inefficient for that.
12:35:05 <onetom> thefox: r mitchels stuff also available 2 zthe public?
12:35:16 <thefox> amresearch.com ?
12:36:06 <I440r> thefox siemens has some meters that use 386ex's that ause some RTOS and huge ammounts of c code
12:36:12 <thefox> yes, http://www.amresearch.com worked nicely
12:36:28 <I440r> i told em i could recode the whole thing on an 8051 and they laughed at me
12:36:46 <thefox> al has cool stories like Forth Inc. about beating the crap out of other vendors.
12:37:37 <thefox> why make it simple when you can make it wonderfully complex?
12:37:55 <I440r> thefox does forth inc get alot of business then ?
12:38:04 <I440r> or are they still a struggling company
12:38:24 <thefox> al stuff is commercial, but he has given lots of people good ideas.  His systems are very cheap and his hardware is well made.
12:39:37 <thefox> He built very nice tethered control cross development systems for a wide range of tiny micros long before it became fashionable.
12:40:10 <onetom> hmm hmmm -- he is talented enough :)
12:41:32 <thefox> Phil Koopman once wrote an article about determinism where he showed that the problem with 386 on real-time was that about one time in a thousand an interrupt would take 100x the average interrupt!  So you only needed a few extra orders of magnitude of hardware and software to meet any hard spec!
12:43:36 <thefox> Forth Inc. has had hard times and boom times, I don't know how many people work there today and they are always looking for more business.
12:44:54 <I440r> i wouldnt mind working there myself but i doubt they would hire me, im too outspoken agains ANS (non)forth
12:45:07 <thefox> But they have those great stories about coming in as the back-up team competing against the main team that was falling behind schedule and then demonstrating how efficient Forth is for both development time and runtime, beating the competition by an order of magnitude in each.
12:45:41 <I440r> or more
12:45:52 <thefox> If you want to work at Forth Inc. you would need to become very familar with their products.
12:46:41 <I440r> i like the story of that compettion where they had to make a hacksaw blade with LED's on it swing back and forth and display some character. the forth team had a scrolly banner before anyone else had even found the resonant freqency of the hacksaw blade :)
12:46:42 <thefox> yes, not just program size but speed and power consumption etc.
12:47:40 <thefox> The first time I used Forth at one job the engineer had not made it back to his desk before I called him back with the software solution to his latest hardware problem.  Before that he hated programmers.
12:48:30 <onetom> :))
12:48:49 <I440r> thefox lol. first time i used forth at work was for joy mining machenery. the JNA RS-20 long wall coal mining roof support system was coded entirely in forth...
12:48:58 <I440r> chipforth for the 8051 to be exact
12:49:38 <thefox> I would worry about windows running the software!
12:49:54 <I440r> i found a bug in their code after 2 weeks of being there. the bug had been known about but undiscoverd for 10 years
12:50:13 <I440r> for some unknown reason a unit would just SHUT DOWN
12:50:29 <thefox> an expert bug hunter
12:50:56 <I440r> it was totally unpredictable, didnt happen very often (yes i love hunting bugs but this i discovered by accident)
12:51:10 <thefox> not another bug hunt? (Aliens)
12:51:13 <I440r> they needed a way to flash priority messages on the LCD then restore whaqt was there previously
12:51:17 <I440r> so i had to code a dpuble buffer
12:51:58 <I440r> and with the double buffer in place the aforementioned "inexplicable HALT" happned every time
12:52:14 <I440r> so i just put debug stuff in the code and traced it back to the bug
12:52:55 <I440r> the problem was with the ADC - the same register you use to program it had a "halt" bit in it
12:53:04 <I440r> that was being accidently set under certain conditions
12:53:06 <onetom> thefox: :))) nice saying
12:54:12 <I440r> they should realy have found and fixed that bug the day found out abou tit
12:54:28 <I440r> for a roof support to just HALT every now and again is VERY dangerous
12:55:05 <thefox> We had one programmer at iTV who found most of the bugs.  It really bothered other people that this guy kept finding their killer bugs.  And he didn't like doing it either. He much prefered writing bug free code to finding nasty hiding bugs that other people created.
12:56:05 <I440r> thefox sounds like my kind of coder. i had a little discussion in #tunes the other day about releasing buggy code and letting the end users beta test it for me....
12:56:14 <I440r> if its a bug i know about i FIX it before release
12:56:24 <thefox> That's what I meant about doing it with PC's running windows.... please insert the installation CD AND ...
12:56:35 <I440r> actually thats not quite true, i let one slip till just recently
12:57:11 <I440r> tehfox lol that would work great 2 miles down a mineshaft on a PC with NO keyboard, NO monitor and NO cd :)
12:58:03 <I440r> you couldnt have any of the above down a mine shaft, might make something go BOOM
12:58:49 <I440r> you cant even carry ANYTHING made of aluminum down a mine, it starts to spark whenh you get deep
12:58:57 <thefox> One time we had what the programmers thought was a funny bug.  Users and sales didn't like it.  We had an 'inhouse' switch which when set caused the system to print out completely accurate error messages for us in development, or when off the politically correct error messages for the users.  It went out the door one day with the switch wrong and the users had no sense of humor about their own errors!
12:59:00 <I440r> i dont know if they understand WHY it does it
12:59:29 <I440r> lol
12:59:36 <I440r> thats a good way to lose customers :)
13:00:37 <thefox> we got 200 phone calls by 8:00 am
13:00:56 <I440r> tehfox you didnt answer my question about "5 # ax mov" in a forth assembler. does chuck prefer it thaty way ?
13:01:02 <thefox> we sent out a new release by noon. but we thought it was funny... bad bad programmers!
13:01:05 <I440r> i know alot of forth coders do
13:01:12 <I440r> do you perfer it that way ?
13:01:29 <thefox> chuck uses that kind of syntax in assembler
13:01:43 <I440r> actually, i think having a "debug" version of software is bad
13:02:00 <I440r> you can never guarantee that the bug you fixed in the debug version is fixed in the release version
13:02:13 <I440r> or that the two will act thte same way under the same conditions
13:02:42 <I440r> i cant get it straight in my head. i started with assembler and moved on to forth. whenm i code forth i think forth. when i code assembler i think assembler
13:02:53 <thefox> our development version gave more verbose messages, "the idiot user just tried to ..." oops.
13:03:08 <I440r> lol
13:03:34 <I440r> i dont think i would allow that sort of error message in a debug version just incase it got released by accident :)
13:03:43 <I440r> very unprofessional i thin
13:03:45 <thefox> when I think assembler/forth on forth chips it doesn't come up.  but you can use any syntax you want in a Forth assembler.
13:03:47 <I440r> think
13:04:52 <I440r> thefox what i realy want for isforth is an assembler extension that can take the existing NASM sources with VERY LITTLE modification....
13:04:58 <I440r> prolly a pipe dream
13:04:58 <thefox> it was a policy long before I got there.  I guess it was just a type of error (setting directives wrong) that no one imagined could ever happen.
13:05:02 <I440r> and WAY beyond what i can do
13:05:16 <I440r> self in foot shoot :P
13:05:40 <I440r> lol thats the ONLY comment on my "recurse" word
13:05:45 <I440r> heh
13:05:59 <thefox> In the old days, I like using debug assembler/dissasembler format so I could use the rom code.
13:06:26 <onetom> thefox: i was discussing ur thoughtful programming article w my friend nowdays
13:06:30 <thefox> I meant DOS code
13:06:47 <onetom> thefox: but they cant believe what u state.
13:07:01 <thefox> I have been working on a short essay inspired by something that Chuck said here a while back.
13:07:09 <I440r> debug assembler/disassembler format ?
13:07:12 <onetom> thefox: they need an evidence. a kind of example code-pair
13:07:36 <thefox> many people go into denail.  What don't they believe? the many examples of large ratios?
13:07:44 <thefox> denial
13:07:58 <onetom> thefox: so could u help me out w such a pair of inefficient-efficient code pair?
13:08:33 <onetom> thefox: they would like 2 see the sources of those examples
13:08:50 <onetom> thefox: so they can reveal the causes of those high ratios
13:08:53 <thefox> people who don't want to believe will always say that you have to publish code owned by other people to prove them wrong, knowing that it can' be done.
13:09:28 <thefox> I have tried to explain it before, but usually if someone doesn't want to hear it they will shout you down and not listen anyway.
13:10:07 <I440r> thefox people dont like hearing "your doing it the DUMBASS way"
13:10:42 <thefox> As Chuck has said, he will continue to publish examples that will clearly demonstrate to anyone who is interested ...
13:11:39 <thefox> Yes, the problem is simply that Forth is too good.  If done right it apparently shames other people so much that they feel personally insulted even if you make no comparisons at all.
13:12:02 <thefox> You have heard my story about the first time Chuck gave a public speech about F21?
13:12:11 <Fare> I haven't
13:12:16 <I440r> nope
13:12:26 <onetom> thefox: dont know which 1 do u think of
13:12:53 <thefox> Chuck came down to SVFIG.  He talked about the development of F21, what he was doing with OKAD, how fast the chip was etc.
13:12:58 <onetom> thefox: ive already prepared all the stuff from ultratech being watched
13:13:25 <thefox> I was very worried that the might make comparisons to C or Intel as he often does and offend someone.  I watched very carefully for that.
13:13:47 <I440r> lol
13:14:01 <onetom> loll
13:14:14 <thefox> I was pleased that Chuck made no comparisons to C or ANS Forth or Intel architecture or VHDL or any of that stuff.  I was pleased that he had not said anything that gets him into trouble.
13:14:46 <thefox> I saw this guy in back who I did not know, had never seen before.  After Chuck's presentation I went up to him and asked him what he thought of Chuck's presentation.
13:14:56 <thefox> He said he HATED it!
13:15:06 <onetom> "ezzel azangol szoveggel, hatha tenyleg azt kerhetek amit akarok ;)"
13:15:11 <onetom> sorry
13:15:21 <I440r> my guess is that when the spade is a spade chuck isnt too scared to state it as such :)
13:15:29 <thefox> He said that he felt very insulted because Chuck had called him an idiot for using C.
13:15:49 <thefox> Chuck had never mentioned C once.
13:16:18 <I440r> did he refer to "other languages" instead of stating "c" outright ?
13:16:52 <thefox> Apparently people listen to Chuck and a voice in the back of their mind says to them, "If what this guy is saying is true, if he can do all that then I am an idiot for using C and for only being able to do 1% of what he can do.
13:17:09 <thefox> Then they go into denial.  Therefore what he says cannot be true.
13:17:14 <I440r> :)
13:17:14 <I440r> they are correct
13:17:43 <thefox> This problem is even worse in the Forth community where so many people want to think that they know more about Forth than Chuck.
13:18:00 <I440r> the trouble is that acadamia has sold the industry a crock of shit and they have baught into it hook line and sinker
13:18:41 <I440r> thefox they probably DO know more... thats the baauty of chuck, he filters out all the CRAP other people build in...
13:19:00 <thefox> chuck had only talked about what he was doing.  He had not mentioned other people using C, VHDL, Intel, etc. etc. not a word.
13:19:51 <I440r> when your a smart guy standing next to someone like nikola tesla you REALY FEEL DUMB!!!!!
13:20:01 <I440r> i would put chuck up there with tesla
13:20:02 <thefox> I think industry has sold academia a crock and acedemia bought it and made it what they teach and perpetuate.
13:20:30 <I440r> thefox a vicious circle :)
13:20:37 <I440r> were both probably right :)
13:20:45 <thefox> I think Chuck is one of the smartest people I ever met.
13:20:57 <thefox> Chuck says that that isn't it, that he just tries harder.
13:21:20 <thefox> Chuck is usually right, so he is probably right about that too, it makes a nice Zen Koan.
13:21:21 <I440r> he is clear thinking. he has the ability to SEE through all the CRAP and get right to the heart of the matter
13:22:16 <I440r> thats also part of it... "this is acceptable to me so ill release it now"  <<--  chuck has a higher standard than other people
13:22:21 <thefox> yes, and he tries very hard, perfectionistically, almost compulsively to get the essentials.
13:22:29 <I440r> probably TOO high, 5 years to write 4 lines of code is too much :)
13:23:18 <thefox> he did write a little other code in those 5 years, but it is a funny line. ;-)
13:23:19 <I440r> i dont like using users as beta testers but ill release not so optimum WORKING code and refine it over time...
13:23:21 --- quit: Stepan ("Client Exiting")
13:23:42 <I440r> thefox i was exaggerating for effect :)
13:24:14 <onetom> thefox: so what could u advice 4 me. how can i break the denial?
13:24:31 <thefox> For those who don't know Chuck was asked how long it took to write OKAD.  OKAD had actually been written and rewritten about 5 times in 5 years and kept getting the inner loop cleaner and cleaner until it got down to 4 lines of code. thus Chuck said four lines of code in five years.
13:24:38 <onetom> thefox: there is no reference source 4 such a comparision available?
13:25:37 <I440r> thefox is oakad saleable ?
13:25:49 <I440r> would any chip manufacturers want to buy ?
13:26:04 <thefox> yes, no, apparently not.
13:26:33 <thefox> onetom: what comparison?
13:26:38 <I440r> doh
13:27:08 <onetom> thefox: sw comparision what proves that ratio
13:27:26 <onetom> thefox: src of 2 programs what solve the same problem
13:27:38 <onetom> thefox: but 1 of the is terribly faster
13:28:19 <thefox> the examples I refer to in that paper were the many examples where the same module got coded twice.  Once by an ANS Forth programmer and once by a MachineForth programmer. iTV owns the code.
13:28:47 <onetom> i c :(
13:29:37 <thefox> Many people have denied the numbers, some even deny the stories.  But there were a dozen of us, some very active in the Forth community where were in all those meetings for years.
13:31:15 <thefox> We had many many cases where we had modules that did the same thing from the systems point of view, but where performance made all the difference in the world.  The first jpeg decode, translated from C was so big, so slow, and produced such ugly images that it was removed from the app for a year.
13:32:26 <thefox> The well coded machineforth version was about 100 times smaller, 1000 times faster, and looked about 10 times better.  It was impressive to see these instant decode jpeg that looked better than vga in b/w.
13:33:28 <onetom> but the original C version and its 4th rewrite produced absolutely the same result?
13:33:50 <thefox> You can't really talk about this stuff in public just because people go into denial and argue that it can't be true, or they come up with their own explanations, oh, you must have meant .... instead.
13:35:17 <onetom> but how about the origianl code?
13:35:23 <onetom> isnt it free?
13:35:27 <thefox> jpeg was our best application module example.  the use of ANS Forth, Forth Scientific Library, code for C style structures was perhaps the strongest example.
13:35:59 <onetom> 4 the 1st place, eg, my friends said u have choosen the wrong C code
13:36:07 --- join: miket2 (Mike@62.60.91.78) joined #forth
13:36:17 <thefox> But just bringing it up upsets people and then you are not allowed to explain.  We even had FIG board members denying that they ever heard of the FIG Pres and VP.
13:36:17 <onetom> it was an inefficient implementation even in C
13:36:56 <I440r> hi miket2
13:37:06 <thefox> yes, but C allows a nieve implementation to get 'optimization' by the compiler.  It is expected.  There is also lots of assumptions that people make about hardware that they won't admit to themselves.
13:37:06 <miket2> hello
13:37:52 <I440r> miket2: is a forth coder ?? :)
13:38:15 <onetom> thefox: Forth Scientific Library is a strong example 4 what?
13:38:26 <thefox> Well it was Chuck who made the statement that if it isn't 100 times smaller than C it isn't Forth.  Again, people feel insulted, the screen it out, they don't want to understand.
13:39:30 <thefox> A strong example of how ANS Forth practice was usually at least 10x larger, often 100x larger, and usually 100x sometimes 10,000x slower than what the machineForth programmers would write given the same requirement for a module.
13:39:52 <I440r> that has a tendancy to insult both forth and c coders :)
13:40:23 <thefox> not to mention java, perl, etc.
13:40:39 <I440r> they dont count
13:40:43 --- join: bob4th (~bob4th@adsl-63-197-120-243.dsl.sktn01.pacbell.net) joined #forth
13:40:43 <I440r> they are scripting languages :P
13:40:50 <thefox> It also upset them that the ANS Forth modules had 50 times more bugs.
13:41:16 <thefox> But there were the best experts that the company could hire!
13:41:46 <thefox> people will tell you that any language can do anything that any other language can do. ;-)
13:41:52 <I440r> hi bob :)
13:41:54 <onetom> does http://www.ultratechnology.com/mfp21.htm is good 4 learning machine forth?
13:42:23 <I440r> thefox thats like saying you can build a house out of balsa wood and have it as strong as an oak house
13:43:18 <thefox> there are a number of machineForth tutorials at my site, more to come.  That is the text from one of them.
13:44:46 <onetom> thefox: NOP 7FFF # A! NOP 1 # A NOP NOP +  <-- why is there those nops? (quote from the above article)
13:44:50 <thefox> I plan to go back over the many tutorials that we posted many years ago that apparently needed more introductory tutorials.  The examples of GUI components with various features were meant as hands on tutorials for the simulators.
13:46:09 --- quit: bob4th ()
13:46:21 <thefox> nops are used to align multiple data loads or stores into one instruction word to reduce instruction bandwidth and maximize on-page memory access.  nop is required BEFORE + if carry has to propagate more than a certain number of bits in the
13:46:33 <thefox> + instruction the way Chuck's adder works
13:46:45 --- join: bob4th (~bob4th@adsl-63-197-120-243.dsl.sktn01.pacbell.net) joined #forth
13:47:37 <I440r> wb bob :)
13:48:05 <I440r> thefox ans people wouldnt like that, they would want the compiler to INJECT those nops for you
13:48:07 <onetom> thefox: so what happens wo those nops?
13:48:26 <onetom> I440r: thats it :)
13:48:28 <I440r> i think ans is a conspiracy to make forth more c like
13:49:20 <I440r> i usually dont type "Forth" with an upper case F because im lazy. i never type c with an upper case c because it doesnt deserve it:)
13:50:55 <thefox> ans is just an effort to standardize on certain people's practices to add more validity to their products.  They also chose to market products more closely tied to the prevailing C based environments.  It's just business.  It is not a conspiracy.  But Forth Inc. sponsored the idea, send multiple repsentatives, chaired every meeting, and now have final authority to interpret what ANS means in all those things where they intentionally 'waffle'
13:52:13 <thefox> the compiler can inject nops if you want.  I prefer not.  W/O them the code is smaller, but slower, or depending on the arguments to the + it might not produce the proper result.
13:52:37 <thefox> Chuck could have included the nop delays in the + instruction so you would never have to add them.
13:53:12 --- part: miket2 left #forth
13:53:35 <thefox> BUT Chuck does scaled math. This is one of the most important things in Forth.  He can use + differently to save a few nano-seconds when he knows the size of the arguments.
13:54:07 <thefox> he can use + or nop + or nop nop + or nop nop nop + (on the 20 bits) etc.
13:54:50 <onetom> oy, c... tho, i have 2 study these stuff deeper 2 understand what did u mean by "reduce instruction bandwidth and maximize on-page memory access" :/
13:54:57 <thefox> gee that's hard to read  "+" or "nop +" or "nop nop +" etc.
13:55:05 <I440r> but.... you have to THINK about what your coding... nonononono that will NEVER do!
13:55:32 <I440r> :)
13:56:10 <onetom> thefox: yeah, it looks pretty ugly
13:56:16 <thefox> getting close to pentium hardware is difficult, many instructions, deep pipelining, several layers of cache, etc.  Very hard to really get close.  But with these chips it is easy to get close to the limits of memory.
13:56:23 <I440r> whats more you have to understand the harware your running on. people dont want to bother with that sort of stuff, let the compiler worry about it...
13:56:33 <onetom> thefox: and hides the actual algorithm, making it harder 2 modify...
13:56:34 <thefox> In either case that may be the biggest factory, the layout and use of memory.
13:56:43 <I440r> tehfox i never liked intel processors
13:56:49 <onetom> thefox: ceratinly its a cost 4 effectivity, tho...
13:56:50 <I440r> love their controllers tho :)
13:58:33 <onetom> thefox: shall we go back 4 the sci lib 4 a while?
13:58:51 <onetom> thefox: u say, thats inefficient, right?
13:59:03 <thefox> If you try to discuss actual hardware details many people go into denial.  They seem to think that the hardware will do lots of things fast that it can't.  They don't want to be bothered with how many layers there are between what the hardware really does and what they can get to.
13:59:08 <onetom> thefox: so is there a machine forth counterpart of it?
13:59:27 <thefox> what about the fsl?
13:59:58 <onetom> is there a machine4th implementation of the fsl available somewher?
14:00:23 <thefox> that's an oxymoron.
14:00:37 <onetom> oh, really? ... hmm
14:01:13 <onetom> libs in machine 4th is an oxymoron in general?
14:01:17 <thefox> machineforth is about embracing the machine. machineforth was about making Forth 1:1 hardware software.
14:01:48 <thefox> fsl was about translating portable c code into portable Forth code by adding lots of stuff to do runtime testing to get portability.
14:02:12 <I440r> a definate oxymoron
14:02:19 <I440r> portable code always is
14:02:27 <onetom> runtime testing??? like what?
14:02:43 <thefox> They represent opposite ends of the spectrum.  Every .f file had a .h file associated with it.  Everything was mostly transcribed from C into Forth script.
14:02:45 <onetom> i have the fsl page open in my browser
14:03:04 <onetom> aha..
14:03:12 <I440r> a terminilogical inexactitude to be exact (did i spell that rite?)
14:04:12 <thefox> ANS 'waffles' on about 100 things that you just can't count on.  So runtime code has to test for this and this and this and then decide which version of the code is appropriate for this platform...  and they didn't do it at compile time, like in C, but at runtime! ugh.
14:04:30 <I440r> tehfox im torn as to weather i should port some of the c header files to isforth the #defines would be converted to consts which will eventually be discarede on turnkey - my #reson for NOT doing so is it will generate inter-dependancies in the sources
14:05:13 <I440r> also, naming EVERY fscking BIT in a byte makes th4e code almost unreadable
14:05:30 <I440r> you have to go hunting for the header file that defines the bit equate to find out what it equates to
14:05:41 <I440r> and sometimes even THAT isnt possible
14:06:18 <I440r> #define foo (fudge << 5 ^ (bar  >> 6) & (wtf << 3))
14:06:21 <I440r> ugh
14:06:24 <thefox> For instance in the code for C style structures there was an option diagnostic message to be printed out by every instance of a strucure when enabled.  It used a double number.  So instead of DOES> + @ ; or something like that it had a 2DUP near the start and lots of 2OVERs and 2SWAPs throughout and a vectored diagnostic at the end that was revectored to 2DROP for embedded use.
14:06:51 <I440r> lol
14:07:04 <thefox> all those 2over 2swap etc were not needed just to do a 2drop but there it was.  The code was 20 times more complex than it needed to be, but in C style grab it from a library and paste it in!
14:07:40 <I440r> cut and paste/drag and drop coding
14:07:48 <onetom> and cant it b coded smarter?
14:07:48 <I440r> they way of the future !!!
14:08:29 <I440r> so instead of coding your code to fit the problem you code it to fit existing libraries
14:08:29 <I440r> lol
14:09:23 <thefox> the code was so lazy and hogish in its stack use that the author never considered doing it machineforth, he gave up the first 50x by requiring ANS, then he grabbed from a 'SHINING EXAMPLE' of an ANS Forth library and pasted resulting an another 20x factor.  Instant 1000x without trying.
14:10:28 <thefox> but people go into denial.  former fig president? never heard of him!
14:10:40 <I440r> thats why it was 1000x
14:10:43 <I440r> NOT trying
14:11:25 <thefox> We were told 1000 times that it was just the idea behind ANS, code re-use, portable libraries, not having to deal with machine specifics, not having to try.
14:11:51 <thefox> it's just called 'embracing common practice' not a conspiracy.
14:12:10 <I440r> i.e. turn forth into something c already is
14:12:28 <I440r> something that any IDIOT can code
14:12:45 <onetom> thefox: but using libraries is not always a drawback, isnt it?
14:12:55 <thefox> that was one of hundreds of similar examples.  People who don't want to believe it, or don't want others to believe it just won't.
14:12:57 <onetom> eer, is it? :)
14:13:22 <I440r> some guy gave me a rating of "gree belt forth coder" he said if i could code this recursive function itterativly he would give me blue :)
14:13:22 <thefox> I say use libraries.  But don't plagurize!
14:13:34 <I440r> i said at least he didnt rate me as white or yello :)
14:14:07 <I440r> thefox if the library fits the problem yes. but its like... dont RELY on the library doing so
14:14:31 <I440r> my father says software engineers today rely on debuggers and in circuit emulators FAR TOO MUCH
14:14:35 <onetom> thfox: plagurize?
14:14:39 <I440r> they want the debugger/ICE to find the bug for them
14:14:40 <thefox> You should have your own library that you can use because you know what is in it.  You should have other libraries that you can learn from.  Blindly cutting and pasting from other's libraries may not be the source of all evil, but ...
14:14:46 <I440r> they dont what to have to THINK
14:15:07 <thefox> Yes, libraries are for learning, not for stealing.
14:15:48 <onetom> thefox: i just didnt understand the word & failed 2 look it up in the dict..
14:15:51 <thefox> grabbing from libraries w/o thinking is what I complain about, it is not what Forth is about.  Forth requires thoughful programming, not programming w/o thought.
14:16:00 <I440r> c coders keep asking me why i have to keep reinventing the wheel
14:16:10 <I440r> i usually answer "because the last coder coded it square"
14:16:45 <thefox> plagurize  (sp?)  to copy someone else's work without giving them credit.  To claim that someone else's work is your work.
14:16:49 <onetom> thefox: but time is so pressing usually, that not thoughtful programming have 2 be used
14:16:54 <I440r> thefox THAT,and the rpn is what makes most coders go screaming for mommy when they look at forth
14:17:09 <onetom> thefox: eg, i know i have 2 do some matrix transformations
14:17:15 <thefox> you can invest time now, or later.
14:17:37 <I440r> thefox thats not always the case tho, people can reinvent... im sure my hanoy solution isnt MINE exactly...
14:17:42 <onetom> thefox: i have 2 produce a program what needs those transformations
14:17:43 <thefox> people want to skip the understanding and design phases and jump to coding.  Then they spend 90% of there time debugging.
14:17:47 <I440r> but i did dream it up all by myseph
14:18:20 <I440r> "do you want it right now.... or right"  or "if we dont have the time to do it right first time where the hell are we going to get the time to do it over"
14:19:15 <onetom> thefox: but i can understand a problem pretty well wo knowing how 2 implement efficient matirx transformator/operator code
14:19:19 <I440r> tehfox actually i dont think its possible to FULLY design any code without trying to write it...
14:19:38 <I440r> there are implecations that do not make themselves apparant until you start coding
14:19:46 <I440r> the design should be a general overview i think
14:20:02 <thefox> I found that by doing it in Forth instead of ... I could write it, and rewrite it, get it right, then port it to the target language they wanted.  Fast, compared to ... and higher quality.  That always kept them happy.
14:20:11 <onetom> I440r: design & implementation have 2 b iterated
14:20:18 <I440r> a place to start thinking about how to code... not a "nitty gritty down to the finest detail design"
14:20:39 <I440r> onetom AFTER you have the implementation
14:20:41 <thefox> I440r: that's why Chuck says that some code deserves to be rewritten.
14:20:53 <I440r> you start making demands on implementation at design time and you are going to cripple the design
14:21:27 <I440r> thefox i think ALL programs should be written at least twice as a minimum :)
14:21:40 <I440r> write it, get it to work. DELETE IT and start over :)
14:21:42 <I440r> the second time round will be better
14:22:22 <onetom> thefox: but if the primary problem is carrying out some matrix operations, probably u dont want 2 solve already solved problems, and wont go and implement your own matrix opertaion. aint i right?
14:22:29 <thefox> in my short 'forth methodology of Chuck Moore' essay I suggest that we design, redesign, redesign, code, redesign, recode, and then go back and re-think the problem.
14:22:56 <I440r> thefox but not add-infinitum :)
14:23:02 <thefox> I suggest that we do that until we get it right.  This is not needed for everything, but it is the nature of Forth.
14:23:13 <onetom> k
14:23:22 <thefox> only until your dead
14:23:23 <onetom> thefox: thanks 4 the lesson :)
14:23:51 <I440r> thefox coding is like art... its never realy finished, just abandoned :)
14:24:05 <thefox> yes, you want, and need, a library of canned things to re-use.  But you must understand them at least once.
14:24:56 <onetom> thefox: i only have 2 understand their surface, their API 2 b able 2 use them...
14:25:18 <onetom> thefox: thats why they r called libraries :)
14:25:27 <onetom> oooor.... :D
14:25:39 <thefox> Chuck has libraries.  I have libraries.  I have some of Chuck's code in my libraries, with some of my changes to it.  Chuck has looked at my code and used a couple of ideas, but not actual code in his libraries.  He says that what libraries are good for.
14:25:51 <onetom> they rather b read & studied :))
14:27:04 <onetom> so u state, there is no "omipotent" form of code for a problem in forth?
14:27:15 <thefox> when you only understand the api you can only use it in certain ways.  It may also lead to problems that you won't easily find.
14:27:21 <onetom> in a specific implementation of 4th @ least...
14:28:10 --- join: BadCog (spamtrap@1Cust158.tnt1.beaverton.or.da.uu.net) joined #forth
14:28:12 <onetom> thefox: aha... api is also a constraint... sure, ive already understood it, ithink
14:28:23 <thefox> Forth is the only omipotent part.  libraries are like extensions that you have not compiled on top of Forth.  Forth isn't the libraries.  The libraries depart from Forth toward application domains.
14:28:37 <I440r> hi BadCog
14:28:40 <onetom> :)) omnipotent, actually
14:28:44 <BadCog> hi
14:29:41 <BadCog> what's the subject?
14:29:52 <I440r> forth :)
14:29:53 <onetom> low fat programming, ithink :)
14:29:59 <I440r> and how it beats the crap out of everything else :)
14:30:01 <thefox> Chuck says that given almost any problem he would start the same couple K of code known as Forth.  Other people can't write a hello world without megabytes of extra baggage at the start.
14:30:08 <onetom> I440r: stoopid ;p
14:30:47 <BadCog> as far as extensions go, in the book Thinking Forth, they talk about components, not librarys
14:31:00 <BadCog> i think that components are smaller than librarys
14:31:07 <I440r> tehfox thats the problem with linking to libc to do the I/O and ncurses to do all the extra fancy cursor control to put the "hello world" in the right place :)
14:31:10 <BadCog> a component could consist of only a few Forth words
14:31:11 <thefox> Some of these people write in Forth, but say that 'REAL' forths must have every bell and whistle available off the shelf to everyone else.
14:31:59 <BadCog> One should atleast get a metacompiler with your Forth, so you can strip out the unnneded words if necessary
14:32:02 <thefox> When you write a complete system you can change the lowest level as easily, more easily, than the highest level.
14:32:08 <I440r> sounds like tom zimmer :)
14:32:23 <I440r> badcog thats whats missing from isforth :(
14:32:28 <I440r> i CANT strip them out yet :(
14:32:36 <BadCog> too bad
14:32:51 <I440r> need an assembler first and thats beyond me :(
14:32:53 <onetom> thefox: why isnt there a much more complete forth system available?
14:33:03 <onetom> thefox: w networking support mainly...
14:33:21 <onetom> thefox: and w more correct video interface
14:33:40 <BadCog> you mean a Linux Forth, correct?
14:33:48 <onetom> thefox: so u shouldnt use linux or windows - like u ;p - 4 irc-ing.
14:33:52 <thefox> Chuck knows how every word in his Forth works and will change the lowest level words to make an application easier.  But people who require 50M of someone else's code to start, and code that they can't change I might add, will end up having to write much much more code than Chuck because they are writing for the restrictions that they choose.  The more problems that they introduce, and then solve the better the feel.
14:34:01 <thefox> Chuck just avoids as much of those problems as possible.
14:34:08 <I440r> badcog isforth IS a linux forth - NOT written in c
14:34:26 <thefox> onetom: that's a straw man argument.
14:35:07 <BadCog> there is another assembly Linux forth i beleive, just a sec
14:35:42 <I440r> eforth
14:35:46 <thefox> I did use only Forth code that we wrote to do everything on the net a while back.  If I say that I can write 100 times more efficiently than ... that I should expect to duplicate the work of ten million other programmers.
14:36:30 <thefox> there is a thing called bootstrapping.  Did you write the compiler? did you write the OS? did you design the
14:37:21 <I440r> and lib4th of corse
14:37:33 <thefox> chip? did you debug it? did you design a board? did you debug it? Did you then duplicate the work of 10M other programmer? No, we just claimed 100/1 not 10,000,000/1
14:39:50 <onetom> :) okay
14:39:51 <BadCog> what about bigForth
14:40:14 <I440r> badcog bigforth is coded in c
14:40:19 <I440r> at least partially
14:40:20 <onetom> thefox: do u know enth/flux?
14:40:23 <I440r> and its ans :)
14:40:31 <I440r> and the assembler is a bit weird,,,,
14:40:38 <I440r> tho i do have permission to use it
14:40:40 <CaffeineJunkie> thefox: hi.
14:40:48 <BadCog> I'm looking at a site that says its a Native Code Forth
14:40:54 --- nick: CaffeineJunkie -> Speuler
14:40:55 <onetom> I440r: but u can use it a non-ans way 2, if u r smart enough
14:41:14 <Speuler> hey
14:41:21 <Speuler> i typed that 3 minutes ago ...
14:41:21 <thefox> onetom: I help Sean design and develop it and posted the first information about it.  But I have't used it.
14:41:25 <I440r> badcog it produces native code i believe... isforth doesnt. it produces direct threaded FORTH code
14:41:31 <I440r> it extends the kernel with the application....
14:41:34 <onetom> I440r: anyway, its pretty fast and compiles native code, so u shouldnt really blame it, just because its ans
14:41:56 <I440r> onetom bigforth is "ok" by me
14:42:03 <I440r> but its still not a TRUE forth
14:42:07 <Speuler> thefox: do you know other sources for i21/f21 than mr ting ?
14:42:43 <Speuler> (if mr ting sells f21 boards at all i mean)
14:42:48 <I440r> isforth is has the potential of becoming the ONLY true forth for linux
14:42:57 <I440r> but im deficient one assembler and one meta compiler
14:43:00 <onetom> thefox: do think enth/flux is the way 2 go i want a small forth only environment 4 pc?
14:43:24 <BadCog> why does having native code generation make it not true forth?
14:43:48 <onetom> thefox: probably color4th is smaller, but that reduced keyboard... and its highly incompatible w most of the hw
14:43:53 <I440r> badcog it does not compile forth sources ONTO itself, its a target compiler only...
14:43:57 <onetom> s
14:44:01 <thefox> Dr. Ting owns Offete Enterprises Inc.  Offete owns MuP21.  I own UltraTechnology.  UT owns F21.  iTV owned i21.  I was a stockholder in iTV.  Ting and I are also together on other projects like 25x.
14:44:04 <I440r> i dont even know if it can create stand alone applications
14:44:25 <I440r> can you create an executable to run on any x86 machine even if it doesnt have bigforth installed there ?
14:44:38 <onetom> thefox: why is all those named ...21?
14:44:51 <BadCog> i don't know
14:44:57 <onetom> neither i :)
14:45:00 <I440r> badcog you CAN with isforth
14:45:05 <onetom> but guess so
14:45:05 <thefox> they are all in the x21 family, originally x20 family, 20 bit data bus, 21 bit alu.
14:45:10 <I440r> but you cant make coded defs yet:(
14:45:28 <BadCog> I'm just using Win32Forth in windows 98
14:45:30 <I440r> nor metacompile. you can rebuild the kernel though... with the nasm assembler
14:45:44 <BadCog> I might make the switch to SwiftForth, but I don't know yet
14:45:56 <I440r> badcog use FPC before that ?
14:46:02 <I440r> swiftforth isnt free
14:46:10 <I440r> isforth IS
14:46:17 <I440r> but isforth is linux only
14:46:21 <thefox> There was also p32, P32, P8, and v21 made for Chuck's other clients.  p21 and F21 were the only chips intended for outside sales.
14:46:22 <BadCog> No, Forthmacs on Atari St
14:46:23 <Speuler> thefox: i tried 3 times to order an i21 prototype board. so i wonder whether there's another possibility to get one
14:47:09 <thefox> swiftforth has a free evaluation version
14:47:12 <I440r> BadCog: isforth is being designed to have the same FEEL as FPC, but the code base is 100% written from scratch and my inner interpreter is VERY different from fpc's
14:47:15 <I440r> which is obfuscated
14:47:17 <Speuler> i think i spoke to mr tings family. not quite sure, as communication wasn't easy
14:47:18 <I440r> thefox i know
14:47:24 <I440r> can you create turnkeys wit it ?
14:48:22 <BadCog> You can turnkey with Win32Forth, but not with the Free eval Swift.  You gota pay for that
14:48:26 <BadCog> i think
14:48:46 <I440r> badcog i know... it was a rhetorical question :)
14:49:00 <BadCog> :)
14:49:26 <I440r> badcog isforth is free. you can create stand alone turnkey apps with it. you can even create NON GPL applications with it
14:49:40 <BadCog> Win32Forth is free
14:49:42 <I440r> and isforth is lisenced under a modified LGPL
14:50:01 <thefox> Dr. Ting's father runs the mail order store.  His English is very weak and he knows nothing about technology.  He can fill an order but not answer quesions.  You might be able to reach Dr. Ting at chenhting@yahoo.com.tw
14:50:01 <I440r> but.... its windows
14:50:09 <I440r> automatic loser on that score heh
14:50:15 <BadCog> Does isforth have a gui interface for creating other gui interfaces?  Does it write object orientated code while you fiddle around in the gui builder?
14:50:38 <I440r> badcog not yet. but an xlib interface IS planned
14:50:55 <BadCog> ok
14:50:56 <I440r> and NO it will NEVER write object obfuscated code
14:51:04 <thefox> speuler: i21 will be hard to get since iTV seems to be out of business.  Harder if you try to get it from a compeditor.
14:51:12 <I440r> and it will NEVER SUPPORT drag and drop gui building
14:51:18 <I440r> EVER
14:51:40 <I440r> thefox who owns itv's intelectual rights ?
14:51:46 <BadCog> thats not maybe so bad if you have good docs for the xlib interface
14:51:48 <Speuler> thefox: ok. that was about 5 years ago. what's f21 and x25 availability state ?
14:51:58 <thefox> itv's prefered investors
14:52:20 <I440r> badcog isforth is deficient in the dox area too so far :(
14:52:46 <I440r> my problem is that i can code ANYTHING. but writing the dox for it is a drag :P
14:52:52 <I440r> but i WANT it fully doccumented
14:53:19 <I440r> my code makes me look 1337 - my english makes me look dumb
14:53:21 --- part: ASau left #forth
14:53:37 <thefox> x25 has not been prototyped. It exists only in cad.  there are still a couple f21d prototype chips from 98 availble for any serious use.
14:54:07 <Speuler> thefox: please define "serious use"
14:58:00 <thefox> Well due to four sources of funding all failing at once I ended up with a company that needed to move forward with a handful of chips that each cost a small fortune.   I could have sold them all in a week for $10 each and been out of business.  I gave many away to people who showed serious interest, meaning they intended to do something.  I knew this because they had downloaded the free tools and written code that they could show me.
14:58:34 <thefox> They contributed software tools and/or hardware tools to help move the project forward.  I tried not to sell them to people who would do nothing with them.
14:59:42 <BadCog> its still too bad that I can't walk into K-mart and pick up a half way decient browser for $49
15:00:01 <thefox> People who wrote code or designed hardware showed a serious interest.  I did not want computer-consumers to expect me to give them complete systems with for-idiots type manuals.
15:00:24 --- join: jamc (~dne@as3-6-8.asp.s.bonet.se) joined #forth
15:01:06 <thefox> or better yet a $5 appliance that slips into the battery compartment on a $19 TV.
15:01:25 <BadCog> well yeah
15:01:37 <BadCog> then the kids don't have to tie up YOUR tv when you want to watch the game
15:01:59 <thefox> or a handful of programmable $1 appliances, but that requires enough interest to do production.
15:02:33 <BadCog> with the economy the way it is, seem like youd have a hard time finding investors
15:02:38 <thefox> iTV almost closed a couple of deals for hundreds of millions of such units.
15:03:04 <thefox> hundreds of millions in the testing phase, if they liked them then they wanted large quantities...
15:03:48 <thefox> well I need to go do some stuff off-line. best wishes to all
15:03:55 <BadCog> bye
15:04:01 --- quit: thefox ()
15:04:48 <Speuler> (was looking for a forth cpu core for chip cards)
15:05:04 <Speuler> got me the shboom / psc1000 then
15:05:18 <BadCog> whats in going in?
15:05:23 <I440r> badcog i privmsg'd you
15:05:50 <BadCog> didn't work
15:09:25 --- quit: bob4th ()
15:10:29 --- part: jamc left #forth
15:43:13 --- join: jamc (~dne@as3-6-8.asp.s.bonet.se) joined #forth
15:43:52 <jamc> I440r: back :)
15:44:01 <I440r> :)
15:44:09 <I440r> much better :P
15:44:15 <I440r> how dare you leave !!!!!!!!! lol
15:45:07 <rob_ert> Hej jamc.
15:45:09 * jamc blushes :)
15:45:17 <jamc> yo, robban
15:45:36 <rob_ert> I440r: Swedes are taking over your world.
15:46:03 <I440r> rob that cus they are good coders.
15:46:03 <I440r> *.se are coders :)\
15:46:12 <I440r> the rest of the word are coder wannabe's
15:46:17 <rob_ert> Heh.
15:46:24 <rob_ert> I'm not a real swede then :)
15:46:58 <I440r> lol
15:47:47 <jamc> I440r: if it's nasm, I think openbsd is out of the question... I think it only produces ELF objects, does it not?
15:48:02 * jamc continuing here :)
15:48:18 <I440r> jamc it can produce many different object formats
15:48:42 <I440r> isforth however only produces elf executables right now but THAT is easy to fix :)
15:48:50 <jamc> ok, maybe my copy is too old
15:49:32 <jamc> you seem to be right... i just did nasm -hf
15:49:40 <onetom> openbsd? object production? what?
15:50:10 <jamc> onetom: I440r wanted to port his forth to bsd
15:50:38 <I440r> jamc doesnt open bsd use elf file?
15:50:39 <onetom> ah, got it
15:50:50 <jamc> I440r: nope, still a.out
15:51:13 <I440r> would require someone knowledgable about XXXbsd's (open, free, net) syscalls
15:51:13 <I440r> at the ASSEMBLEr level
15:51:26 <I440r> jamc making isforth save a.out would be a SNAP
15:52:17 <jamc> where can I get the source?
15:52:55 <Speuler> the_rob: sweden plays england tomorrow ?
15:53:09 <rob_ert> Maybe...
15:53:20 * rob_ert is not a football freak :)
15:53:47 <Speuler> could start winning the cup for taking over the world :)
15:54:01 <rob_ert> Hmm... good idea.
15:54:28 <Speuler> some norwegians come here tomorrow. they support the swedish team
15:54:46 <Speuler> provided the match is tomorrow
15:54:46 <onetom> rob_ert: hello, how is goin?
15:54:46 <rob_ert> :)
15:54:58 <rob_ert> onetom: Fine..working on a FAT implementation.
15:55:16 <Speuler> bye, night and over
15:55:24 <rob_ert> Bye.
15:55:28 --- quit: Speuler ("Bye.")
15:58:50 <jamc> I440r: ok, google found your source :)
15:59:37 <I440r> jamc yea hehe thats quite recent, i deleted my sourceforge account ages ago
15:59:48 <I440r> and google only just located my new site :)
15:59:59 <I440r> which im very pleased about :) heh
16:00:08 <jamc> hehe
16:01:21 <I440r> im downloading freebsd right now, the iso. the first iso is going to take another 60 hours to download
16:01:33 <I440r> then i have the second iso to grab too
16:02:49 <jamc> openbsd is what i run here at home
16:04:38 <I440r> does open bsd use the same syscall numbering as free bsd ?
16:04:38 <I440r> and net bsd ?
16:04:49 <I440r> are they all similar kernels?
16:04:49 <I440r> or all different
16:05:02 <jamc> they are similar, but different :)
16:05:48 <jamc> but net/open can run each other's, linux and freebsd's binaries (syscall emulation layer)
16:07:16 <I440r> jamc isforth wont run in freebsd, nasm produces an executable elft type ZERO
16:07:18 <I440r> unknown to free bsd
16:07:43 <jamc> try the linux binary
16:07:57 --- quit: Fare ("Connection reset by pear")
16:08:36 <I440r> jamc i need to be able to assemble the isforth kernel under free bsd (net/open) and produce a NATIVE executable
16:09:25 <jamc> ok.. so it doesn't help to bootstrap it under emulation... i mean it doesn't do "save-image"?
16:10:34 <I440r> if you compile the sources AS IS you basicaly get a linux executable written out to a BSD format
16:10:39 <I440r> bsd will assume its a bsd executable
16:10:50 <I440r> and the syscalls are going to be b0rken for you
16:11:31 <jamc> yeah i see
16:12:41 <I440r> so. you need to be able to assemble the kernel with all the syscalls recoded for what ever OS you run
16:12:54 <I440r> syscalls.1 is where the syscall INTERFACE is
16:12:56 <jamc> just tried "make freebsd", but my nasm doesn't seem to like "%define"
16:13:22 <I440r> but there are references to syscalls throughout the code3 and the syscall numbers referenced etc will need to be changed for YOUR os
16:13:27 <I440r> no
16:13:32 <I440r> your nasm is old
16:13:38 <I440r> you need to update
16:13:38 <I440r> its %xdefine
16:13:51 <I440r> %define in nasm is FUCKED IN THE HEAD
16:13:54 <jamc> oh yeah... still doesn't like it :)
16:13:55 <I440r> %xdefine is a fixed %define
16:14:06 <jamc> macros.1:4: unknown preprocessor directive `%xdefine'
16:14:06 <jamc> macros.1:4: label or instruction expected at start of line
16:14:25 <I440r> %xdefine is not defined in your version  of nasm
16:14:32 <I440r> you need to get the latest version
16:15:05 <I440r> badcog im good at juggling :)
16:15:13 <BadCog> i'm not so good at it
16:15:18 <I440r> jamc hers whats wrong iwth %define in nasm
16:15:24 <I440r> %define x 1
16:15:29 <I440r> %define y 2
16:15:32 <I440r> erm %define y x
16:15:35 <I440r> %define x 3
16:15:47 <I440r> what are the values for X and Y in the above
16:15:53 <I440r> forget the %define y 2 - thas a mistake
16:16:10 <jamc> i have no idea... never coded nasm :)
16:16:20 <I440r> both are the same
16:16:27 <I440r> y is defined as a POINTER to x
16:16:42 <I440r> i.e. you have absolutley NO fucking way to SAVE the value of X in Y for later restoration
16:16:54 <I440r> xdefine fixes this
16:17:06 <jamc> well most macro systems suck rotten eggs
16:19:09 <BadCog> taking off
16:19:12 <BadCog> bye
16:19:24 <I440r> not the a86 assemblers :)
16:19:24 <I440r> that one is good :)
16:19:31 <I440r> later badcog!!
16:19:35 --- quit: BadCog ()
16:19:43 <I440r> a86 is so good i baught it lol
16:25:31 * jamc building newer nasm
16:25:45 <rob_ert> bought :-)
16:26:00 <rob_ert> Anyway... 16-bit NASM 0.98.33 sucks :) Bugs all over.
16:27:20 <jamc> uhm 0.98.08 was the one i found... is it also too old?
16:28:12 <rob_ert> Hehe
16:28:17 <rob_ert> It's a bit incompatible.
16:28:24 <rob_ert> It defaults to near jumps, iirc.
16:28:31 <rob_ert> Or, no
16:28:35 <rob_ert> short jumps
16:29:35 <I440r> robert lol
16:30:50 <jamc> so which version do I need, and where can i find it?
16:31:30 <I440r> erm which one are you building now
16:31:43 <jamc> 0.98.08
16:31:48 <I440r> try nasm.2y.net
16:31:59 <I440r> get their latest. i think that one works... it does in debian at lest
16:32:10 <I440r> 98.08 worked in debian
16:32:31 <I440r> continue building it and see if it makes it possible to build my kernel
16:32:57 <jamc> yup
16:33:15 <I440r> working ?
16:33:23 <I440r> cool
16:33:28 <jamc> not tried it yet
16:33:32 <I440r> make freebsd or make linux
16:33:40 <I440r> freebsd build is ALL FUCKED UP
16:33:50 <I440r> because i do not know what syscall number equates to swhat syscall
16:34:00 <I440r> or now many parameters any given syscall takes in freebsd
16:34:10 <I440r> but i think the freebsd.1 source file is basically correct
16:35:01 <jamc> 0.98.08 at least doesn't complain about %xdefine
16:35:10 <I440r> its just the syscall numbers that are regferenced in there and in the isforth/forthsrc/*.f files that are wrong
16:35:13 <jamc> but now there's the ELF problem :)
16:35:34 <I440r> then you can build just fine
16:35:43 <I440r> yes
16:35:51 <I440r> you need to modify the asmsrc/Makefile
16:35:53 <jamc> yup... got kernel.o in ELF format
16:35:58 <I440r> to write to an a.out
16:36:06 <I440r> the executable STILl wont run tho
16:36:45 <I440r> the only files in asmsrc/*.1 that make ANY references to a syscall are...
16:36:45 <I440r> linux.1  <--- linux specific
16:36:55 <I440r> freebsd.1  <-- freebsd specific
16:37:05 <I440r> no other file in there does ahy syscall stuff
16:37:17 <I440r> i would say work iwth freebsd.1 and change it to work in openbsd
16:37:17 <jamc> gaah now all hell broke loose :)  isforth.asm:25: error: segment name `.head' not recognised
16:37:30 <I440r> change the syscall numbers that are specified in there
16:37:41 <jamc> changed to  flags = $(debug) $(list) -faoutb  in Makefile
16:38:05 <I440r> err
16:38:17 <I440r> oh
16:38:17 <I440r> does a.out allow multiple sections ?
16:38:29 <I440r> or is it all in one section
16:39:08 <jamc> yeah, normally you have text, bss and data, as in traditional unix
16:39:22 <jamc> not sure if you can name your own sections though
16:39:53 <I440r> .head can be renamed to .data
16:40:25 <I440r> maybe
16:40:56 <I440r> yes.
16:41:06 <I440r> edit macros.1 and change .head to .data
16:41:06 <I440r> that will probably work
16:41:42 <I440r> works for me
16:43:08 <jamc> strip -R doesn't work here, so i'll comment out the stripping
16:43:29 <I440r> hrm
16:43:30 <I440r> ok
16:43:32 <jamc> dne@nowhere$ file kernel.com
16:43:32 <jamc> kernel.com: OpenBSD/i386 demand paged executable not stripped
16:43:41 <I440r> cool
16:44:00 <I440r> did you do a make linux or a make freebsd
16:44:05 <jamc> freebsd
16:44:21 <I440r> ok
16:44:26 <I440r> thats a good starting poing
16:44:32 <I440r> let me explain the differences between make linux and make freebsd
16:44:37 <I440r> make linux does
16:44:45 <I440r> cp linux.1 syscalls.1
16:44:52 <I440r> make freebsd does
16:44:59 <I440r> cp freebsd.1 syscalls.1
16:44:59 <I440r> so. dont edit syscalls.1 heh
16:45:16 <jamc> agreed :)
16:45:23 <I440r> look in freebsd.1. make sure all the syscalls specified in there have the correct syscall number
16:47:12 <I440r> let me explain something tho
16:47:39 <I440r> there is a function in there caled init_mem that makes calls to 2 different syscalls
16:47:53 <I440r> sys brk and sys mprotext
16:48:06 <I440r> those are easy enough to understand
16:48:06 <I440r> but look towards the bottom of the source file
16:48:13 <I440r> there are alot of forth syscalls words in there
16:48:22 <I440r> they all follow the format...
16:48:22 <I440r>  call do_syscall
16:48:29 <I440r>  db syscall_number
16:48:33 <I440r>  db number_of_parameters
16:48:40 <I440r> you need to make sure THOSE are correct too
16:48:53 <I440r> once you correc those you can test it by doing
16:48:59 <I440r> make freebsd
16:49:05 <I440r> ./kernel.com
16:49:13 <I440r> if you dont gpf etc you are probably in a working state :)]
16:49:35 <I440r> fload isforth.f will smoke test it some more for you
16:49:42 <I440r> if you see alot of dots being printed very fast then you have a working compiler
16:49:54 <I440r> each dot represents ONE file being compiled
16:50:09 <jamc> ok, i'll probably need to research the kernel source to figure this out
16:50:18 <I440r> that should drop you back to the shell and you should then have a ./isforth in there. but dont run it because it will be b0prken till you fix it :)
16:52:05 <I440r> man 2 works in linux
16:52:15 <I440r> erm and unistd.h tells me the syscall nubmers for the various syscalls
16:52:43 <jamc> yup
16:53:07 <jamc> but are just the numbers enough? i mean the calling sequence may differ too
16:53:23 <I440r> the fact that you got it to build is a big step hehe
16:53:35 <I440r> NOW get it to RUN@!!!
16:53:35 <I440r> lol;
16:53:42 <I440r> the order parameters are passed might be different.
16:53:45 <I440r> the nubmer of parameters might be different
16:54:04 <I440r> the syscall might not even EXIST in bsd
16:54:12 <jamc> yeah that's what i was thinking
16:54:14 <I440r> that would be a problem :P
16:54:19 <I440r> but.
16:54:29 <I440r> there would be a way to effect the same result from bsd
16:54:47 <jamc> of course they exist
16:55:28 <I440r> if you do this i can guarantee youll be an expert syscall person when your done :)
16:56:05 <jamc> lucky me, hehe
17:03:00 --- quit: davidw (Remote closed the connection)
17:03:00 --- join: davidw_ (~davidw@adsl-66-101.38-151.net24.it) joined #forth
17:05:40 <jamc> I440r: hmm why is there a sys_creat? open supports O_CREAT in linux i hope?
17:26:11 --- quit: davidw_ (Read error: 113 (No route to host))
17:32:42 <I440r> jamc
17:32:49 <jamc> yup
17:32:57 <I440r> sorry had to go afk to sort out my sister :)
17:33:09 <jamc> hehe np
17:33:23 <I440r> ]theres a sys open and close, open supports create
17:33:33 <I440r> but theres a sys creat too i belive
17:33:43 <I440r> yes syscall number 8 is creat
17:34:01 <jamc> yeah there is... for backwards compatibility... it's considered obsolete nowadays
17:34:19 <I440r> but open DOES support create of file... thers actuall a bug in isforth relating to this
17:34:39 <I440r> if i do an fsave in isforth it tries to CREAT its own executable
17:34:49 <jamc> right, that's "open" is standard way to create files in modern unix
17:34:58 <I440r> but that tends to crap out if that same executable already exists and is currently OPEN
17:34:58 <I440r> lol
17:35:08 <jamc> bugger :)
17:35:20 <I440r> yea. isforth currently uses creat
17:35:33 <I440r> i was unaware it had been deprecated
17:35:46 <I440r> isforth very rarely needs to resave an executable of the same name
17:36:02 <I440r> but if your running ./isforth and also run ./kernel and try extend
17:36:10 <I440r> the save is going to barf :)
17:36:19 <I440r> that will be fixed when we have an install
17:36:28 <I440r> because the isforth you wil be runnning will be in /usr/bin or something
17:36:28 <I440r> not ./
17:36:37 <jamc> right :)
17:37:40 <jamc> i'm trying to sort out init_mem...
17:39:14 <I440r> yes
17:39:22 <I440r> that would be a good place to start
17:39:30 <I440r> thats the first thing isforth does :)
17:39:39 <I440r> it sys breaks out to one meg
17:39:43 <jamc> here's how the sbrk is implemented in openbsd:  "return (ENOSYS);", hehe
17:39:50 <I440r> then sys mprotects it all to read/write/execute :)
17:40:14 <I440r> chmod +rwx *.memory
17:40:14 <I440r> lo;l
17:40:14 <I440r> its called brk i think
17:40:37 <I440r> there has to be a break of some sort. or else openbsd is crap :P
17:41:24 <jamc> yeah there's plain "break",  sbrk/brk are implemented in libc as wrappers around break
17:42:06 <I440r> aha. well. if you go the libc route im going to SHOOT you :P
17:42:18 * I440r goes to find his 1911 :)
17:42:39 <jamc> hehe just looking there to see how it's done :)
17:43:07 <I440r> yes
17:43:11 <jamc> it's in asm, so if you're statically linking there shouldn't be a problem with bloat
17:43:20 <I440r> i do that... i write a small piece of c code, statically link it to libc
17:43:30 <I440r> then use the interactive disassembler pro to reverse engineer it :)
17:43:46 <I440r> jamc i will NOT use libc in isforth
17:43:46 <jamc> well this is asm already
17:43:54 <I440r> or ANY OTHER library
17:44:00 <jamc> but you use errno?
17:44:14 <I440r> all i/o will go directly to the kernel
17:44:25 <I440r> no fucked up libraries to hold your hand for you
17:44:39 <I440r> no
17:44:39 <I440r> i generate errno
17:44:51 <I440r> you think libc pulls errno out of thin air ?
17:44:51 <I440r> no
17:45:11 <jamc> hehe :) yeah you're right... momentary confusion
17:45:16 <I440r> it generates it based on the returned value of the syscall
17:45:16 <I440r> thats what i do too
17:45:16 <I440r> libc does NOT give me errno. the syscall itself does
17:45:28 <jamc> yup i remember
17:46:13 <I440r> its a bit obfuscated and totally undoc'd
17:47:07 <jamc> well it's how it's been done for ages... described in various books on unix architecture
17:47:45 <I440r> yes but not in MAN pages :P
17:47:59 <I440r> and thats all i got :P
17:48:40 <jamc> well that can be hell, especially on linux (in my experience)
17:49:32 <I440r> linus doesnt want any one to know how to interface directly to the kernel
17:49:46 <I440r> he wants us all to be crippled into using shittly middlemen like libc
17:49:53 <I440r> so he doesnt doccument the calls
17:50:04 <I440r> i say linux is closed source
17:50:33 <jamc> hmmm well
17:51:13 <jamc> source is source, docs are docs
17:51:54 <I440r> and in linux... never the twain shall meet
17:52:01 <I440r> i.e. the sources are totally undoc'd
17:52:11 <jamc> you get what you pay for :)
17:52:17 <I440r> i.e. closed to all but the super 31337
17:52:33 <I440r> whats more the sources are a NIGHTMARE to read
17:52:43 <I440r> oh... its free so the sources have to be a pile of shit ?
17:52:48 <I440r> thats what im doing wrong with isforth ???
17:52:50 <I440r> lol
17:52:59 <jamc> if you try any of the bsd, you'll see that the manpages are of much higher quality and more complete
17:55:27 <I440r> im downloading freebsd as we speak
17:55:34 <I440r> im 9% of the4 way done
17:55:41 <jamc> ouch :)
17:55:44 <I440r> and only 66 more hours to go
17:55:54 <I440r> then i got disk 2
17:59:34 <jamc> you don't have to use an ISO... you can just put the tarballs on a cd, make an install floppy(s)
18:00:31 <I440r> jamcso then the install has to go thru the PPP ?
18:00:31 <I440r> yea... good idea :P
18:00:54 <jamc> no it doesn't have to
18:01:25 <jamc> you just transfer the necessary tarballs... it'll be much smaller that ISOs
18:02:41 <I440r> heh well what i realy want is a bootable freebsd / partion
18:02:49 <I440r> and a mounted small /home :)
18:03:12 <jamc> hmmm, there's     4.5-mini.iso     - Minimal bootable 4.5 installation CDROM image.
18:03:15 <I440r> i wouldnt know what tarballs i NEED
18:03:23 <jamc> 177M only
18:04:57 <I440r> im grabbing 5.0 heh
18:04:57 <I440r> unofficial cd
18:05:28 <jamc> ah, well 5.0 hasn't been released yet
18:05:42 <I440r> hehe
18:05:48 <I440r> i know
18:05:48 <I440r> im fast :)
18:05:53 <I440r> lol
18:06:32 <jamc> well don't expect stability then
18:07:25 <I440r> nah
18:07:31 <I440r> i also ran debian SID for a while
18:07:35 <I440r> but reverted back to woody
18:10:42 <jamc> how do i import an external symbol in nasm?
18:11:59 <I440r> why would you want to
18:12:08 <I440r> please tell me your not trying to use libc?
18:12:18 <I440r> that wont work
18:12:20 <I440r> at all
18:12:40 <I440r> please do not taint isforth by marrying it to that pile of horse shit
18:12:41 --- join: bugslayer (thin@h24-64-175-61.cg.shawcable.net) joined #forth
18:12:49 <I440r> futhin!
18:12:53 <jamc> nope, i need a symbol that points to the end of data segment
18:13:05 <I440r> erm
18:13:08 <jamc> i assume it's simply defined in the linker script
18:13:13 <I440r> err
18:13:16 <I440r> no
18:13:28 <jamc> yes
18:13:31 <I440r> i mean
18:13:34 <I440r> why do yo uneed taht ?
18:13:41 <jamc> to emulate sbrk
18:13:44 <I440r> u lost me
18:13:50 <bugslayer> heya i440r
18:14:09 <jamc> here's sbrk in openbsd:
18:14:14 <I440r> data being head space... look at the macros to create headers, you could figure out "end of data" by knowing how headers are created etc
18:14:18 <jamc> 	movl	4(%esp),%ecx
18:14:18 <jamc> 	movl	curbrk,%eax
18:14:18 <jamc> 	addl	%eax,4(%esp)
18:14:18 <jamc> 	movl	$(SYS_break),%eax
18:14:18 <jamc> 	int	$0x80
18:14:19 <jamc> 	jc	err
18:14:21 <jamc> 	movl	curbrk,%eax
18:14:23 <jamc> 	addl	%ecx,curbrk
18:14:25 <jamc> 	ret
18:14:44 <I440r> oh you need to know where you are breaking FROM ?
18:14:48 <jamc> curbrk is intialized with _end, which point to the end of the data segment
18:15:19 <I440r> OOO you pass parameter by register in openbsd ?
18:15:19 <I440r> then you need to be working on the linux.1 file as a base
18:15:31 <I440r> freebsd passes parameters by stack
18:16:07 <I440r> ok. so you need to knw where your end of memory is currently?
18:16:07 <I440r> that can be arranged
18:16:07 <I440r> hang on
18:16:23 <jamc> you're talking syscalls... now this sbrk code i sent is an asm function in libc
18:16:55 --- join: TheBlueWizard (~tbw@ip-216-25-202-118.vienna.va.fcc.net) joined #forth
18:16:55 --- mode: ChanServ set +o TheBlueWizard
18:17:00 <TheBlueWizard> hiya all
18:17:13 <bugslayer> howdy tbw
18:17:23 <I440r> tbw! :)
18:17:26 <TheBlueWizard> hiya bugslayer
18:17:32 <TheBlueWizard> hiya I440r!
18:17:35 <bugslayer> bluewi = bluewy
18:17:48 <bugslayer> i declare your nickname is bluewy now :P
18:17:52 <I440r> jamc yes but its passing the parameters to int 0x80 (the syscall) via registers, NOT on the stack!!!!
18:17:53 <TheBlueWizard> hmm?
18:18:10 <bugslayer> funner to type "bluewy" instead of "tbw" :P
18:18:19 <jamc> I440r: uhm isn't that just the syscall number that's passed in eax
18:18:35 <I440r> in freebsd it would be push parameter push parameter
18:18:45 <TheBlueWizard> bugslayer: if that makes you happy...fine...I'll stick with mine, thank you
18:19:09 <I440r> jamc no!!! freebsd PUSHES all the parameters. linux puts them ALL in registers. im guessing openbsd does too
18:19:09 <I440r> looking at that syscall code
18:19:20 <I440r> so you need to be using linux.1 as a base
18:19:29 <I440r> not freebsd
18:19:33 <bugslayer> thebluewizard: no i'm just saying instead of me having to type "tbw" i can just type "bluewy"
18:19:50 <I440r> and i need to see how best to determine the end of the .data section
18:20:00 <I440r> tbw is easier
18:20:06 <jamc> as i said... x86 asm makes me dizzy, can't even remember which way mov moves :)
18:20:11 <I440r> :P
18:20:48 <I440r> mov dest, src :)
18:21:08 <TheBlueWizard> jamc: if it helps, remember that the constants are at the end (e.g. mov eax,3)
18:21:57 <jamc> right... but in AT&T syntax (gas) it's the other way around
18:22:12 <jamc> or?
18:22:20 <I440r> gas is fucked in the head
18:22:22 <bugslayer> so what do you guys think about forth? eh!
18:22:26 <TheBlueWizard> hehe
18:22:30 <I440r> whoever designed gas needs to be shot
18:22:30 <I440r> twice!
18:22:52 <bugslayer> i'll shoot him a second time..
18:22:53 <bugslayer> er
18:22:55 <bugslayer> third time
18:23:18 * TheBlueWizard grins
18:23:19 <I440r> futhin we are talking forth
18:23:25 <I440r> were talking fixing isforth for openbsd
18:23:25 <I440r> free bsd
18:23:29 <I440r> and net bsd :P
18:23:30 <jamc> calm now.... the ones who needs to be shot are the x86 designers
18:23:48 * TheBlueWizard looks around for futhin....
18:23:51 <I440r> them too :)
18:23:53 <bugslayer> shh
18:26:52 <bugslayer> dumb debuggers are the worst eh ?
18:27:17 <bugslayer> reading maxim makes me dumb
18:31:38 <I440r> jamc it shouldnt be too difficult to find the end of the data segment, the problem is that the only place the data segment is references AT ALL in the sources is from within macros
18:31:44 <I440r> and they are all in macros.1
18:32:19 <jamc> well there's a linker symbol, i just need to import it
18:34:52 <I440r> hrm
18:35:04 <I440r> you could do
18:35:04 <I440r> .data
18:35:14 <I440r>  foo = $
18:35:14 <I440r> .text
18:35:24 <I440r> at the end of isforth.f
18:35:35 <I440r> erm no
18:35:35 <I440r> do
18:35:35 <I440r> .data
18:35:35 <I440r> foo:
18:35:36 <I440r>  .text
18:35:50 <jamc> i just used it as a variable... seemed to work
18:36:02 <I440r> and somewhere in the soruces you can do mov eax,foo
18:37:01 <I440r> mov eax,[foo] gets contents of address
18:37:01 <I440r> mov eax, foo
18:37:01 <I440r> points eax at address
18:41:17 <bugslayer> anybody have any forth source code to give me?  i am interested in small interesting forth programs that do something amusing, like display something or do something useful..   i'm not particularly interested in mathematical programs or extensions of forth , etc
18:49:04 <TheBlueWizard> gotta go...bye all
18:49:34 --- part: TheBlueWizard left #forth
18:56:10 * I440r is gona go code... ill bbl
18:56:25 <I440r> jamc if u get ./kernel.com working ill buy you a beer nest time i see you :)
18:56:45 <jamc> hehe... i'm busy dumping core :)
18:57:17 <I440r> lol
18:57:21 --- quit: I440r ("the coder whent that way -->")
19:21:34 <bugslayer> anybody here alive?
19:22:01 * jamc nearly asleep
19:22:24 <bugslayer> i'm curious about how i can switch the screen mode from 80x25 to 80x50 in forth
19:22:44 <bugslayer> in dos
19:22:53 <bugslayer> if i were in linux, i'd just use the framebuffer..
19:23:35 <jamc> i don't remember how exactly... but it's like 3 lines of asm, IIRC
19:24:04 <bugslayer> what keywords should i use to google for it?
19:24:29 --- join: tathi (~josh@ip68-9-58-81.ri.ri.cox.net) joined #forth
19:25:14 <jamc> 80x50 dos text-mode change  :)
19:25:43 <jamc> found it.... http://www.cybertrails.com/~fys/faq.htm#dos1
19:25:59 <jamc> damn, 4 lines :)
19:26:14 <bugslayer> heh
19:30:34 <bugslayer> i can type mode co80,50 at the dos prompt too
19:30:46 <jamc> right
19:33:43 <jamc> gotta go to bed... see ya all
19:34:10 <bugslayer> good night
19:34:24 --- quit: jamc ("[x]chat")
19:46:47 --- join: CrowKiller (Vapo_Rulez@cnq5-233.cablevision.qc.ca) joined #forth
19:46:52 <CrowKiller> hi
19:47:01 <tathi> hey
19:48:13 <CrowKiller> does someone here developped something for an ARM processor? do you think we could code a unique virtual forth machine in thumb code and x86 assembly?
19:48:30 <bugslayer> what's thumb code
19:49:09 <tathi> don't see why not...never done much with ARM though...
19:49:54 <CrowKiller> thumb code is from what i've read 16 bits tokenized arm assembly
19:50:10 <CrowKiller> like a 16 bit instruction set for the most common 32bits instructions
19:52:01 <CrowKiller> because there's a portable and programmable 32 bits architecture present named the GBA
19:52:29 <CrowKiller> the gameboy advance is really a nice piece of portable, cheap hardware, with sound and color screen
19:52:46 <tathi> what processor?
19:52:48 <CrowKiller> the processor is 32 bits, as with the x86
19:52:52 <CrowKiller> arm7tdmi
19:53:01 <CrowKiller> with a graphical coprocessor
19:53:05 <tathi> nice
19:53:07 <tathi> speed?
19:53:16 <CrowKiller> hm around 25-30mhz
19:53:22 <CrowKiller> the specs are somewhere
19:53:29 <CrowKiller> but what is really nice
19:53:40 <tathi> yeah, just thought I'd ask you instead of going and looking it up :)
19:53:49 <CrowKiller> is that there's a serial bootstrapper onboard
19:54:09 <tathi> hmm... :)
19:54:26 <CrowKiller> that read an input stream if theres no cart present, dump it in wram and execute it
19:54:52 <CrowKiller> http://www.godsmaze.org/gba/
19:55:18 <CrowKiller> ive got some AVR chips but the source code for the micro is in PIC assembly
19:55:53 --- nick: bugslayer -> futhin
19:56:50 <tathi> ah
19:57:01 <CrowKiller> in my case I would have to port it to the AVR and by using Dataflash serial memories from atmel also, its a much nicer dev platform, limited code memory but with an aha like system it would be quite easy to implement serial compilation on the fly
19:57:15 <CrowKiller> so dont need a 200$ programmable cart ;o)
19:57:22 <CrowKiller> the ram is 256kb large I think
19:57:52 <tathi> wait, source code for the cable's micro is PIC, right?
19:58:02 <CrowKiller> with that you cna easly dump roms, just upload a little programm that output it on the serial port and its done
19:58:04 <CrowKiller> yes
19:58:17 <tathi> oh, ok, I see what you're talking about now
19:59:17 <CrowKiller> but in theory it can be bypassed, because i know some code for windows 95 that can run Ring0 code on the fly, right in the middle of the program
19:59:44 <CrowKiller> so the horrible timing issues associated with windows and the hassels of implementing a driver are eliminated
20:00:08 <CrowKiller> I saw on the colorforth list that some guy wants to code a WinColorForth
20:00:25 <CrowKiller> it might get very interesting in the future
20:00:26 <tathi> yeah, I was reading that the other day
20:01:22 <tathi> I'm more interested in forth being the operating system than running on top of it though...
20:01:58 <CrowKiller> because the editor could run in Ring3, and when we start the colorforth program, a snapshot of memory is taken, the program run in ring zero and on completion or on error, the routine of putting the memory snapshot back there and returning the os to windows is called
20:02:14 <CrowKiller> incognito lol
20:02:56 <CrowKiller> this technique is used by virii makers, its detailed in the VIH source code, this is a great piece of work to learn about viruses and asm programmin
20:02:58 <CrowKiller> g
20:03:03 <tathi> heh
20:03:04 <tathi> yeah
20:03:24 --- mode: ChanServ set +o futhin
20:03:39 --- topic: set to 'can someone tell futhin if a forth chip can emulate the architecture required to run palmos, and run palmos at a faster speed than the current palms out there? :)' by futhin
20:03:58 <CrowKiller> just recode palmos functionalities in forth
20:04:07 <futhin> well that's more work
20:04:08 <CrowKiller> there's an emulator out there I think
20:04:22 <futhin> i just want to see if i can quickly get forth chips produced and out there in the world :P
20:04:31 <tathi> futhin, that would kind of depend on the speed of the forth chip...
20:04:49 <CrowKiller> forth chips like you want thme to be used exist
20:04:55 <CrowKiller> theyre called java stamps
20:04:59 <CrowKiller> or named like that
20:05:42 <CrowKiller> palmos is irrevelant IMHO
20:05:59 <CrowKiller> irrelevant
20:06:16 <CrowKiller> just look at the Gameboy Advance
20:06:22 <CrowKiller> cost lot less than a pda
20:06:51 <CrowKiller> got mroe capabilities built in than some of them, and you can extend it by building custom hardware into a cart
20:08:12 <futhin> well ideally, we want forth chips to be used lots in the real world, so that they are available to us forthers :)
20:08:26 <CrowKiller> design them in FPGAs
20:08:33 <CrowKiller> but they will not run fast
20:08:41 <CrowKiller> or as fast as you could get from a real chip
20:08:52 <CrowKiller> realdigital CPLDs from xilinx look great
20:08:56 <tathi> I'd love to do that
20:09:10 <tathi> unfortunately I don't know _anything_ about hardware at that level...
20:09:11 <CrowKiller> they're all digital CPLDs run on less power and to 300 MHZ
20:09:18 <CrowKiller> i learned the basics in 2 weeks
20:09:20 <tathi> yeah, I was looking at them the other day
20:09:22 <CrowKiller> really easy
20:09:43 <tathi> that's good to know
20:10:12 <CrowKiller> but if we only have foreign hardware, we must code efficient virtual machines
20:10:31 <CrowKiller> a 32 bit one and a 8 bit one
20:10:36 <tathi> yup
20:10:36 <CrowKiller> chuck said that before
20:10:49 <CrowKiller> and i tried myslef to come up with a 8 bit standard
20:10:54 <CrowKiller> for a 32 bit machine
20:10:58 <CrowKiller> for compactness
20:11:05 <CrowKiller> but what i gain there I loose it in speed
20:11:17 <tathi> really
20:11:20 <futhin> what would be a "killer application" for forth chips?  PDAs?
20:11:24 <futhin> embbedded stuff?
20:11:26 <CrowKiller> for
20:11:27 <CrowKiller> me
20:11:28 <CrowKiller> :
20:11:31 <CrowKiller> gameing consoles
20:11:45 <CrowKiller> XBOX, GBA, gamecube, PS2
20:12:17 <tathi> you mean you lose speed vs. a 32-bit standard for a 32-bit machine?
20:12:21 <CrowKiller> XBOX is a x86 and is 32 bits, GBA is a 32bits ARM and the gamecube is a 32bits PowerPc
20:12:38 <tathi> I'll write stuff for gamecube then :)
20:12:45 <futhin> what are you talking about?
20:12:47 <CrowKiller> i own myself a gamecube
20:12:52 <futhin> how would you use forthchips for that sutff?
20:12:58 <futhin> and who really cares about gaming consoles? ;)
20:13:02 <CrowKiller> im not talking about forth chips
20:13:08 <CrowKiller> talking about virtual machines
20:13:11 <CrowKiller> if we developp
20:13:15 <CrowKiller> a virtual machine so good
20:13:20 <CrowKiller> that it performs well
20:13:24 <CrowKiller> on those three consoles
20:13:37 <CrowKiller> then everybody will want to build forth chips
20:14:00 <futhin> so develop emulators ontop of forthchips..  one that emulates x86, one for ARM, one for powerpc?
20:14:03 <futhin> is that what you are saying?
20:14:06 <CrowKiller> because they'll get it in the teeth: forth is the best language around and floppy hardware blocks it
20:14:15 <CrowKiller> futhin
20:14:16 <CrowKiller> lol
20:14:21 <CrowKiller> im saying
20:14:24 <CrowKiller> we must emulate
20:14:28 <CrowKiller> the ideal forth chip
20:14:35 <CrowKiller> on top of those non-perfect hardware
20:14:45 <CrowKiller> so everybody, in their search for speed
20:14:53 <CrowKiller> will look into hardware forth chip design
20:14:58 <tathi> i.e. develop a virtual machine with standard bytecode
20:15:04 <tathi> write code for it
20:15:05 <CrowKiller> java tried this
20:15:14 <CrowKiller> but its wrong and ugly
20:15:19 <CrowKiller> forth is the key to me
20:15:29 <CrowKiller> aha particularly
20:15:41 <CrowKiller> JIT compilation of java
20:15:53 <futhin> okay, so basically say to everyone "forget about java, use forth" and then create high-quality forth virtual machines for all the architectures
20:15:55 <CrowKiller> that what it brings to forth
20:16:28 <CrowKiller> we must code applications in forth, on top of our ideal virtual machine implemented on those 32 bits hardware
20:16:36 <CrowKiller> java is just an example im saying your mixing things
20:16:38 <futhin> why will people look into hardware forth chip design?
20:16:48 <futhin> yes! yes!
20:16:51 <futhin> this is good stuff
20:17:01 <futhin> with a virtual machine, we can have a portable forth operating system..
20:17:10 <futhin> right? :P
20:17:28 <CrowKiller> hey come on! forth will be running on top of INNEFICIENT VIRTUAL INSTRUCTIONS
20:17:41 <CrowKiller> if that code would run on dedicated hardware it woudl run fast
20:17:49 <futhin> yes
20:17:55 <CrowKiller> anyway all the picture is very clear to me now
20:17:59 <futhin> um
20:18:12 <futhin> code a forth os ontop of the forth virtual machines..
20:18:18 <futhin> so that people gain acceptance of the os
20:18:19 <CrowKiller> get lost with the os
20:18:21 <CrowKiller> lol
20:18:26 <futhin> then later on, they switch to forth chips
20:18:27 <CrowKiller> no os at all
20:18:30 * tathi agrees with CrowKiller :)
20:18:31 <futhin> and run the os ontop of the forth chips
20:19:02 <CrowKiller> only a virtual layer, to uniform all forth code, so that a non-platform specific code can run on any of those consoles
20:19:05 <CrowKiller> that include games
20:19:06 <CrowKiller> and all
20:19:30 <CrowKiller> overall acceptance of this programming model would become true
20:19:36 <CrowKiller> because that the ebst way to go
20:19:56 <tathi> so have you started designing a virtual machine yet?
20:20:08 <CrowKiller> we just have to get our ass moving and come up with the best 32 bits virtual machine possible
20:20:11 <CrowKiller> yes
20:20:18 <futhin> can't we just use machineforth?
20:20:18 <CrowKiller> ive got idea for the if
20:20:25 <CrowKiller> machineforth need improvement
20:20:29 <CrowKiller> in clarity
20:20:35 <futhin> how so?
20:20:40 <tathi> ok
20:20:43 <CrowKiller> i dont know all of machine forth
20:21:03 <CrowKiller> but i know I enjoyed very much in my asm programming session on AVRs and SXs
20:21:19 <CrowKiller> the "skip the next instruction of the flag is true"
20:21:25 <CrowKiller> or false or whatever
20:21:33 <CrowKiller> in asm this concept is odd
20:21:45 <CrowKiller> because you can only skip a single machine instruction
20:21:53 <CrowKiller> of a specific size
20:22:05 <CrowKiller> so you can only do this with small jumps values or things like that
20:22:12 <CrowKiller> imagien that concept in forth
20:22:24 <tathi> though that's only a problem on machines with variable instruction sizes...
20:22:43 <CrowKiller> if the "if" skip^the following word only
20:22:46 <CrowKiller> its great
20:22:57 <CrowKiller> cause no more search for the damn then 37 word later
20:23:04 <CrowKiller> no back propagation of the jump offset
20:23:15 <CrowKiller> no more complications
20:23:20 <CrowKiller> just a single word
20:23:31 <CrowKiller> and if you want to make a jump table
20:23:32 <CrowKiller> its easy
20:23:38 <CrowKiller> just put a the start of the called word
20:23:43 <tathi> you still get the same problem with forward branches though...
20:23:46 <CrowKiller> a Rdrop insstruction
20:24:04 <CrowKiller> that drop the first item on the return stack
20:24:24 <CrowKiller> tathi: ?
20:24:53 <tathi> the bit about searching for the then
20:25:14 <CrowKiller> my code would look like: :cacaboudin blopblopblop aliboulalawordfromnowhere if jumptothatotherwordiftrue elsegotothatotherword
20:25:30 <CrowKiller> it would add the flag to the pc
20:25:33 <tathi> d'oh
20:25:34 <tathi> sorry
20:25:39 <CrowKiller> the negative flag
20:25:42 <CrowKiller> or the zero flag
20:25:47 <CrowKiller> or any flag that you implement
20:25:50 <tathi> yeah
20:26:06 <CrowKiller> this is a lot nicer mechanism
20:26:13 <CrowKiller> and really easier to code than thens
20:26:40 <CrowKiller> a lot easier to read and all
20:26:44 <CrowKiller> its all in my .plan
20:27:01 <CrowKiller> now that the if are defined i need to tackle with other control structures
20:27:05 <CrowKiller> like loops
20:27:08 <CrowKiller> and all
20:27:16 <CrowKiller> an efficient way of implementing them
20:27:27 <CrowKiller> maybe a special "count register"
20:27:33 <CrowKiller> or by using the A register
20:27:56 <CrowKiller> the A would have store+ or fetch+
20:28:03 <tathi> A register would probably be better
20:28:18 <CrowKiller> a register dedicated to looping constructs
20:28:30 <CrowKiller> i thought it would be great also
20:28:42 <futhin> woah woah
20:28:54 <CrowKiller> and also made the data stack idem to the return stack
20:29:02 <tathi> wait, I meant I'd prefer using the A register to having a dedicated count register
20:29:03 <CrowKiller> so we can do drops overs and stuff on the return stack
20:29:25 <futhin> how does machineforth need improvement in clarity?
20:29:31 <futhin> let's get back to my question :P
20:29:43 <futhin> i don't know much asm
20:29:47 <CrowKiller> in clarity i dont know but at least the code ive read is at someplace hard to understand
20:30:14 <CrowKiller> looping could be easly implemenetd with "dec if Rover" construct
20:30:20 <tathi> dedicated count register just seems like one more thing that you don't _really_ need
20:30:28 <tathi> one more bit of state that you might need to save
20:30:31 <CrowKiller> hahahaha the if is really powerful that way
20:30:49 <futhin> crowkiller: why not talk to chuck moore?  maybe ask him for some help in building your virtual machine ;)
20:31:07 <CrowKiller> I would like very much if he could come here
20:31:23 <CrowKiller> but you people on this chan just bore him with stupid things IMHO
20:31:35 <futhin> like what?
20:31:41 <CrowKiller> i dont know
20:31:46 <CrowKiller> hes advanced a lot
20:32:00 <CrowKiller> we shpould try to look into what he's interested
20:32:06 <futhin> i haven't had a chance to ask some deep questions, in fact, i don't really want to do that until i've read all the papers on ultratechnology and at other places
20:32:08 <CrowKiller> and try to discuss with him
20:32:15 <CrowKiller> much like jeff did all those years
20:32:25 <CrowKiller> ive read all those papers
20:32:42 <CrowKiller> and ive read all the stuff I could read
20:32:56 <CrowKiller> on the net and elsewhere
20:33:30 <CrowKiller> and really you dont need that do do something useful or intelligent or to help gurus like thefox or chipchuck
20:34:15 <CrowKiller> just try to know what theyre talking about and see things like they see them and try to come up with something new fresh and new that would be interesting
20:34:21 <futhin> i hope whatever virtual machine you design..  i hope that you keep it FORTH!
20:34:23 <CrowKiller> hackiong away gaming consoles is great
20:34:33 <futhin> don't bastardize forth :P
20:34:33 <CrowKiller> it will be forth
20:34:42 <futhin> ans forth isn't forth in chuck's view
20:34:43 <CrowKiller> im the only one here that dont bastardize it
20:34:48 <CrowKiller> i know all that stuff
20:35:02 <futhin> i hope i don't bastardize it either
20:35:09 <CrowKiller> and i find it a pity personnaly that isforth exist, that someone put his time in it
20:35:11 <futhin> i really agree with chuck's philosophy
20:35:21 <futhin> it's a hard path to walk, but it's the higher path
20:35:29 <futhin> what don't you like about isforth?
20:35:33 <CrowKiller> chuck's path is the easiest
20:35:46 <CrowKiller> the simple reason, it looks complicated
20:35:58 <CrowKiller> and its not like i want foth to be
20:36:04 <CrowKiller> its not the incarantion im looking for
20:36:13 <futhin> no, it is complicated, people won't understand you, and it's hard to achieve the simplest solutions
20:36:16 <futhin> it takes work
20:36:28 <futhin> "it takes a genius to be simple" -- one of my favorite sayings
20:36:35 <CrowKiller> i dont want to code into that, i dont want to learn how its working on the inside
20:36:40 <CrowKiller> on to cope with its bug
20:36:44 <futhin> oh you are talking about isforth
20:36:52 <CrowKiller> learn how to come up with something
20:37:00 <futhin> what bug?
20:37:14 <CrowKiller> i dont know, i never used it
20:37:15 <tathi> whatever bugs it may have...
20:37:19 <CrowKiller> its a shame in fact
20:37:25 <CrowKiller> but im not interested
20:37:34 <CrowKiller> i always got better things to do than try it
20:37:38 <futhin> well, that's fine and all
20:37:42 <CrowKiller> like linux
20:37:47 <CrowKiller> i once believed in linux
20:37:50 <CrowKiller> i installed it
20:37:54 <CrowKiller> 2 years ago
20:38:05 <CrowKiller> when slackware 4.0 has hit the shelves
20:38:21 <CrowKiller> i build my own system from scratch it took me my full summer
20:38:36 <CrowKiller> this was a ream system for me, well almost, i learned a lot about linux and all i sacrified time
20:38:58 <CrowKiller> and when my hd got formated because of a stupid partition error and the a format command
20:39:05 <tathi> ouch
20:39:41 <CrowKiller> i was so discouraged I got windwos installed, and im using it, i was tired of ripping my ass in 4 just
20:39:48 <CrowKiller> to come up with something IN LINUX
20:39:54 <futhin> yeah
20:39:59 <futhin> i sympathize
20:40:01 <CrowKiller> because linux is bla bla bla bla bla... you know the stuff
20:40:03 <CrowKiller> lol
20:40:07 <futhin> linux can take a lot of time
20:40:18 <CrowKiller> i tried to install it on my friend,s computers
20:40:24 <futhin> a learning curve, time to customize things & tools to feel comfortable with it, etc
20:40:25 <CrowKiller> i wasnt even able
20:40:31 <CrowKiller> in the time i had
20:40:41 <CrowKiller> because kernel would not compile on their particular machine
20:40:44 <CrowKiller> and bla bla b;a
20:40:54 <CrowKiller> linux is really enslaving people as much or even more as windows
20:41:03 <CrowKiller> forth is the key
20:41:16 <futhin> yes true
20:41:23 <CrowKiller> all i want is simplicity damn it lol
20:41:24 <futhin> i don't like linux, and i don't like windows
20:41:32 <futhin> and i don't like just about every os out there :P
20:42:02 <futhin> my main interest is coding an os
20:42:05 <tathi> yes, and you're goung to replace them all with a forth os that's going to take the world by storm, we know :P
20:42:09 <futhin> a better os..
20:42:12 <futhin> in forth
20:42:13 <CrowKiller> its so much gratifyng coding a AVR or a SX to do your owm biddings, to code a complete solution to a problem, its great
20:42:26 <CrowKiller> we dont need an OS
20:42:31 <futhin> AVR or SX?
20:42:32 <CrowKiller> OS = shit
20:42:41 <CrowKiller> www.atmel.com for the avr
20:42:43 <futhin> crowkiller:  WHO IS "we" ???
20:42:50 <CrowKiller> www.ubicom.com for the SX chip
20:42:58 <CrowKiller> everyone computing
20:43:01 <CrowKiller> all of us
20:43:05 <CrowKiller> even the begginers
20:43:31 <futhin> so what do you propose?
20:43:41 <CrowKiller> they dont need an os, in fact they need to learn how to use a simple, small, powerful forth in 5-10 minutes
20:43:47 <futhin> how are we web browsing and talking on irc without an os?
20:43:48 <CrowKiller> and stop wasting their lives
20:43:52 <CrowKiller> in front of computers
20:44:04 <CrowKiller> because the time you're apssing in front of the screen
20:44:28 <CrowKiller> doing nothing that you want to do, you do nothing useful
20:44:46 <CrowKiller> lol its that simple
20:44:56 <CrowKiller> wasting time in a chan
20:44:57 <CrowKiller> like this
20:45:00 <futhin> my idea for the forth os is 99% about how the general person will be able to easily code in forth after learning it in 10 minutes..  my inspiration comes from an easy scripting language called MUF (multi-user forth)
20:45:03 <CrowKiller> bragging into thin air blablas
20:45:17 <tathi> heh
20:45:18 <futhin> whatever, irc can be a tool
20:45:21 <futhin> web browsing is a tool
20:45:22 <futhin> etc
20:45:27 <futhin> people want tools
20:45:28 <CrowKiller> is really annoying for people like jeff fox and chuck
20:45:38 <futhin> i also want to give them the ability to modify their tools on the fly
20:45:42 <CrowKiller> they want some innovation
20:45:43 <futhin> or to code their own
20:45:44 <futhin> etc
20:45:47 <CrowKiller> they're actually used to it
20:45:49 <CrowKiller> lol
20:47:00 <futhin> lets get back to web browsers, spread sheets, communication tools, etc
20:47:11 <futhin> where are those things going to be when we eliminate the os?
20:47:17 <CrowKiller> in source code
20:47:20 <CrowKiller> all source
20:47:30 <futhin> yes
20:47:30 <CrowKiller> no more distributed object code
20:47:33 <futhin> that's fine
20:47:36 <futhin> that's still an os
20:47:39 <CrowKiller> only source that compiles itself on the fly
20:47:39 <futhin> that's what i was thinking..
20:47:41 <futhin> opensource
20:47:47 <futhin> source for everything..
20:47:51 <CrowKiller> opensource need living source code
20:47:56 <futhin> people can quickly, easily get into the source
20:47:59 <futhin> modify it, etc
20:48:01 <CrowKiller> that can evolve rapidly and efficiently
20:48:07 <futhin> (but more likely, extend it)
20:48:13 <CrowKiller> because having to compile something is frustrating
20:48:17 <CrowKiller> you ghet errors in the source
20:48:19 <CrowKiller> etc etc etc
20:48:27 <CrowKiller> if the app is stored as source no more need to compile
20:48:39 <futhin> yes
20:48:44 <CrowKiller> just execute the compiler when you want to use it and hop, its not traditionnal compilation
20:48:47 <futhin> it's still an os in my point of view
20:49:06 <CrowKiller> yeah to me an os is code that supervise other code
20:49:17 <CrowKiller> so thats why i dont feel a need to that
20:49:29 <futhin> my vision was that people would be using something like irc, and maybe they want to change how it behaves..  so they just code a few lines on the fly, and boom, they have extended the irc client
20:49:34 <futhin> and they can throw away that code later
20:49:43 <futhin> because the language is so simple & easy to understand
20:50:02 <CrowKiller> imagine kids coding in forth
20:50:06 <futhin> or they want to code their own tools, etc
20:50:06 <CrowKiller> on top of their gaming consoles
20:50:12 <CrowKiller> coding brilliant things
20:50:13 <futhin> imagine grandmas coding in forth! :P
20:50:21 <CrowKiller> forth is the lego of programming language
20:50:23 <CrowKiller> ;p
20:50:24 <futhin> yes
20:50:46 <futhin> yes, i was thinking along those lines with the "forthos"
20:50:54 <CrowKiller> that a dream to me, i would be in a perfect world if all my friend would be able to code
20:51:06 <CrowKiller> we could share things easly
20:51:18 <CrowKiller> they could learn faster, open their minds to a lot of things
20:51:22 <futhin> man, my first exposure to forth was MUF
20:51:34 <CrowKiller> my first was colorforth
20:51:35 <futhin> i learned it in like 5 minutes, and i was coding immediately..  it was exhilarating
20:51:44 <futhin> simple string based stuff
20:51:52 <futhin> but it was very easy to learn
20:51:54 <CrowKiller> i got frustrated that i could not do that lol
20:51:58 <futhin> and use
20:52:17 <CrowKiller> anyway ill come up with something, i must before going to college
20:52:21 <CrowKiller> because i wont have time after
20:52:32 <futhin> the interactive nature of forth is a greater advantage that most people realize..  most people do not understand how much of an advantage it actually is
20:53:02 <futhin> sure you'll have time
20:53:08 <futhin> as long as you don't party :P
20:53:21 <CrowKiller> i wont, actually I hate bars and stuff like that
20:53:35 <futhin> you only need to study about an hour or two every day
20:53:59 <CrowKiller> paryys arenot great because to me most of the time its losing time, smoking pot drinking beer and all, funny stuff come out, but to a greater loss: lost of time and health lol
20:54:07 <futhin> it's not as hard, but it's a trick.. they lull you in the first two months, then the last month is the real shit
20:54:54 --- quit: tathi ("later all...")
20:54:56 <futhin> imho, school is a joke
20:55:21 <futhin> highschool, university, whatever..  its a joke, and also a waste of time, inefficient teaching methods, etc
20:55:31 <CrowKiller> exactly
20:55:40 <futhin> you wanted to lose 30 pounds?
20:55:46 <CrowKiller> all knowledge must be accessible
20:55:56 <futhin> i read this very interesting book recently
20:56:22 <futhin> called The Carbohydrate Addict's Diet..
20:56:40 <CrowKiller> it is from an old women
20:56:55 <CrowKiller> my mother owns a library, she talked in her book about sugar addiction
20:57:07 <futhin> do you think maybe you are a carbohydrate addict? ever get any craving for sweets or breads?
20:57:11 <futhin> no this isn't really about sugar
20:57:13 <CrowKiller> its was a diet book the name was in french, i dont recall
20:57:17 <futhin> it's about carbohydrates
20:57:22 <futhin> the writers are american
20:57:35 <CrowKiller> it was a translation
20:57:47 <futhin> this book is not a translation..
20:57:48 <CrowKiller> the author's name was english
20:58:08 <futhin> there were two authors with this book..
20:58:11 <futhin> anyways
20:58:12 <CrowKiller> anyway i know you book isnt translation lol go on
20:58:36 <futhin> with carbohydrate addicts the body has a tendency to produce excess insulin
20:59:27 <futhin> generally, you eat, the body produces insulin, you finish the meal, the insulin is absorbed by the muscles and fat cells (and the cells use insulin to get fuel from the bloodstream)
20:59:28 <CrowKiller> so...?
20:59:59 <futhin> and then serotonin is produced when the insulin levels have gone down to pretty much zero
21:00:15 <futhin> and that gives you a satisfied feeling, and tells you that you are full
21:00:33 <futhin> however, with 80% of obese people, their body produces excess insulin
21:01:25 <futhin> and so they won't get the feeling of being satisfied..  2 hours after a meal, they might feel like snacking, even if it was a really big meal
21:01:40 <futhin> the bad thing is that with all the excess insulin, more fat is stored
21:01:50 <futhin> so you can actually eat less than a skinny person but still store more fat
21:02:11 <futhin> but generally, with all this insulin, you'll get lots of craving for carbohydrates
21:02:15 <futhin> erm
21:02:33 <futhin> when eating carbohydrates, the body produces insulin..
21:03:01 <CrowKiller> also a lot of obese people drink pops all day containg aspartame (synthetic sugar) that breaks up into ascorbic acid, phtenaltaine and methanol when temperature is more than 30deg celcius
21:03:16 <CrowKiller> those chemical too make insulin imbalances in the body
21:03:34 <futhin> so eating carbohydrates causes insulin to be produced and excess insulin will result in cravings for more carbohydrates, unsatisfied feelings, and greater fat uptake
21:03:48 <CrowKiller> in fact aspartame is an excitotoxin, killing your neurons into thinking that what you eat is sweet
21:03:56 <futhin> so here's how to manipulate the body to lose weight & reduce cravings
21:04:31 <futhin> the breakfast and lunch meals are low-carbohydrates, and then the dinner can be high-carb, eat as much as you want..
21:05:10 <CrowKiller> i usually dont eat much in the mroning or at noon, but i get a hundred snacks when at home lol
21:05:35 <futhin> when you eat, insulin is released within a few minutes
21:05:47 <futhin> the amount of insulin released is based on how much insulin needed previously
21:05:51 <CrowKiller> that may explain my proportionned weight even if i dont do exercise or sport
21:06:21 <futhin> so because your previous two meals were low-carbs, the insulin released during dinner is very low
21:06:39 <futhin> you are allowed to eat as much as you want during dinner, but here's the catch:
21:06:45 <futhin> you have to eat it within an hour
21:06:59 <futhin> because what happens is that if you are still eating an hour and 15 minutes later from the start of eating
21:07:06 <futhin> a second amount of insulin will be produced
21:07:22 <futhin> and you'll end up with excess insulin, cravings for more snacks, etc.. and be basically screwed  :P
21:07:50 <futhin> but the cool thing
21:08:03 <futhin> is that because the first amount of insulin released is so small
21:08:07 <futhin> that you can pig out
21:08:12 <futhin> and still lose weight
21:08:29 <futhin> because there won't be much insulin to tell the fat cells to uptake fat
21:08:52 <futhin> if you don't eat for morning or lunch, but eat when you get home, try this:
21:09:00 <futhin> come home, and eat a low-carb meal/snack
21:09:09 <futhin> and then wait for dinner, and make sure you finish that off in an hour
21:09:21 <futhin> and don't eat anything after that, unless it's low-carb
21:09:38 <CrowKiller> bread is high carbohydrates?
21:09:49 <CrowKiller> i know at least that gatorade is full of that lol
21:09:58 <futhin> low-carb stuff are:  meat, cheese, lettuce, spinash, most vegetables
21:10:36 <futhin> high-carb stuff are:  bread, starchy foods (grain products, spaghetti, etc), sugars, fruit, pop, juice, etc
21:11:06 <CrowKiller> the basic of a good alimentation is a vegetarian one and not very much food, and we should all drink our urine form time to time lol
21:11:38 <futhin> so anyways, when you get home, just have a low-carb snack/meal and wait for dinner
21:11:38 <CrowKiller> some allergies can be fought with that particular practice
21:11:48 <CrowKiller> dacodac ;p
21:12:11 <futhin> i wouldn't be surprised if you lost more than 2 pounds a day because you are fasting the morning & early afternoon
21:12:21 <futhin> er
21:12:22 <futhin> a week
21:12:30 <futhin> more than 2 pounds a week
21:12:31 <CrowKiller> a day lol
21:12:46 <futhin> but becareful about weighing yourself once a week
21:12:49 <futhin> it won't be accurate
21:12:58 <CrowKiller> i am 5'10" and weighting now about 140lbs
21:13:17 <futhin> eh?
21:13:17 <futhin> wtf?
21:13:21 <futhin> you aren't overweight
21:13:25 <CrowKiller> i know lol
21:13:34 <futhin> um
21:13:38 <CrowKiller> im actually in great physical condition
21:13:54 <futhin> who the hell was i talking to that needed to lose 30 pounds for some air force thing
21:13:58 <CrowKiller> i sprinted the fastest last year at my svhool and jumped the highest
21:14:05 <CrowKiller> dont know
21:14:11 <futhin> lol!
21:14:13 <futhin> sorry
21:14:23 <CrowKiller> that was great info
21:14:26 <futhin> wasted your time :P
21:14:35 <CrowKiller> I myself get lost on a lot of subject sometimes
21:14:38 <CrowKiller> not at all
21:14:44 <futhin> except you could tell obese ppl around you about the insulin thing ;)
21:14:51 <CrowKiller> exactly
21:15:01 <CrowKiller> and those girls around me
21:15:07 <CrowKiller> that try to lsoe weight
21:15:07 <CrowKiller> lol
21:15:12 <futhin> yeah
21:15:21 <futhin> but the diet only applies to carbohydrate addicts
21:15:33 <futhin> people who's bodies are slightly out of whack
21:15:51 <futhin> (not a bad thing, it's probably an old survival trait)
21:15:57 <futhin> s/who's/whose
21:16:22 <CrowKiller> that thing is that a regexp?
21:16:30 <CrowKiller>  s/who's/whose
21:18:17 <CrowKiller> I dont know about them, and they come by every now as "previously known by the reader" references when reading stuff on the Net
21:18:20 <futhin> it's a vi thing
21:18:41 <futhin> linux, vi :P
21:18:47 <futhin> you could get gvim for windows
21:18:52 <futhin> but bleh, vi has a learning curve
21:19:17 <CrowKiller> ha like search who,s and replace by whose
21:19:22 <futhin> even though after you get past that, you'll find yourself working a lot faster, etc
21:19:30 <CrowKiller> its kind easy to understand in fcat, it was in my 1500 pages linux book
21:19:41 <CrowKiller> but i skipped this section
21:20:07 <futhin> well it's understandable, but working with vi and becoming comfortable with it takes some time
21:22:22 <CrowKiller> I woudl put as topic "Actual Project: To design the easiest, simplest and mightiest 32bit Virtual Foth Machine for x86/ARMthumb/PowerPC possible"
21:22:46 <CrowKiller> or something like that
21:23:01 <CrowKiller> because once we have it
21:23:11 <CrowKiller> when everybody agreed on that
21:23:16 <CrowKiller> we can actually share code for it
21:23:20 <CrowKiller> and work together
21:23:33 <CrowKiller> i think thats the standard people are looking for
21:23:42 <CrowKiller> anyway thats what im looking after lol
21:24:47 <CrowKiller> http://www.godsmaze.org/gba/
21:24:56 <CrowKiller> this is the gba code uploader
21:25:05 --- topic: set to 'Actual Project: To design the easiest, simplest and mightiest 32bit Virtual Foth Machine for x86/ARMthumb/PowerPC possible' by futhin
21:25:10 <CrowKiller> they got a NES emulator
21:25:13 <CrowKiller> a cart dumper
21:25:17 <CrowKiller> a savegame dumper and uplaoder
21:25:23 <CrowKiller> all coded serially
21:25:42 <CrowKiller> hmmm
21:25:47 <CrowKiller> not all at all in fact
21:25:49 <CrowKiller> lol
21:25:58 <CrowKiller> i mean those are samll piece of code
21:26:09 <CrowKiller> that can be uplaoded to the gba at startup
21:26:27 <CrowKiller> so that he can do functions that woudl require special hardware
21:27:46 <CrowKiller> the best thing would be an sram cart, storing apps or roms in inexpensive dataflash and sending them to sram could be great
21:28:31 <CrowKiller> on the gamecube theres the high speed io port that is connected to the bus that send the bootcode to the gekko processor, the 485 MHz PowerPc
21:28:37 <CrowKiller> this beast got embedded ram
21:28:46 <CrowKiller> with a one cycle latency or less i think lol
21:29:01 <CrowKiller> it would outperform anything if it could run forth
21:29:05 <futhin> hey, the virtual machine should be also an actual forth chip design
21:29:19 <CrowKiller> you finally understood my point lololol
21:29:34 <CrowKiller> forth is truly a hardware design
21:29:39 <futhin> no, always understood it :P
21:29:53 <CrowKiller> ok lol
21:30:10 <futhin> don't worry, english gets in the way of our communication of ideas, points, etc
21:30:21 <futhin> any language does..  need direct cortical links :P
21:30:54 <CrowKiller> i can understand anything, but i lack "outward communication skills" in english ;p
21:31:18 <CrowKiller> in college i will get to Ottawa/Hull region
21:31:26 <CrowKiller> s/get/go
21:31:28 <CrowKiller> ;p
21:32:04 <CrowKiller> so my english will surely get better
21:32:05 <futhin> well for outward communication skills, try to stick with straightforward logic, and make sure to break up your message so that it is a step-by-step logic thing or whatever
21:32:30 <futhin> i dunno
21:32:36 <futhin> you could've said earlier
21:32:56 <futhin> my point is that forth is also a hardware thingie
21:33:01 <futhin> ;)
21:33:05 <futhin> but i know that
21:33:11 <futhin> and i think most ppl here know that
21:33:19 <futhin> especially onetom
21:33:29 <futhin> and perhaps kc5tja afaik
21:33:50 <futhin> and goshawk
21:34:00 <futhin> that idling bastard :P
21:34:42 <CrowKiller> so if we get the best virtual machine around, it will be the best hardware ;p all free and opensource of course =)
21:35:20 <futhin> well
21:35:39 <futhin> if the virtual machine is also going to be hardware, you'll have to spend a lot of time designing forth chips
21:35:48 <CrowKiller> chuck is the expert
21:35:54 <futhin> learn VLHD and stuff
21:36:01 <CrowKiller> if we come up with a good virtual machine and if he believs in it
21:36:14 <CrowKiller> then hell design it for us im sure
21:36:18 <futhin> fine, we come up with what we think is a good virtual machine, and present it to him
21:36:27 <futhin> and he'll modify it
21:36:29 <CrowKiller> his approach to chip engineering is the best around
21:36:32 <futhin> if necessary
21:36:53 <futhin> in order to fit the real world realities of how his forth chips function
21:37:03 <futhin> and the limits and advantages of the chips
21:37:44 <CrowKiller> look at the instructions for the x18 his masterpiece core
21:38:04 <CrowKiller> I dont understand how it works completely
21:38:14 <futhin> the virtual machine is the same as the forth chip design, i doubt that you can create an effective forth chip design AFTER the virtual machine has been designed
21:38:16 <CrowKiller> at least i came up with a simple to understand if statement
21:38:31 <futhin> it'll be more effective if the virtual machine & forth chip design are both created at the same time
21:38:41 <CrowKiller> i dont understand how the call and and all are implemented
21:38:54 <CrowKiller> thats what i tried to do
21:39:32 <CrowKiller> adding a flag to the stack can implement a lot of thing, if, -if, odd if or even if
21:40:00 <CrowKiller> -if and if are the only one in fact
21:40:12 <CrowKiller> we need in hardare because the otehrs can be made from them
21:40:34 <CrowKiller> to do a is the number odd if statement
21:40:56 <CrowKiller> only rotate the 32 bits of TOS right
21:41:01 <CrowKiller> and do a -if
21:41:30 <CrowKiller> when using asynchronous logic you can afford to synthetise instructions with forth words
21:42:12 <CrowKiller> if the chip design is the bare minimum running at a good enough speed great things can be achieved using the minimum of ressources
21:43:00 <CrowKiller> http://www.colorforth.com/X18.html
21:43:10 <CrowKiller> its supposed to be its masterpiece
21:43:28 <CrowKiller> the if jumping to a nowhere to be found then is confusing
21:43:55 <CrowKiller> a no operand call
21:44:06 <CrowKiller> can only be call what's on the datastack
21:44:42 <CrowKiller> the or opcode is confusing
21:44:57 <CrowKiller> because its really a xor
21:45:02 <CrowKiller> theres such little details
21:45:20 <CrowKiller> that need to be adressed to make it even better
21:45:25 <futhin> hey i'm going to be away for a bit
21:45:34 <futhin> for 45 minutes or so
21:45:42 <CrowKiller> k
22:03:53 <CrowKiller> look at this!!!!!!!!!! http://www.charmedlabs.com/ my dream dev hardware come true lol
22:03:59 <CrowKiller> for the gba of course
22:04:26 <CrowKiller> but it applies for anything from gamecube to xbox, just change the controller
22:04:31 <CrowKiller> i meant the connector
22:05:02 <CrowKiller> imagine forth running on this baby lol
22:10:18 <CrowKiller> I think aha redefined what a forth virtual machine need to be
22:10:44 <CrowKiller> I'll try to come up with something before going to bed
22:14:29 <CrowKiller> ARM vs X86: Arm is a lot nicer when it comes to instructions, look at section 4.4 of http://www.arm.com/techdocs/56QGGJ/$File/ARM7vC.pdf
22:14:57 <CrowKiller> only have to look now at the powerpc architecture
22:16:55 <CrowKiller> this file resumes well the arm7 architecture and is small
22:17:20 <CrowKiller> ill try to get a powerpcp one for the gamecube but since its proprietary it might be difficult
22:33:19 --- join: sbk_ (~500@dsl-65-184-98-221.telocity.com) joined #forth
22:45:15 <CrowKiller> an account on that site would be crucial http://www.warioworld.com/
22:45:32 <CrowKiller> they surely got the specs for the gekko processor up there
22:45:57 <futhin> why you messing with gaming consoles? :P
22:48:32 <CrowKiller> the powerpc in the gamecube is the cheapest abaible processor of this class around
22:50:38 <CrowKiller> and for the virtual machine I'll see right away if my ideas are good enough, if they work well on a poerpc, and arm and a x86, ill keep them
22:56:11 --- quit: sbk_ ("Leaving")
22:59:34 <CrowKiller> I think in the virtual machine there should be only no operand opcodes
23:00:06 <CrowKiller> like in aha, the concept of dictionary could be extended to forth hardware
23:00:35 <CrowKiller> its already in hardware in fact
23:00:52 <CrowKiller> because aha's dictionary are indexed calls
23:01:09 <CrowKiller> ha ill try to come up with something
23:08:41 <CrowKiller> i must go to bed its 2:10 am here, good night everyone
23:09:07 --- quit: CrowKiller ("The 32 bit machine need to be done, we cant do anything without it!")
23:32:56 --- quit: futhin ("bye")
23:59:59 --- log: ended forth/02.06.01