00:00:00 --- log: started forth/04.03.22
00:07:07 --- quit: Serg ()
00:20:22 --- join: crc (crc@1Cust136.tnt1.levittown.pa.da.uu.net) joined #forth
00:33:04 <hovil> anyone know how old chuck is?
00:34:08 <hovil> 66?
00:38:31 <hovil> kc5tja, did you want people interested in FBKestrel to mail you, to get approximate numbers?
00:40:50 --- quit: ayrnieu (Read error: 104 (Connection reset by peer))
01:04:50 --- join: Serg (~z@212.34.52.142) joined #forth
01:07:19 <kc5tja> Approximate numbers for what?
01:07:27 <kc5tja> Pricing?
01:07:49 <kc5tja> For now that is the best way; however, when such numbers become available, I will release them on my website.
01:08:09 <kc5tja> Damn, it's late!!  I need to get going here pretty soon.  :/
01:08:19 * kc5tja got wrapped up in reading information on the PL/M and PL/I languages.
01:14:18 <hovil> kc5tja, approximate number of people interested
01:18:57 <kc5tja> Ahh, well, yes, definitely.  Market research is always a critical component of any business endeavor.
01:21:10 <hovil> looked at colorforth again today
01:21:15 <hovil> the interface still boggles the mind
01:22:17 <hovil> think I might wuss out and try one of the qwertified derivatives
01:23:36 <hovil> well, it doesn't boggle the mind. I just find it quite hard to fluently use a the keyboard layout
01:23:38 * kc5tja nods
01:23:43 <kc5tja> Well, I have to get to bed.
01:23:48 * kc5tja has work and school tomorrow.
01:23:54 <Serg> damn timeshift !
01:23:58 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK")
01:43:02 * Serg is screwing brains w/ W2K inet gate
01:55:26 <hovil> sounds dangerous
02:03:03 <Serg> i try to install it
02:03:20 <Serg> after few crashes, it's partition table was madman's paint
02:04:12 <Serg> i washed it w/ zeroes and attempt from scratch
02:06:28 --- quit: Serg ()
02:43:54 --- quit: crc (Nick collision from services.)
02:43:54 --- quit: ASau (Read error: 104 (Connection reset by peer))
02:44:42 --- join: crc (crc@1Cust218.tnt1.levittown.pa.da.uu.net) joined #forth
03:25:12 --- join: Downer (~Pen@host213-122-34-1.in-addr.btopenworld.com) joined #forth
04:05:12 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth
05:18:12 --- join: ASau (~asau@158.250.48.196) joined #forth
05:25:46 --- quit: hovil ("Leaving")
05:26:47 --- quit: Downer ("brb")
05:55:29 --- join: Herkamire (stjohns@c-24-62-120-109.ne.client2.attbi.com) joined #forth
05:59:41 <ASau> Dobryjj vecher, Herkamire!
05:59:54 <Robert> Privet. :)
06:00:12 <ASau> God kvaell.
06:00:25 <Robert> It's not evening yet :)
06:00:54 <ASau> 5:01 --> Evening.
06:01:20 <mur> privet
06:01:32 <mur> 5.00 is morning!
06:01:33 <mur> ;)
06:02:11 <ASau> It's another "5:00". ;)
06:05:50 --- join: ayrnieu (julian@65.169.246.30) joined #forth
06:06:03 <ASau> Dobryjj vecher, ayrnieu!
06:07:01 <mur> terve ayrnieu 
06:12:43 <qFox> [15:01:30] <mur> 5.00 is morning!
06:13:03 <ASau> Gy!
06:13:09 <ASau> 4-lines editor.
06:14:09 <ASau> How do you like it?
06:31:54 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
06:32:08 <ASau> Dobryjj vecher, proteusguy!
06:57:05 --- quit: crc ()
07:00:35 --- join: Downer (~Pen@host213-122-144-57.in-addr.btopenworld.com) joined #forth
07:17:27 --- quit: I440r (Read error: 104 (Connection reset by peer))
07:37:23 <ASau> qFox. How do you like another "Hacking FIG Forth" tour?
07:37:39 <ASau> How would you like?..
07:38:06 <ASau> I've got a fresh idea.
07:41:40 <qFox> yay
07:41:41 <qFox> :p
07:41:57 <ASau> Well.
07:42:06 <ASau> The main idea is this.
07:42:24 <ASau> Initially you've got bare bone system.
07:42:38 <ASau> You have practically nothing.
07:42:47 <qFox> uhu
07:42:51 <ASau> But you have underlying DOS.
07:43:00 <ASau> Though you can't access it.
07:43:06 <qFox> not directly
07:43:17 <ASau> Well.
07:44:01 <ASau> You have initial plain binary COM file, which is simple memory dump.
07:44:20 <ASau> When you define new words, you change memory.
07:44:33 <qFox> with you so far..
07:44:42 <ASau> So why not to dump memory in another plain binary COM file?
07:45:17 <ASau> So why not dump memory to another plain binary COM file?
07:45:23 <qFox> mmmm
07:45:56 <qFox> you mean write from one area to the other in the memmory? 1 on 1 to a com file?
07:46:07 <ASau> Yes.
07:46:28 <qFox> so expand the fig forth, and save it..
07:46:29 <ASau> You get another plain binary COM file with _extended_ Forth dictionary.
07:46:49 <qFox> i havnet worked with files yet in forth
07:46:53 <qFox> so i wouldnt know how
07:46:58 <ASau> No matter.
07:47:04 <qFox> but i suppose it should be possible, if that theory is correct
07:47:14 <ASau> I've done this many times.
07:47:17 <ASau> :)
07:47:30 <qFox> ok, so if this isnt anything new...whats the hack?
07:47:48 <ASau> The problem is to express solution in as little words as possible.
07:47:57 <qFox> thats always the problem :p
07:48:16 <qFox> but atm i dont even know how to open a file for read/write/whatever
07:48:26 <qFox> so i dont think i'm the right target for htis practice
07:48:27 <ASau> It seems I've achieved less than one screen.
07:48:57 <ASau> qFox. We need only two DOS API functions: CREAT and WRITE.
07:48:57 <qFox> uhm, i dont know, but opening files for writing shouldnt be long
07:49:11 <qFox> and copying one area of the memmory to a file shouldnt be much
07:49:14 <qFox> so why is it so big?
07:49:53 <ASau> You've forgotten about low level coded primitives.
07:50:36 <ASau> Another problem is that Forth restores dictionary state from so called "boot up parameters."
07:50:45 <ASau> Hence you need to change them.
07:50:51 <qFox> yes but if you use that.... oh right
07:53:25 <ASau> Actually you need to change initial DP value and initial LATEST.
07:54:47 <ASau> OK. Are you interested in this?
07:55:10 <qFox> i'll follow your lead, but you are warned that i have very little experience with this
07:56:32 <ASau> Well.
07:56:39 <ASau> How the stuff works.
07:58:57 <ASau> When you've run FORTH.COM it does register initialization, copies initial values to USER variables.
08:00:01 <ASau> You have initial values (in boot parameters area) for these variables:
08:00:14 <ASau> S0  R0  TIB  WIDTH  WARNING  FENCE  DP  VOC-LINK
08:00:24 <qFox> k
08:01:17 <ASau> When your dictionary grows, you get different DP and VOC-LINK values.
08:01:39 <qFox> dp = datapointer, voc-link is dictionary offset?
08:02:13 <ASau> DP is for "dictionary pointer", voc-link is for linked list of vocabularies.
08:02:21 <ASau> : HERE DP @ ;
08:03:52 <qFox> : HERE DP @ ; HERE MSG # 4  OK
08:04:03 <qFox> i guess that means it already existed
08:04:06 <ASau> No. Do not do it.
08:04:13 <qFox> oh heh
08:04:27 <ASau> I just show how it's implemented.
08:04:30 <qFox> k
08:05:00 <ASau> HERE is done as it has been defined like : HERE DP @ ;
08:05:27 <qFox> ok
08:05:42 <ASau> You see why you have to change initial DP.
08:06:14 <qFox> because the dicionary is bigger..?
08:06:21 <qFox> when you use the custom one i mean
08:06:22 <ASau> Right.
08:07:16 <ASau> BTW. You have VLIST word.
08:07:29 <ASau> It lists vocabulary.
08:07:47 <ASau> Like WORDS in ANS or "83".
08:08:09 <ASau> You also have +ORIGIN word.
08:08:17 <qFox> heh
08:08:18 <qFox> ok
08:08:30 <qFox> what does that do?
08:08:58 <ASau> It adds constant displacement.
08:09:12 <ASau> 0 +ORIGIN is the start of Forth system image.
08:09:34 <ASau> Well.
08:10:00 <qFox> like hte offset where the forth system starts in the memmory?
08:10:02 <ASau> Let's suppose we have defined FIN as the latest word.
08:10:15 <ASau> Right.
08:10:32 <qFox> winforth doesnt recognize it, but thats not fig anyways
08:10:45 <qFox> ok with you so far
08:10:46 <qFox> :)
08:11:13 <ASau> When FIN is least recently defined word, LATEST gives its NFA.
08:11:43 <ASau> You can print this word with ID.
08:11:47 <ASau> LATEST ID.
08:12:01 <ASau> ID. ( nfa -- )
08:12:18 <ASau> prints word name.
08:12:33 <qFox> word name at that memmory location?
08:12:41 <ASau> Right.
08:12:52 <qFox> ok and what if there's no definition at that location?
08:13:06 <ASau> It prints what it finds.
08:13:12 <ASau> Well.
08:13:26 <ASau> How you save you new DP.
08:13:53 <ASau> Initial DP is located at:  DECIMAL 30 +ORIGIN
08:14:18 <qFox> whats in the first 30 bytes?
08:14:18 <ASau> Hence you need:  HERE DECIMAL 30 +ORIGIN !
08:15:13 <ASau> There're another boot parameters, two jump commands, current system version etc.
08:15:20 <qFox> k
08:15:59 <ASau> When system boots up it sets FORTH vocabulary as CONTEXT and CURRENT
08:16:33 <ASau> Any other vocabularies are referred from FORTH one.
08:17:04 <ASau> (If you have not hacked your dictionary at low level that's absolute true.)
08:17:14 <qFox> CURRENT is the vocabulary where words are added to you define with : ,and CONTEXT is.... i dont know.. ?
08:17:36 <ASau> CONTEXT is searched first when you interprete word.
08:17:44 <qFox> ah ok
08:18:02 <qFox> so yo ucan have multiple vob's loaded, and he'll search all of them?
08:18:09 <ASau> You may define word in CURRENT voc. but look into CONTEXT
08:18:20 <qFox> ok
08:18:31 <ASau> Well.
08:19:41 <ASau> Other boot parameters concern memory layout: bottoms of stacks, user area address and start of disk buffers area.
08:19:48 <ASau> So we don't need them.
08:19:55 <qFox> ok
08:20:18 <ASau> But you have changed FORTH vocabulary.
08:20:58 <ASau> And system forgets everything when reboots (see COLD word).
08:21:15 <qFox> yea, i understand
08:21:17 <ASau> So you have to update corresponding parameter.
08:21:44 <ASau> To do this means:
08:22:05 <ASau> a) swithing CONTEXT to FORTH vocabulary;
08:22:26 <ASau> b) extracting latest word of CONTEXT voc.;
08:22:42 <ASau> c) updating boot parameters.
08:22:47 --- join: networm (~networm@L0656P14.dipool.highway.telekom.at) joined #forth
08:23:04 <qFox> but isnt CONTEXT automaticly set to FORTH voc?
08:23:35 <ASau> Code:  FORTH  CONTEXT @ @  DECIMAL 12 +ORIGIN ! 
08:23:39 <qFox> or does it have at least two voc's by default?
08:23:46 <qFox> the forth and the user one, or something...
08:24:11 <ASau> Automatically FORTH is set as CONTEXT and CURRENT at boot.
08:24:45 <qFox> so why do we need to switch CONTEXT to FORTH?
08:24:48 <ASau> But you may define as many vocabularies as you have memory.
08:25:32 <ASau> If you have another CONTEXT at that time, you don't get latest word in _FORTH_ voc.
08:25:42 <qFox> ok..
08:28:19 <ASau> You get latest word for you CONTEXT voc.
08:28:26 <ASau> Well.
08:28:58 <ASau> After you've updated these two boot parameters you can bravely save memory dump to COM file.
08:29:58 <qFox> ok, understood
08:30:49 <ASau> Now how to save.
08:31:07 <ASau> Let's take a break.
08:31:11 --- join: rO| (~JBouncer@pD9E59C89.dip.t-dialin.net) joined #forth
08:31:13 <qFox> :)
08:33:34 <qFox> let me try to recap, see if i got everything right... CONTEXT is the first vocabulary that forth looks up words in. CURRENT is the vocabulary where newly defined words will be added into. +ORIGIN is the offset where the FORTH system starts in the memmory. 30 +ORIGIN is the offset of the first word of the FORTH vocabulary.
08:34:21 <qFox> and here simply puts the address of the dp on the stack
08:39:09 <ASau> First approximation. OK.
08:39:16 <ASau> Well.
08:40:06 <ASau> To see how it works in reallity you can do this.
08:40:06 <ASau> 1. Define several words and 
08:40:06 <ASau> call WARM
08:40:16 <ASau> 2. Define several words and call COLD
08:40:50 <ASau> See difference in VLIST results.
08:41:30 <ASau> 3. Define several words, change initial DP and initial LATEST (see above), call COLD.
08:41:42 <ASau> See difference in VLIST result.
08:42:26 <ASau> That's how boot parameters influence on system.
08:43:45 <qFox> first WARM then COLD? i would think the other way around.... COLD indicating the system isnt booted yet, WARM when its done...
08:44:13 <ASau> Order of 1 and 2 does no matter.
08:44:24 <qFox> oh
08:44:42 <ASau> That're examples.
08:45:16 <ASau> You can see in reality, why I was explaining all these.
08:46:14 <qFox> ok
08:49:05 <ASau> Well.
08:49:10 <ASau> Now DOS.
08:49:42 <ASau> First rule for Intel machine code: Intel machine code is _octal_.
08:49:56 <ASau> Not hexadecimal, but _octal_.
08:50:12 <ASau> That's why I will use octal numbers.
08:50:22 <ASau> DECIMAL : OCTAL 8 BASE ! ;
08:50:38 <qFox> k
08:51:39 <ASau> Well. You know that DOS API call is interrupt service.
08:51:54 <qFox> yes, and that ints are slow
08:51:56 <qFox> :)
08:52:19 <ASau> We need: load registers, call for interrupt, save registers.
08:52:46 <ASau> Of course we will load registers from stack. And save them back there.
08:52:52 <qFox> uhu
08:53:37 <ASau> Registers: AX CX DX BX SP BP SI DI ,--- from 0 to 7.
08:53:51 <ASau> Or: AL CL DL BL AH CH DH BH
08:54:06 <qFox> why not just ax bx cx dx...
08:54:10 <ASau> 3 bits for registed in op code.
08:54:22 <qFox> oh thats what you just said, lol sorry
08:54:29 <ASau> It's longer to explain.
08:55:35 <ASau> PUSH command is 12<r>Q, POP command is 13<r>Q.
08:55:45 <qFox> <r> ?
08:55:50 <ASau> <r> is digit for register.
08:55:53 <qFox> oh ok
08:55:53 <ASau> See above.
08:56:14 <ASau> 120Q is for PUSH AX, 130Q is for POP AX.
08:56:35 <ASau> What else we need.
08:56:46 <ASau> PUSHF is 234Q.
08:56:54 <ASau> INT is 315Q
08:57:04 <ASau> DOS intr. is 41Q.
08:57:49 <ASau> MOV.B <r>, <imm> is 26<r>Q
08:58:01 <ASau> MOV.W <r>, <imm> is 27<r>Q
08:58:10 <ASau> Immediate operand follows.
08:59:02 <ASau> Declaring primitive words is done like I explained yesterday.
08:59:12 <ASau> : CODE CREATE SMUDGE ;
09:00:11 <ASau> HEX ( exception ) : NEXT E9 C, ( your NEXT address ) HERE 2+ - , ;
09:01:57 --- join: wossname (wossname@HSE-MTL-ppp77708.qc.sympatico.ca) joined #forth
09:03:24 <ASau> CODE (CREAT)  ( create DOS file )  OCTAL 121 C, 132 C, 264 C, 74 C, 315 C, 41 C, 230 C, 234 C, NEXT
09:04:12 <ASau> I.e.: cx pop dx pop  74 # ah mov  41 int ax push pushf next
09:04:35 <ASau> : NAME  BL WORD HERE COUNT OVER + 0 SWAP C! ;
09:04:57 <ASau> : CRE  NAME 0 CREAT IF ." DOS ERROR # " . ABORT THEN ;
09:05:21 <ASau> Now you can create as many files as you have free file handles.
09:05:30 <ASau> CRE filename
09:05:38 <ASau> returns file handle.
09:06:15 --- quit: Downer (Read error: 60 (Operation timed out))
09:07:14 <ASau> And I've forgotten:
09:07:22 <ASau> : CREAT (CREAT) 1 AND ;
09:08:16 <qFox> ye
09:08:16 <qFox> :)
09:09:25 --- join: Downer (~Pen@host213-122-144-57.in-addr.btopenworld.com) joined #forth
09:10:01 <qFox> why the 1 AND btw?
09:11:38 <ASau> Well.
09:12:12 <ASau> You have carry flag set when error, but other flags are set of reset inpredictably.
09:12:38 <ASau> When you PUSHFed, you have carry flag as the least significant bit.
09:12:45 <qFox> oh ok
09:13:25 <ASau> Well, writing is easier as you need not to deal with file name.
09:14:03 <ASau> CODE (WRITE) 133 C, 131 C, 132 C, 264 C, 100 C, 315 C, 41 C, 230 C, 234 C, NEXT
09:14:23 <ASau> ( addr cnt fh -- cnt/err 0/1 )
09:14:53 <ASau> : WRITE (WRITE) 1 AND ;
09:15:14 <ASau> To dump Forth you:
09:15:45 <ASau> 0 +ORIGIN HERE OVER - ( file handle ) WRITE . .
09:16:21 <ASau> First printed number should be 0, second one should be equal to  HERE 0 +ORIGIN - .
09:17:12 <ASau> "Altogether now!" :)
09:18:13 <ASau> If you do everything right, you're able to run extended Forth.
09:19:49 <qFox> hmmm crash. wasnt that B88 my NEXT addr?
09:20:36 <ASau> No.
09:20:43 <ASau> You have 159.
09:21:00 <ASau> HEX 159 ? should give AD
09:21:24 <qFox> gives me ... crap
09:21:37 <qFox> HEX 159 ? -7453 OK
09:21:46 <ASau> Oi.
09:21:52 <ASau> HEX 159 C@ .
09:21:58 <ASau> should be AD
09:22:04 <qFox> yep
09:22:05 <qFox> :)
09:24:18 <qFox> still crashes, except it restarts fig now, rather then closing it
09:25:21 <ASau> Reread what you typed in. You can decompile all machine code.
09:34:24 <ASau> My does not.
09:34:35 <ASau> FORTH  CONTEXT @ @  DECIMAL 12 +ORIGIN !
09:34:36 <ASau> HERE DECIMAL 30 +ORIGIN !
09:34:36 <ASau> : CODE CREATE SMUDGE ;
09:34:37 <ASau> HEX : NEXT E9 C, 128 HERE 2+ - , ;
09:34:37 <ASau> : OCTAL 8 BASE ! ;
09:34:38 <ASau> CODE (CREAT) OCTAL 121 C, 132 C, 264 C, 74 C, 315 C, 41 C, 230 C, 234 C, NEXT
09:34:38 <ASau> : NAME  BL WORD HERE COUNT OVER + 0 SWAP C! ;
09:34:40 <ASau> : CRE  NAME 0 (CREAT) 1 AND IF ." DOS ERROR # " . ABORT THEN ;
09:34:42 <ASau> CRE 22MAR04.COM .
09:34:44 <ASau> CODE (WRITE) OCTAL 133 C, 131 C, 132 C, 264 C, 100 C, 315 C, 41 C, 230 C, 234 C, NEXT
09:34:46 <ASau> : WRITE (WRITE) 1 AND ;
09:34:48 <ASau> : WHOLE 0 +ORIGIN HERE OVER - ;
09:34:50 <ASau> WHOLE 5 WRITE . .
09:35:06 <ASau> You have to change 128 in NEXT to your 159.
09:36:52 <qFox> um
09:36:59 <qFox> i think i just borked up my keyboard
09:37:01 <qFox> or somethjing
09:37:29 <qFox> nm
09:43:24 <ASau> Actually I mistyped 3 times:
09:43:25 <ASau> : CODE CREATE SMUDGE ;
09:43:26 <ASau> HEX : NEXT E9 C, 128 HERE 2+ - , ;
09:43:26 <ASau> : OCTAL 8 BASE ! ;
09:43:27 <ASau> CODE (CREAT) OCTAL 131 C, 132 C, 264 C, 74 C, 315 C, 41 C, 120 C, 234 C, NEXT
09:43:28 <ASau> : NAME  BL WORD HERE COUNT OVER + 0 SWAP C! ;
09:43:28 <ASau> : CRE  NAME 0 (CREAT) 1 AND IF ." DOS ERROR # " . ABORT THEN ;
09:43:29 <ASau> CODE (WRITE) OCTAL 133 C, 131 C, 132 C, 264 C, 100 C, 315 C, 41 C, 120 C, 234 C, NEXT
09:43:31 <ASau> : WRITE (WRITE) 1 AND ;
09:43:33 <ASau> : WHOLE 0 +ORIGIN HERE OVER - ;
09:43:35 <ASau> : SAVE WHOLE CRE WRITE IF ." DOS ERROR # " . ELSE . THEN ;
09:43:37 <ASau> FORTH  CONTEXT @ @  DECIMAL 12 +ORIGIN !
09:43:39 <ASau> HERE DECIMAL 30 +ORIGIN !
09:43:41 <ASau> SAVE 22MAR04.COM
09:43:54 <ASau> This works.
09:44:06 <wossname> woah
09:44:17 <wossname> what forth needs is a published magazine
09:44:38 <wossname> and you could stick all this shit in it, and we could rush home to power up our dinky basic machines and type in the code
09:44:47 <wossname> that would be :cool:
09:46:11 <qFox> i think the problem with forth now
09:46:14 <qFox> is that its old
09:46:19 <qFox> and because of that
09:46:34 <qFox> there are still quite alot of forth programmers that know all kind of tricks
09:46:37 <qFox> so to come up with something new
09:46:43 <qFox> is probably impossible
09:47:35 <ASau> "There's nothing new in this.
09:47:37 <wossname> computer science is always creating new stuff
09:47:46 <ASau> Because there's nothing new at all."
09:47:54 <wossname> at the least, the magazine could show how to fit every new thing in a screen ;l
09:48:38 <ASau> I've written 4 line editor today.
09:55:54 --- quit: networm (Read error: 113 (No route to host))
10:11:02 --- join: networm (~networm@L0649P20.dipool.highway.telekom.at) joined #forth
10:13:51 --- quit: wossname ("brb")
10:14:48 <qFox> if i include a file, how do i make sure i dont include the same header file over and over again?
10:15:35 <qFox> something like in C, #ifndef HEADER_INCLUDED #define HEADER_INCLUDED .... #ENDIF
10:16:16 <qFox> but i'm not sure how to skip it from processing the file
10:17:02 <Herkamire> qFox: put the #endif at the bottom of the file
10:17:20 <qFox> thats c. or is it the same style in forth?
10:17:33 <Herkamire> oh
10:17:35 <qFox> :)
10:17:40 <Herkamire> try NEEDS
10:17:55 <qFox> checks if a file is loaded and loads it if not?
10:18:03 <Herkamire> iirc
10:18:07 <qFox> tnx :)
10:18:21 <Herkamire> might not be a standard word. but check for it
10:18:21 <qFox> yay
10:18:25 <qFox> works for winforth
10:19:07 <qFox> gonna continue on that set game ;)
10:19:18 <Herkamire> :)
10:22:47 --- quit: proteusguy (Read error: 54 (Connection reset by peer))
10:22:56 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
10:43:50 --- join: MrReach (~spam@209.181.43.190) joined #forth
10:44:29 --- quit: MrReach (Client Quit)
10:48:16 <Herkamire> qFox: what sort of graphics will your set game have?
10:50:43 --- join: _proteus (~proteusgu@65.166.249.10) joined #forth
10:52:15 --- quit: fridge (orwell.freenode.net irc.freenode.net)
10:52:15 --- quit: ChanServ (orwell.freenode.net irc.freenode.net)
10:52:15 --- quit: proteusguy (orwell.freenode.net irc.freenode.net)
10:52:15 --- quit: SDO (orwell.freenode.net irc.freenode.net)
10:52:15 --- quit: l0C0 (orwell.freenode.net irc.freenode.net)
10:52:15 --- quit: chandler (orwell.freenode.net irc.freenode.net)
10:52:16 --- quit: ayrnieu (orwell.freenode.net irc.freenode.net)
10:52:16 --- quit: Downer (orwell.freenode.net irc.freenode.net)
10:52:16 --- quit: hefner (orwell.freenode.net irc.freenode.net)
10:52:24 --- quit: ASau ()
10:52:36 <qFox> i honestly have no idea yet. i was thinking about how to do that, but seeing as i cant seem to get a dos display mode where i can control pixels myself, i'll have to do it via the windows sdk, but i dont know anything about it yet.... so thats gonna be a long way
10:52:54 <qFox> of course i have a character output, but thats not very practical for the game
10:53:44 --- join: hefner (~hefner@pool-151-196-19-248.balt.east.verizon.net) joined #forth
10:54:18 --- quit: hefner (orwell.freenode.net irc.freenode.net)
10:54:52 <qFox> Set table:
10:54:52 <qFox>  /   1    2    3    4
10:54:52 <qFox>  A  1212 2131 2123 1321
10:54:52 <qFox>  B  3333 1331 2211 3112
10:54:52 <qFox>  C  1311 2121 2132 3231
10:54:52 --- join: fridge (~fridge@dsl-203-33-165-6.NSW.netspace.net.au) joined #forth
10:54:53 --- join: chandler (~chandler@64-145-60-36.client.dsl.net) joined #forth
10:57:37 --- quit: chandler (Killed by fontana.freenode.net (Nick collision))
10:58:13 --- join: chandler_ (~chandler@64-145-60-36.client.dsl.net) joined #forth
10:58:15 --- join: ChanServ (ChanServ@services.) joined #forth
10:58:15 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
10:58:15 --- join: l0C0 (loco@bor.gs) joined #forth
10:58:15 --- join: SDO (~SDO@68.170.20.253) joined #forth
10:58:15 --- join: chandler (~chandler@chandler.registered.freenode) joined #forth
10:58:15 --- mode: irc.freenode.net set +o ChanServ
10:58:28 --- join: ayrnieu (julian@65.169.246.30) joined #forth
10:59:06 --- quit: ayrnieu (Remote closed the connection)
10:59:07 --- quit: proteusguy (Success)
10:59:09 --- quit: chandler (Connection reset by peer)
10:59:10 --- join: hefner (~hefner@pool-151-196-19-248.balt.east.verizon.net) joined #forth
11:00:47 --- join: ayrnieu (julian@65.169.246.30) joined #forth
11:04:04 <Herkamire> SET!
11:04:33 <Herkamire> 1C, 2B, 4A
11:04:46 <qFox> :p
11:04:53 <Herkamire> neat
11:05:07 <qFox> but this is a bit harder, or at least unusual, compared to the images version
11:05:20 <Herkamire> yeah, it's way harder for me to find sets
11:06:27 --- quit: warpzero (orwell.freenode.net irc.freenode.net)
11:06:42 <Herkamire> SET!
11:06:58 <Herkamire> 2A, 2B, 4C
11:09:12 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth
11:12:10 --- nick: chandler_ -> chandler
11:49:48 --- join: Downer (~Pen@host213-122-29-243.in-addr.btopenworld.com) joined #forth
12:09:26 --- join: Mark4 (~Mark4@64.47.44.254) joined #forth
12:10:22 --- nick: Mark4 -> I440r
12:10:35 --- mode: ChanServ set +o I440r
12:22:38 --- quit: Downer ("brb")
12:23:52 --- quit: I440r ("Leaving")
12:40:24 --- nick: _proteus -> proteusguy
12:48:56 --- quit: rO| (Read error: 60 (Operation timed out))
13:08:46 --- quit: SDO (Read error: 104 (Connection reset by peer))
13:39:20 <Herkamire> SET!
13:39:34 <Herkamire> 1A, 1B, 2C
14:04:10 --- quit: networm ("Client exiting")
14:27:54 <qFox> haha
14:28:18 <qFox>  /   1    2    3    4
14:28:18 <qFox>  A  3232 3112 1211 2311
14:28:18 <qFox>  B  1332 1122 2131 3312
14:28:18 <qFox>  C  1331 1113 3231 3311
14:28:21 <qFox> :p
14:28:38 <qFox> i'm still working on the code to find the sets, being distracted so its not going fast ;)
14:29:48 <qFox> what does CELLS do again? appearantly i mixed it up with CELL so the outcome wasnt quite what i expected :p
14:30:09 <qFox> CELLS doesnt do anything visual for winforth
14:30:16 <qFox> (and is code..)
14:44:19 --- quit: proteusguy (Read error: 113 (No route to host))
14:44:21 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
14:57:27 --- join: chandler_ (~chandler@64-145-60-36.client.dsl.net) joined #forth
14:59:33 --- quit: chandler (Nick collision from services.)
14:59:40 --- nick: chandler_ -> chandler
14:59:50 --- join: kc5tja (~kc5tja@66-91-231-74.san.rr.com) joined #forth
14:59:57 --- quit: kc5tja (Client Quit)
15:00:03 --- join: kc5tja (~kc5tja@66-91-231-74.san.rr.com) joined #forth
15:00:27 --- mode: ChanServ set +o kc5tja
15:03:12 --- join: skylan (~sjh@vickesh01-4960.tbaytel.net) joined #forth
15:08:49 --- quit: proteusguy (Read error: 113 (No route to host))
15:09:48 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
15:16:02 <Herkamire> qFox: : CELL 4 ; : CELLS CELL * ;
15:16:18 <Herkamire> CELL is the number of bytes in a stack item
15:16:27 <Herkamire> so on palm pilot, CELL is 2
15:16:53 <qFox> yes but appearantly its the behaviour of CELLS to not put anythign on the stack if there isnt anything on it in the first place
15:16:59 <qFox> no error either
15:17:02 <qFox> (in winforth)
15:17:25 <Herkamire> does winforth usually give you an error for stack underflow?
15:17:34 <qFox> normally yes
15:17:37 <qFox> not for CELLS
15:17:43 <Herkamire> weird
15:17:59 <qFox> .s  empty  ok
15:17:59 <qFox> cells  ok
15:17:59 <qFox> .s  empty  ok
15:18:01 <Herkamire> CELLS should not add a stack item.
15:18:12 <qFox> replace it 
15:18:14 <qFox> sorry
15:18:48 <Herkamire> sounds like a bug. probably some optomization in CELLS does not check stack depth because it does not change the stack depth
15:19:34 <qFox> maybe, it confused me
15:41:46 <Herkamire> 3B 3C 1C
15:46:59 --- join: matt_ (1000@adsl-64-160-166-15.dsl.snfc21.pacbell.net) joined #forth
15:47:05 --- nick: matt_ -> Sonarman
15:54:03 <kc5tja> Maybe CELLS in winforth is defined as : CELLS DEPTH IF CELL * THEN ;
15:54:08 <kc5tja> so as to prevent stack underflow.
15:54:16 <kc5tja> I don't know winforth, so I don't know.
15:54:25 <kc5tja> But that being said, I now have to cruise -- time to go to work.  >:(
15:54:28 <qFox> yea probably
15:54:39 <qFox> see cells
15:54:39 <qFox> CELLS IS CODE
15:54:39 <qFox>         401540 8D1C9D00000000   lea     ebx, 0 ( "400000" : 0xFFBFFFFC )
15:54:39 <qFox> [ebx*4]
15:55:40 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK")
16:24:27 --- join: blockhead (default@dialin-224-tnt.nyc.bestweb.net) joined #forth
16:31:39 --- quit: proteusguy (Read error: 54 (Connection reset by peer))
16:32:10 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
16:41:21 --- join: Sonarman_ (1000@adsl-67-113-234-164.dsl.snfc21.pacbell.net) joined #forth
16:41:45 --- quit: Sonarman (Nick collision from services.)
16:41:47 --- nick: Sonarman_ -> Sonarman
16:42:29 <Sonarman> 20:47 < ASau> I don't know anyone who likes FCBs.
16:42:29 <Sonarman> 20:48 < I440r> ugh no
16:42:29 <Sonarman> 20:48 < I440r> fcb's are horrible
16:42:29 <Sonarman> 20:48 < ASau> kc5tja. NO!
16:42:29 <Sonarman> 20:48 < ASau> kc5tja. DOS 2.0 has file handles API.
16:42:32 <Sonarman> 20:48 <@kc5tja> ASau: SO DOES CP/M 86!
16:42:38 <Sonarman> what are FCBs?
16:43:12 <Sonarman> never mind, sorry
16:51:18 <blockhead> fcbs? ... oh: file control blocks.  Dang.  That brings back memories of my old z80 CP/M system,  :D
16:52:39 <Sonarman> :)
16:53:49 * blockhead learned about fbcs when I was typoing in a program for file undeletion.
16:55:51 --- join: imaginator (~George@georgeps.dsl.xmission.com) joined #forth
17:11:05 <blockhead> http://www.1112.net/lastpage.html
17:13:59 <imaginator> Hi Forth'ers
17:14:52 <blockhead> hi
17:15:39 <Sonarman> hi
17:19:08 <blockhead> 'nn all
17:19:14 --- quit: blockhead ("Client Exiting")
17:21:27 * imaginator was outside playing, after reaching the "Last Page"
17:23:03 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth
17:45:12 <qFox> ok someone please clear up my confusion. if a word expects 3 parameters, 1 2 and 3 to be on the stack, and 1 should be the TOS, whats the proper stack description?
17:46:38 <qFox> is it ( 1 2 3 -- ) or ( 3 2 1 -- )
17:47:12 <tathi> ( 3 2 1 -- )
17:47:22 <qFox> and for output? when 1 is tos?
17:47:46 <qFox> i mean result, return, whatever..
17:48:41 <qFox> ( -- 3 2 1 ) ?
17:50:13 <tathi> yes
17:50:18 <qFox> oki :)
17:57:14 --- quit: proteusguy (Read error: 113 (No route to host))
17:58:12 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
18:00:48 --- quit: proteusguy (Read error: 54 (Connection reset by peer))
18:02:15 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
18:04:29 --- quit: proteusguy (Connection reset by peer)
18:04:51 <slava> hi all
18:05:33 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
18:06:14 <Sonarman> hi slava 
18:09:16 <ianp> hello slava 
18:11:56 --- join: I440r (~mark4@64.3.99.130.ptr.us.xo.net) joined #forth
18:13:37 --- quit: proteusguy (Read error: 113 (No route to host))
18:14:01 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
18:14:37 --- quit: tathi ("leaving")
18:20:25 <qFox> can i safely do a loop in a loop without screwing up?
18:20:40 <qFox> 10 0 do 10 0 do 10 0 do loop loop loop
18:21:04 <qFox> and is the counter in each loop still i ? or...
18:21:22 <qFox> and... how do you get the i from the other loops? :p
18:24:44 <qFox> ok you can
18:24:58 <qFox> and you get the i by putting it on the stack before entering the next loop? :)
18:25:06 --- join: ayrnieu_ (julian@65.169.246.30) joined #forth
18:25:24 --- quit: ayrnieu (Remote closed the connection)
18:27:57 <Herkamire> qFox:  you can nest loops
18:28:01 <Herkamire> i is from the innermost loop
18:28:20 <Herkamire> some forths have J wich is lik I but from the next loop out.
18:28:32 <Herkamire> some forths have K which is the third loop out etc...
18:28:36 <qFox> k
18:30:46 <slava> is there any forth where you can write something like : some-word << hello world >> execute ;
18:30:50 <Herkamire> if you can, you're probably better of factoring more, so you don't have to use J or K
18:30:54 <slava> and have << ... >> push an XT of the code inside those
18:31:11 <slava> Herkamire, how do you pick good names for factors?
18:31:31 <qFox> factoring?
18:31:33 <Herkamire> I just try to think of something that makes sense
18:31:47 <qFox> oh breaking up code thing?
18:31:51 <Herkamire> if it really doesn't do anything helpful except help you implement xxx, then I usually name it -xxx
18:32:40 <slava> i use xxx* for a more lower-level form of xxx that possibly takes additional parameters
18:32:56 <slava> and sometimes [xxx] (i know this has diff. meaning in forth)
18:33:07 <slava> [xxx] for custom control strucutres that help impl. of word xxx
18:33:48 <qFox> this is for checking for a match, which is the straightforward approach imo. 3 loops for each card... 1 2 3, 1 2 4, 1 2 5, etc
18:33:57 <qFox> checking for a set, i should say
18:34:34 <Herkamire> qFox: you mean checking if there is a set anywhere on the board?
18:34:37 <qFox> yes
18:34:49 <qFox> basicly just check each card combination
18:35:15 <Herkamire> right
18:35:23 <Herkamire> you're storing a card in one cell right?
18:35:24 <qFox> it probably wont be very fast, but it wont have to be done very often so thats ok
18:35:26 <qFox> yep
18:35:43 <qFox> i hope i can safely asume that a cell is 4 bytes?
18:35:46 <qFox> on most systems..?
18:36:10 <Herkamire> on desktops and laptops
18:36:14 <Herkamire> not on PDA
18:36:25 <qFox> i can live with that :p
18:37:59 <qFox> i'll just have three loops, one for each card positition, and check for all equal or odd
18:38:36 <qFox> continue with next loop if neither are encountered, at the end of checking, make sure all 4 properties are equal/odd
18:38:41 <qFox> and do some output 
18:39:21 <qFox> although you dont need to check at the end since you'll continue if neither were encountered so...
18:39:47 <Herkamire> why three loops?
18:39:52 <qFox> three cards
18:39:53 <Herkamire> I only see why you'd want/need two
18:40:24 <Herkamire> oh, you're right
18:40:28 <qFox> 12 cards, 3 cards per check, 1 2 3, 1 2 4, 1 2 5 ... 1 3 4
18:40:46 <qFox> although i'm positive there's a entirely different approach to this problem
18:40:54 <qFox> faster/easier/shorter one
18:41:05 <Herkamire> what are you going to do if there's no match?
18:41:11 <qFox> but optimization comes later :p
18:41:38 <qFox> well the odds on that are very small, but if soo, display a message about the rarity and shuffle and deal again
18:42:27 <qFox> 0.1%
18:42:44 <qFox> hm
18:42:47 <qFox> is it 12 or 15 cards?
18:43:09 <qFox> ohj
18:43:15 <qFox> right
18:43:38 <qFox> so if after 12 cards no set was made, you deal another 3 cards, and if even THEN no set is made, ...
18:43:43 <qFox> well sod that, i'll do that later :p
18:43:55 <qFox> for now, if there's no set at 12 cards i'll deal again
18:47:46 <qFox> hm
18:47:49 <qFox> i just realized 
18:48:07 <qFox> that i could do things a bit easier, by just referencing the cards on the table by their number
18:48:30 <qFox> rather then actually copying the properties to the array that contains "cards on the table"
18:48:44 <qFox> just card x from the deck
18:49:15 <qFox> and a variable that holds how many cards have been taken from the deck so far (so starting at 12, incrementing every time one card is taken... etc)
18:56:19 --- quit: Sonarman (Read error: 60 (Operation timed out))
19:03:03 <Herkamire> interesting
19:03:13 <Herkamire> is it possible to have 15 cards with no match?
19:03:17 <qFox> yes
19:03:19 <qFox> 0.1% 
19:03:31 <Herkamire> how'd you calculate that?
19:03:33 <qFox> (according to http://www.reed.edu/~mcphailb/applets/set/ , i havent done the math myself)
19:03:34 <qFox> :p
19:03:37 <qFox> i cheat
19:03:57 <qFox> nah its just probability
19:05:43 <Herkamire> cool
19:05:52 <qFox> #math ;)
19:06:06 <qFox> those folks can explain it much better then i could
19:06:19 <qFox> its all in my head, but dont ask me to explain...
19:08:03 --- join: arke (~Chris@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth
19:08:23 <arke> hi
19:08:25 <arke> :)
19:09:18 <arke> hiya
19:11:53 --- join: Sonarman (1000@adsl-67-113-234-67.dsl.snfc21.pacbell.net) joined #forth
19:12:08 <arke> hi matt
19:12:18 <Sonarman> hi chris
19:12:38 <arke> :)
19:12:41 <arke> (:
19:13:02 <Sonarman> 69
19:13:20 <arke> ok, lets go
19:16:29 <qFox> Herkamire> can i manipulate the i of the outer loops?
19:20:17 <Sonarman> anyone know what SMUDGE does in FIG?
19:20:50 * qFox points at ASau
19:21:23 <imaginator> qFox: you could probably build your own looping constructs 
19:21:23 <Sonarman> i would have asked him if he were here :)
19:21:26 <ayrnieu_> sonarman - the same thing it does in other forths?  temporarily bastardizes the new colon-definition so that you can refer to previously defined colon-definitions of the same name.
19:22:21 <Sonarman> ah, thanks ayrnieu.  how long does it remain "bastardized"; in other words, how do you get to the new colon definition?
19:22:51 <Fractal> Well the word ; unsmudges the last colon definition
19:23:04 <Sonarman> ok, thanks
19:24:56 <ayrnieu_> fractal - I think that you should have said 'legitimizes' rather than 'unsmudges'
19:26:06 <imaginator> What is the theory behind how Moore's latest Forth works?
19:26:19 <imaginator> IIRC in the video I watched he said that ; was like a return.
19:26:36 <Fractal> Er well ya whatever the word is in the forth you're using. Personally I don't see any reason for the whole smudging myself... If you do absolutely need to use a previously defined word, there are ways around it
19:27:03 <Fractal> : myword' myword ;   : myword ... myword' ... ;
19:27:05 <Fractal> and so on
19:28:02 <ayrnieu_> imaginator - Moore's colon-definitions live in a contiguous word-space of N named entry points and M exit points.
19:28:47 <ayrnieu_> imaginator - I prefer it =) since it makes it natural to produce 'safe' and 'raw' versions of words.
19:29:22 <ayrnieu_> : safe-/ dup 0= if freak-out-and-die ; then : / (/) ;
19:29:47 <imaginator> oh, I see.
19:30:35 <ayrnieu_> : hello-world s" world" : hello-you ." Hello, " type [char] ! emit CR ;
19:30:54 <ayrnieu_> s" imaginator" hello-you
19:31:40 <imaginator> That's wild.  :)
19:32:45 <arke> eh?????
19:32:47 <arke> ???
19:32:47 <arke> ??
19:34:04 <madgarden> ayrnieu_, is there a ; missing in hello-world?
19:34:16 <ayrnieu_> madgarden - no.
19:34:34 <madgarden> well... where is hello-you defined?
19:34:40 <imaginator> madgarden: he said: "contiguous word-space of N named entry points and M exit points."
19:34:57 <madgarden> Ohh...
19:34:59 <ayrnieu_> madgarden - right after I define hello-world -- please read back a bit to get the context
19:35:19 <madgarden> So it's like a weird multiple entry-points thing going on.
19:35:28 <madgarden> chaining word definitions
19:35:42 <ayrnieu_> in Color Forth my : there would have a different color, so you wouldn't have to worry about whether I meant to compile a call to it into the hello-world or to call it interactively or whatnot.
19:35:52 <madgarden> Right.
19:35:55 <madgarden> Neato.
19:36:39 * arke doesn't get it
19:37:04 <madgarden> arke, hello-world executes "hello-world" followed by "hello-you".
19:37:13 <madgarden> hello-you just executes "hello-you'
19:37:24 <madgarden> They both share a ;
19:37:35 <madgarden> two : definitions, one ; between them.
19:38:10 <ayrnieu_> looking only at that, yes, you see two entry points -- the first supplying a default argument.
19:39:58 --- quit: proteusguy (Read error: 54 (Connection reset by peer))
19:40:06 --- join: _proteus (~proteusgu@workstations.pinnaclesports.com) joined #forth
19:40:36 <arke> eh...
19:41:09 <arke> so something like : set-base base : ! (!) ;
19:41:14 <arke> (lame example)
19:44:23 --- quit: ayrnieu_ ("Changing server")
19:44:26 --- join: ayrnieu (julian@65.169.246.30) joined #forth
19:44:40 <arke> eh...
19:44:40 <arke> so something like : set-base base : ! (!) ;
19:44:41 <arke> (lame example)
19:45:03 <madgarden> I've often thought that using some syntax like this might make Forth code more intuitive:
19:45:07 <madgarden> foo :bar
19:45:07 <madgarden> 10 :constant: fred
19:45:07 <madgarden> variable: bob
19:45:23 <slava> how does this work?
19:45:54 <madgarden> The colon indicates where the arguments are coming from (if they exist).
19:46:13 <madgarden> :bar takes the argument that comes before.
19:46:32 <madgarden> :constant: takes argument before and after (parsing a token)
19:46:45 <madgarden> variable: only parses.
19:46:54 <slava> parsing words really add a whole new dimension to forth
19:47:03 <madgarden> Yes, and they can make the code confusing.
19:47:16 <madgarden> I think, at least, all parsing words should have a colon appended to them.
19:47:25 <madgarden> 10 constant: foo
19:47:33 <madgarden> This would appeal to many non Forthers I think.
19:47:47 <madgarden> It's obvious that the text that follows is associated with constant:
19:47:49 <Herkamire> yeah, I think parsing words make forth way cooler, and more confusing
19:47:57 <Herkamire> my forth doesn't have parsing words. it's more like colorforth
19:48:27 <madgarden> RPN is easy to learn and understand. It's the random mixing of RPN and parsing that messes people up.
19:48:31 <Herkamire> I also like the multiple entry points thing, but unfortunately, it doesn't work on PPC
19:48:46 <imaginator> Herkamire: why?
19:48:49 <Herkamire> madgarden: yeah, people regularly come in here confused about parsing words
19:48:55 <Herkamire> imaginator: PPC does not have a CALL instruction
19:49:26 <madgarden> That multiple entry points is a cool idea. I've made note of that for later. :)
19:49:26 <slava> i don't have parsing words
19:49:26 <imaginator> Why can't you just jmp to some memory?
19:49:39 * imaginator is probably going to port some assembly code to PPC, so ...
19:50:04 <imaginator> I guess I don't see why you would need a call.
19:50:16 <Herkamire> imaginator: you can jump, but you (in most cases) need to push the return stack when you enter a word
19:50:21 <madgarden> : for single token parsing, and " ..." or ( ...) for non-whitespaced strings.
19:51:05 <Herkamire> on a PPC native forth you put branch to the begining of a definition. At the begining of definitions you compile some instruction to push the return stack
19:51:15 <slava> i had a syntax idea
19:51:17 <madgarden> I think that Forth is held back by its ancient "syntax". The old standards damn it somewhat. It could use a makeover in that respect.
19:51:19 <slava> : { >r ;
19:51:21 <slava> : } r> ;
19:51:46 <slava> the syntax is fine, but parsing words can be confusing.
19:51:54 <imaginator> madgarden: I get the impression that Moore thinks the same.
19:52:00 <madgarden> Yep. 
19:52:03 <Herkamire> slava: that's part of the syntax
19:52:09 <Herkamire> me too
19:52:17 <Herkamire> colorforth doesn't have parsing words right?
19:52:21 <madgarden> I use { and } for vocabulary stuff sometimes.
19:52:27 <Herkamire> or most of the things I dislike about ans forth
19:52:31 <slava> everybody seems to have their own concept of how to improve forth, hence we have n implementations in #forth alone
19:52:42 <madgarden> namespace blah
19:52:43 <madgarden> blah{ ... }
19:52:57 <slava> madgarden, how does that work?
19:53:00 <madgarden> (yes, blah has a { appended to the name when created)
19:53:10 <madgarden> Well, it's just a vocabulary stack.
19:53:14 <slava> blah{ pushes a dynamic extent scope that is popped when } appears?
19:53:15 <slava> interesting
19:53:21 <madgarden> blah{ pushes the vocab, } pops it.
19:53:26 <madgarden> So I can do stuff like this:
19:53:27 <slava> i have something very similar!
19:53:55 <madgarden> foo bar math{ + * }
19:54:12 <madgarden> And you'd also define "methods" for a vocabulary this way...
19:54:17 <madgarden> namespace math
19:54:17 <slava> <namespace> @bits-and-pieces
19:54:27 <madgarden> math{
19:54:27 <madgarden> : + ... ;
19:54:27 <madgarden> : * ... ;
19:54:27 <madgarden> }
19:54:36 <slava> $bits-and-pieces [ "value of x is " print $x print ] bind
19:54:46 <slava> "value of global x is " print $x print
19:54:57 <slava> yours allows word definitions in the namespace?
19:55:02 <Herkamire> I don't do anything with namespaces at this point. I just do prefixes
19:55:05 <slava> i have 2 different namespaces
19:55:09 <slava> 1 global for words
19:55:09 <madgarden> Yea. That's standard Forth, really.
19:55:19 <slava> and arbitrary many ones for variables
19:55:21 <madgarden> My namespaces are dead simple.
19:55:42 <slava> the word namespace is really just symbols; a symbol is an internalized name, that maybe holds a pointer to a word definition.
19:56:19 <slava> madgarden, do you implement OO in this way?
19:56:22 <madgarden> Each word entry has a field for which vocabulary it belongs to. When searching the dictionary using the vocabulary stack, this field is matched to selected vocabs.
19:56:26 <slava> madgarden, where each instance has its own namespace?
19:56:39 <madgarden> slava, sure it could be used that way.
19:56:47 <slava> how would the code look?
19:56:53 <madgarden> You'd have to pass around a "this"
19:57:00 <slava> an XT?
19:57:03 <slava> for a blah{ word?
19:57:29 <madgarden> I mean, if it was to be used as OO methods. They'd need something to operate on, an object... an XT or address, etc.
19:58:20 <imaginator> I've been thinking of something like: Object obj  obj : incr 1 + .  obj set 'var' 123
19:58:30 <madgarden> Objects data would be seperate from the methods... like modules.
19:58:32 <imaginator> So, obj would have a builtin : compile symbol.
19:59:04 <madgarden> imaginator, you'd be tying data and words together in an object?
19:59:09 <imaginator> yes
19:59:53 <madgarden> Each object would have a field pointing to a seperate dictionary then, for method lookup?
19:59:57 <imaginator> The idea behind my OO design is that having an object is almost like having a separate namespace within the global.
20:00:09 <madgarden> Right.
20:00:24 --- quit: _proteus (No route to host)
20:00:37 <imaginator> I haven't decided whether to use a separate dictionary or just concat methods and variables to the object name.
20:00:56 <madgarden> That's basically what I was getting at with my namespace. Such namespaces made immediate.
20:01:30 <madgarden> Do the methods need to be tied to the data? Seems to be more complicated than is necessary.
20:04:19 <imaginator> Maybe it is more complicated.
20:04:56 <madgarden> The object would have to access its own dictionary at compile time as well.
20:05:29 <madgarden> So, all objects would have to have compile time and runtime semantics.
20:05:35 <imaginator> It would be costly.  This is part of why I haven't done it yet.
20:05:37 <madgarden> (I think, if I understand you correctly)
20:06:25 <imaginator> Well, code is inlined usually, and it's all relocatable, so it would probably be compile time.
20:07:08 <imaginator> Maybe I'm not seeing the runtime specific semantics this would impose.
20:08:35 <imaginator> Fun, my mom is editing grades over the Internet, and the dialogs popup "Javascript error ... Would you like to debug?"  She says yes, because she wants the bugs gone.
20:08:48 <madgarden> Hehe.
20:09:43 <madgarden> Oooh...
20:09:51 <madgarden> Tell me what this syntax means:
20:09:54 <madgarden> foo. bar.
20:09:54 <madgarden> 10 constant: fred
20:09:54 <madgarden> variable: bob 10 baz bar.
20:11:22 <imaginator> invoke foo, invoke bar, define fred as a constant of 10, define a variable bob that contains 10 baz and invokes bar. ?
20:11:42 <madgarden> Not quite.
20:11:48 <madgarden> constand and variable behave normally.
20:11:56 <madgarden> constant
20:12:06 <madgarden> : indicates parsing ahead... constant: fred
20:12:17 <madgarden> . indicates a word with no return values. foo.
20:12:37 <madgarden> So, the flow of the data through the program is rather obvious.
20:12:43 <madgarden> Compare...
20:12:46 <madgarden> foo. bar. 10 constant: fred variable: bob 10 baz bar.
20:12:54 <madgarden> foo bar 10 constant fred variable bob 10 baz bar
20:13:21 --- quit: qFox ("if at first you dont succeed, quit again")
20:14:17 <madgarden> If Forth "words" are supposed to make readable "sentences" then this kind of syntax makes them more sentence-like.
20:14:39 <imaginator> In my Forth variables are used via set and get.
20:15:01 <imaginator> 10 'a' set 20 'b' set 
20:15:02 <madgarden> So get and set instead of @ and !
20:15:14 <madgarden> 10 a ! 20 b !
20:15:17 <madgarden> What's the difference?
20:15:23 <imaginator> then: 'a' get emit.char  (prints 10)
20:15:42 <imaginator> Well, my get and set expect a string on the stack.
20:15:43 <madgarden> a @ emit.char
20:15:45 <slava> change : to begin-colon-definition :)
20:15:56 <madgarden> slave, how about word:
20:16:11 <madgarden> imaginator, ahh Ok. So hash-tabled variables, like Lua?
20:16:36 <imaginator> yes, but eventually they will use a 4-bit association engine which my machine code prototype uses.
20:16:41 <imaginator> That's more like a trie from what I've been told.
20:16:56 <madgarden> trie?
20:16:59 <imaginator> Then set.pointer and get.pointer will be used for building structures.
20:17:08 <arke> any MIPS people here?
20:17:17 <madgarden> Men In Pink Suits?
20:17:27 <arke> no :)
20:17:29 <arke> MIPS
20:17:30 <madgarden> :P
20:17:36 <arke> the absolutely sweetest processor ever
20:17:41 <imaginator> madgarden: http://www.xmission.com/~georgeps/var/trie_playground.c
20:17:44 <madgarden> Not I.
20:18:27 <madgarden> imaginator, want to give me a quite overview?
20:18:46 <imaginator> It uses the values of the chars as paths.
20:18:49 <imaginator> Rather than hashing.
20:18:58 <imaginator> No collisions can occur, and it's faster than a b-tree.
20:19:07 <madgarden> quick=quite. I am all typoey tonight.
20:19:16 <madgarden> What if two strings are the same?
20:19:24 <imaginator> Then they map to the same node.
20:19:29 <imaginator>  struct trie_s *ar[16];
20:19:29 <imaginator>  int val;
20:19:29 <imaginator> } Trie;
20:19:51 <imaginator> 16 is pow(2,4)
20:20:03 <imaginator> So, say we have the string ABC 
20:21:05 <imaginator> A = 65  we take the lower bits by doing 'A' & 15
20:21:19 <imaginator> That gives us 1
20:21:36 <imaginator> So, we use ar['A' & 15] and change our node pointer to match.
20:21:55 <madgarden> Ahh.
20:21:57 <imaginator> Oh, and we also allocate another node if it points to NULL
20:22:27 <imaginator> The higher bits we get with ar['A' >> 4'] and then do the same.
20:22:50 <imaginator> It ends up making a map through the memory pointers to the very end, and then we fill in val (which in my latest is a void *).
20:23:12 <madgarden> Interesting.
20:23:34 <madgarden> I might have to play with that.
20:23:45 <imaginator> :)
20:24:40 <imaginator> My first implementation I called a "Pig structure" because the pointer array for each node was 256 or pow(2,8) (for 8 bits).
20:24:47 <madgarden> Heh.
20:25:09 <imaginator> With the 4-bit version more nodes are shared, and fewer overall pointers are created.
20:25:36 <madgarden> I'd maybe even mess with a 2-bit version.
20:25:41 <imaginator> The cost is a little more computing for ar['A' & 15] and ar['A' >> 4] vs. ar['A']
20:27:39 <imaginator> 2 bits is closer to what some people do with tries.
20:27:48 <madgarden> I wonder how this compares overall to a standard hash table.
20:28:08 * madgarden is not too knowledgable in the data structures stuff.
20:28:10 <imaginator> Well, with hashing you may never have a perfect hash (at least not with good speed).
20:28:24 <imaginator> So, collisions occur, and then you must allocate ->next nodes.
20:28:34 <imaginator> or store in near bucket.
20:28:46 <madgarden> Right. Hmm.
20:28:52 <madgarden> How about memory usage?
20:28:57 <madgarden> trie seems a bit more hefty.
20:29:17 <imaginator> Ficl and Tcl both have algorithms that analyze their hash tables and recreated if it would improve performance (collided buckets).
20:29:35 <imaginator> Well, it uses more memory, but I don't mind :)
20:30:10 <madgarden> How much more though? I'd like to see a graph or something. :P
20:30:47 <imaginator> It depends on the number of values stored/associated.
20:31:44 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
20:32:56 <imaginator> And whether more paths are reused.
20:33:07 <imaginator> 68 bytes for a system with 4 byte pointers would be used per node.
20:33:15 <imaginator> 408 bytes for "ABC" and then another 68 for the root.
20:33:27 <imaginator> A MB is pow(2,20 = 1048576
20:34:44 <madgarden> So around how many entries would that be?
20:35:40 <imaginator> entries?
20:37:26 <madgarden> keys
20:38:03 <imaginator> "ABC" is 6 path nodes.  In a hash table it would be one key.
20:38:50 <madgarden> Well what do you call the "ABC" bit then? :)
20:39:06 <madgarden> The string?
20:39:27 <madgarden> So how many strings (guesstimate) would fit into your 1M.
20:39:43 <imaginator> It would be 1 string.
20:39:55 <madgarden> 1MB for 1 string?
20:39:56 <imaginator> oh, I don't know how many would fit in 1M
20:40:06 <madgarden> Oh.
20:40:10 <imaginator> no, 476 for that "ABC" string.
20:40:29 <imaginator> But some strings may follow the same path as the A or B or C.
20:40:42 <madgarden> Right.
20:41:06 * imaginator wonders how he could accurately measure the number that would fit in 1 MB
20:41:09 <madgarden> So if you wanted to get the most bang for your buck, you'd use variables like AAA, AAB, AAC, etc. :P
20:41:24 <imaginator> yes
20:41:37 <imaginator> and making them all upper/lower case would help too.
20:41:47 <madgarden> I'd probably do a random string generation test.
20:42:07 <imaginator> Well, when/if you do I'd like to see the result :)
20:42:16 <madgarden> Multiple passes, see how many each pass takes to fill up 1M of memory, then average the results out.
20:42:38 <madgarden> Heh, it's your damn trie! ;P
20:42:46 <madgarden> I like the idea though, if it's not too bloaty.
20:43:13 <imaginator> ;P
20:43:15 <madgarden> Like say, 250K for 1000 strings wouldn't be too shabby.
20:43:31 <madgarden> So what did you think of my super duper Forth syntax idea? Seems so obvious to me.
20:44:19 <imaginator> Well, I mostly work with LISP-like languages and C, so Forth is somewhat foreign to me even though I'm building my own.  It isn't so obvious to me.
20:44:54 <imaginator> But a programmer can learn to use almost anything.
20:44:59 <imaginator> :)
20:45:01 <madgarden> Yep. 
20:45:19 <madgarden> Forth just lacks context in its code. That's the main problem I see.
20:45:42 <I440r> context ?
20:45:43 <imaginator> That's why I have set and get instead of @ fetch and ! for store.
20:45:46 <madgarden> If a word isn't well described, then the program can be hard to decipher.
20:46:08 <I440r> @ and ! are perfect
20:46:13 <madgarden> Well yea, that's an example, though at least @ and ! are quickly learned, unlike new word definitions.
20:46:36 <imaginator> I440r: %%#*##&**#!@
20:46:45 <imaginator> ;P
20:46:45 <madgarden> I440r, did you see my syntax ponderings earlier?
20:47:40 <imaginator> I can read ! and @, but I like set and get.  I also understand the view that "If you use them often, you will remember."
20:58:44 <I440r> er no
20:58:50 <I440r> i was sorta elsewhere lol
20:59:20 <madgarden> OK, I'm gone, but take a peek real quick:
20:59:24 <madgarden> ( Simple but descriptive syntax for parsing words and no stack-result words ) 
20:59:25 <madgarden> foo. bar.
20:59:25 <madgarden> 10 constant: fred
20:59:25 <madgarden> variable: bob 10 baz bar.
21:01:28 <arke> I440r: best processor in the world, in your opinion?
21:01:37 <I440r> hmmm
21:01:39 <I440r> 8051
21:01:40 <I440r> :)
21:01:46 <arke> ack :)
21:01:50 <arke> try again
21:04:24 <I440r> hmmm
21:04:30 <I440r> nope
21:04:32 <I440r> 8051 
21:04:34 <I440r> definatly :)
21:05:30 <arke> naaaaaaah
21:05:31 <arke> not even close
21:05:34 <arke> MIPS my friend :)
21:11:50 --- nick: hefner -> hefzzz
21:17:35 --- quit: proteusguy (orwell.freenode.net irc.freenode.net)
21:17:35 --- quit: ayrnieu (orwell.freenode.net irc.freenode.net)
21:17:35 --- quit: warpzero (orwell.freenode.net irc.freenode.net)
21:17:35 --- quit: ChanServ (orwell.freenode.net irc.freenode.net)
21:17:35 --- quit: arke (orwell.freenode.net irc.freenode.net)
21:17:35 --- quit: chandler (orwell.freenode.net irc.freenode.net)
21:17:35 --- quit: l0C0 (orwell.freenode.net irc.freenode.net)
21:18:06 --- join: hovil (~hovil@CommSecureAustPtyLtd.sb1.optus.net.au) joined #forth
21:18:32 --- join: ChanServ (ChanServ@services.) joined #forth
21:18:32 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
21:18:32 --- join: ayrnieu (julian@65.169.246.30) joined #forth
21:18:32 --- join: arke (~Chris@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth
21:18:32 --- join: chandler (~chandler@chandler.registered.freenode) joined #forth
21:18:32 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth
21:18:32 --- join: l0C0 (loco@bor.gs) joined #forth
21:18:32 --- mode: irc.freenode.net set +o ChanServ
21:31:22 --- join: ASau (~asau@158.250.48.196) joined #forth
21:31:30 <ASau> Dobroe utro!
21:38:10 <arke> dobryjj vecher
21:50:02 --- quit: arke ("Leaving")
21:56:38 --- join: Serg (Serg_Pengu@212.34.52.140) joined #forth
22:01:56 <ASau> Privet, Serg
22:02:07 <Serg> hi
22:02:25 <ASau> Kak dela?
22:03:01 <Serg> i bought a CD of MP3 proggies to convert my files into lower bitrate for player...
22:03:08 <Serg> none did it well, dhamn !
22:03:30 <ASau> Hm.
22:03:35 <Serg> so i had to dig for my old CD-R backup and find Winamp w/ MP3-write plugin ;))
22:03:40 <ASau> There should exist free ones.
22:03:58 <ASau> Maybe, sox.
22:03:59 <Serg> it did, perfectly !
22:04:19 <Serg> so now i have all the Scorpions in my 128M NEXX-3 player ;))
22:04:32 <Serg> 32kB/s
22:05:07 <Serg> yesterday i rised from dead my inet gateway at work
22:05:35 <ASau> NAT?
22:05:37 <Serg> it does gatind ;)) but Mdaemon glitches slightly
22:06:01 <Serg> no, WinGate+Mdaemon+Sygate personal firewall
22:06:11 <ASau> OK.
22:06:25 <ASau> Fine.
22:06:42 <ASau> No, I think GPRS is not needed.
22:06:52 <ASau> Now,..
22:06:53 * Serg will like OpenBSD more, but it has to provide networked HASP key
22:07:12 <Serg> GPRS + dialup at home, XDSL at work ;)
22:07:27 <ASau> Ah...
22:08:23 <ASau> A ja sobralsja v aspiranturu.
22:08:33 <Serg> damn XP does not cope w/ my Siemens M50, saying 'DTR low'
22:08:42 <Serg> and 'CTR low'
22:09:05 <Serg> hardware hackers say i need to short some pins in COM port connector
22:09:35 <ASau> Maybe you've low power?
22:09:44 <ASau> That's strange.
22:10:04 <Serg> ASau: education sucks at this time in this cuntry ;((( did it help many losers w/ 'engineers' diplomes ?
22:10:17 <ASau> I'm not engineer.
22:11:27 <ASau> And organic chemistry research is payed good enough.
22:11:48 <Serg> 'business' (at margin of crime) rules, EDU sucks by our nowdays life
22:12:43 <ASau> It depends on your will.
22:13:19 <ASau> If you want risky life, you go to "business."
22:13:57 <ASau> If you want another life, you go another way.
22:14:04 <Serg> i gonna go "middle way" like buddhists say :
22:14:32 <Serg> invest in being "tuff" and "agile", but...
22:14:56 --- join: kc5tja (~kc5tja@66-91-231-74.san.rr.com) joined #forth
22:15:06 --- mode: ChanServ set +o kc5tja
22:15:23 <Serg> not rip tons of bucks and go to far, restrict self to middle-class income level ;))
22:15:50 <ASau> I don't think you succede.
22:16:08 <ASau> Try.
22:17:08 <kc5tja> ASau: Thank you very much for the information on PL/I.  It was very educational.
22:17:59 <imaginator> Hi kc5tja 
22:18:11 * kc5tja has decided to use Oberon as the basis of the Kestrel's alternative programming language instead of Forth.  The implementation difficulties for the two appear to be about the same, and a lot more people who are familiar with home computing are more familiar with Pascal than with PL/I or PL/M.
22:18:41 <kc5tja> imaginator: Hello.
22:18:43 <ASau> There's an educational dialect, so called XPL.
22:18:55 <ASau> It was published in book.
22:19:06 <ASau> Sorry, I can't recall authors.
22:19:16 <imaginator> kc5tja: does Oberon have interactive compilation?  How dynamic of a language is it compared to Forth?
22:19:35 <ASau> imaginator, it's classic compiled p.l.
22:19:36 <kc5tja> ASau: Yeah, I haven't researched that yet.  But Oberon isn't too hard to implement.  It and PL/M have *very* similar overall structure, so considering how similar the two are, I had to go with Oberon.
22:20:17 <kc5tja> imaginator: Oberon, the language, is like C in that compilation is not interactive.  However, its modularity provides vastly more interactivity during development than C does.
22:20:31 <kc5tja> But it still doesn't compare with Forth for interactivity.
22:21:33 <kc5tja> I'm calling my Oberon dialect Charon though, which gives me the flexibility to evolve the language independently of Oberon proper.
22:23:07 --- quit: Sonarman ("bye")
22:25:11 <ASau> kc5tja, try looking at Algol-68.
22:25:57 <ASau> It is more "prefix".
22:26:27 <ASau> You know type before you create variable.
22:26:51 <ASau> When reading left to right and not looking ahead.
22:27:32 <ASau> Of course, it's huge enough, but you don't need everything.
22:27:49 <kc5tja> Well, same thing in Oberon; variables must be declared before used.
22:28:00 <ASau> No.
22:28:44 <ASau> I mean this:  [1:10] INT A
22:29:06 <kc5tja> Ahh, so that would define an array then?
22:29:07 <ASau> vs.: array(1..10) of int.
22:29:12 * kc5tja sees.
22:29:42 <kc5tja> Although, I'd look rather silly resurrecting Algol for this computer.
22:29:44 <kc5tja> :)
22:29:53 <ASau> And Algol-68 is more close to Forth nature.
22:29:55 <kc5tja> PL/* is awkward enough for most people, it seems.  :)
22:30:30 <kc5tja> Well, remember this alternative language environment is for people who are not interested or who actively eschew Forth (as a few gave ample hints towards me about).
22:30:36 <ASau> Every BEGIN ... END has (may have) return value.
22:32:20 <ASau> Then why not BASIC?
22:34:39 <kc5tja> BASIC will be offered as well.  However, BASIC is interpreted.
22:34:46 <kc5tja> It offers no means of producing compiled applications.
22:35:22 <ASau> kc5tja, BASIC is FORTRAN in main part.
22:35:45 <kc5tja> Yes, I know this.
22:35:46 <ASau> You can eliminate parts which require interpreting.
22:36:06 <ASau> And you'll get FORTRANoid p.l.
22:36:10 <ASau> Compiled.
22:37:05 <kc5tja> I'd rather offer a "real" programming language than stick with compiled BASIC though.
22:37:57 <kc5tja> Compiled BASIC still needs its run-time environment to be carried along too, as BASIC is garbage collected.
22:38:24 <ASau> Make it more FORTRAN.
22:38:30 <ASau> Use static allocation.
22:39:01 <ASau> But, I don't see why garbage collection.
22:39:26 <ASau> I can't recall one existed.
22:40:23 <ASau> I recall only "Out of memory" error message.
22:40:39 * imaginator doesn't remember BASIC being garbage collected
22:40:56 <ASau> Arrays are allocated explicitly and freed the same way.
22:41:11 * imaginator remembers accidental loop (memory exceeded) lockup
22:41:36 <ASau> Strings are allocated implicitly but they can't be longer than 256 bytes.
22:42:27 <kc5tja> BASIC is very garbage collected.
22:42:55 <kc5tja> You've never had a BASIC system just halt dead while running some compute-intensive application (especially one that involved strings), to pick up later again a few seconds later?
22:43:50 * kc5tja has observed this behavior with TRS-80 BASIC and Commodore 64 BASIC.  Commodore 128 had a vastly superior garbage collector (it was actually concurrent!), so it never really halted the system.
22:44:38 <kc5tja> Also, BASIC wouldn't be BASIC if it were more Fortran and more static.
22:44:49 <kc5tja> The whole point of BASIC was to move away from Fortran's static declarations.
22:57:41 --- quit: Herkamire ("Good work. Sleep well. I'll most likely kill you in the morning.")
22:59:49 <ASau> I don't think you need GC.
23:00:02 <ASau> Memory allocation is just semi-automatic.
23:00:45 <ASau> You create variables on request, but never free memory till special request.
23:04:37 <kc5tja> ASau: Variable space is never freed; only compacted.
23:04:44 <kc5tja> And only when it runs out of space.
23:05:03 <kc5tja> Once a variable instance is created, it's there forever.
23:05:08 <kc5tja> (until you enter NEW)
23:05:27 <kc5tja> This is, at the least, the behavior of C64 BASIC and C128 BASIC.
23:05:32 <kc5tja> So, suppose I have:
23:05:38 <kc5tja> A$="Hello"
23:05:48 <kc5tja> That creates a new variable descriptor and string containing Hello.
23:06:13 <kc5tja> Then, if I say, A$=A$+" World!", it will create a *new* A$ descriptor in memory, leaving the original present in memory.
23:06:33 <kc5tja> But becase the latter A$ is more recent, any reference to A$ will result in the "Hello World!" string, not the original.
23:07:31 <Serg> till it run out of mem ?
23:07:38 <kc5tja> When garbage collection occurs, variables with the same name are compacted over by memmove()ing (basically) the most recent portions of the descriptor data over the old data.
23:07:58 <kc5tja> Since this is an O(n^2) way of doing things, it can take up to several minutes on machines like the Commodore 64.
23:08:21 <kc5tja> Serg: Generally, yes, it will *completely* exhaust memory before invoking the compactor.
23:09:00 <Serg> heh, why not just allocate N bytes for every string var and not fsck brains ?
23:09:08 <Serg> or implement ring buffer
23:09:08 <kc5tja> Commodore 128 BASIC is a notable exception: it was widely hailed as a major advancement in 8-bit BASICs because it had a concurrent compactor(!!).  It compacted one variable evern couple of VSYNC interrupts.
23:09:25 <kc5tja> Serg: Because it's not right.
23:09:51 <kc5tja> Serg: If I set A$="Serg", and later on I find that A$ suddenly contains "Foo" without my changing it explicitly, well, that can have major repercussions.
23:10:09 <Serg> we waste ram, if we gone to end we circle to beginning....
23:10:23 <kc5tja> Serg: That's not correct behavior.
23:10:30 <kc5tja> And is highly undesirable.
23:10:39 <imaginator> It's what MachineForth does.
23:10:41 <Serg> if we run into var still used, we 'compact' it
23:10:51 <kc5tja> imaginator: No it isn't.
23:10:54 <imaginator> The registers wrap around.
23:11:14 <kc5tja> imaginator: The *REGISTERS* wrap around (and even then, *ONLY* on the P21 and P18 series CPUs).
23:11:27 <kc5tja> imaginator: Strings are not registers, and registers are not string variables.
23:11:39 <Serg> see,  let's '.' is waste var, and 'u' is good var:
23:11:51 <Serg> ...u...u....u..n
23:11:54 <imaginator> :) ok 
23:12:11 <Serg> at N, we need to insert but we hit end of ram, so we do:
23:12:15 <kc5tja> But the concept applies to all data types, not just strings.
23:12:25 <Serg> N..u...u....u...
23:13:12 <Serg> next, let's say we need to allocate _really_long string of 5 dots, but we have only 2 be4 next used var, see ?
23:13:22 <Serg> so ve compact _one_ used var:
23:13:35 <Serg> Nu.....u....u...
23:13:50 <Serg> voila, we made room for new long string !
23:14:19 <Serg> NuNNNNNu....u...
23:14:21 <Serg> got idea ?
23:15:09 <Serg> and no garbage collection ;))
23:16:17 <hovil> http://www.kuro5hin.org/story/2004/3/17/93442/8657
23:16:21 <kc5tja> What are you talking about, no garbage collection?
23:16:27 <kc5tja> What do you think garbage collection *IS*?
23:16:32 <kc5tja> That is just incremental compaction!
23:17:06 <kc5tja> however, I do like it; it's much faster than what the Commodore 64 used.
23:17:25 <kc5tja> But to say it's not garbage collection would be lying.
23:18:24 <Serg> kc5tja: heh, do you like my algorithm ?
23:18:58 <ASau> Serg, how do you suggest to fight against fragmentation?
23:19:18 <Serg> as i shown - defrag on demand
23:19:29 <ASau> Well.
23:19:40 <Serg> if not enough room to place next var, defrag to make room
23:19:48 <ASau> Compaction is easier than reallocation.
23:19:59 <Serg> ?
23:20:33 <ASau> To allocate memory you only increase heap pointer.
23:21:25 <hovil> can't wait to get home
23:21:28 <ASau> When you reallocate memory, you need find free memory range and mark it.
23:21:43 <hovil> the monitor for my a2000 arrived today
23:23:33 <hovil> Wikibooks is dedicated to developing and disseminating free, open content textbooks and other classroom texts. 
23:23:43 <hovil> If only educational institutions would take them up on that.
23:30:06 --- quit: Serg ()
23:55:24 <kc5tja> Sorry, was helping out with household chores
23:59:59 --- log: ended forth/04.03.22