00:00:00 --- log: started forth/04.04.08
00:45:11 <Serg> how else one may spam bad guys ?
00:45:20 <Serg> sms ? fax ?
00:48:34 <yeoh> snail mail?
00:49:31 <yeoh> why not let the spammers spam each other? get all their names, addresses and e-mails and circulate them to all othe spammers.
00:53:43 --- quit: Herkamire (Read error: 60 (Operation timed out))
00:58:59 --- quit: Serg ()
01:56:14 <yeoh> bye for now. See you all later. 
01:56:22 --- part: yeoh left #forth
03:08:53 --- join: yeoh (~yeoh@219.95.6.105) joined #forth
03:41:43 --- quit: yeoh ("Client exited")
04:14:18 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth
04:43:56 --- join: far- (~far@tony11-1-228.inter.net.il) joined #forth
04:44:01 <far-> hello
04:44:07 <far-> forth is a low level programming language?
04:50:14 <Robert> Kind of.
04:54:50 <far-> Forth binary files are smaller than Assembly binary files?
04:56:27 <far-> ?
04:58:40 <Robert> Depends. Usually not.
04:59:00 <far-> depends on what?
04:59:08 <Robert> How you write the program...
04:59:21 <Robert> But if you want speed and small size, use assembly language.
04:59:39 <far-> i dont want speed, only small size
04:59:59 <far-> are you sure there isnt another language that its binaries are smaller than asm binaries?
05:00:24 <Robert> Some sort of self-decompressing program...
05:01:05 --- join: segher (~segher@blueice3n1.de.ibm.com) joined #forth
05:01:07 <Robert> You _can_ make small Forth binaries if that's your design goal.
05:01:19 <Robert> But if the program is small, I wouldn't bother.
05:01:29 <segher> it's hard to make huge binaries, even
05:01:35 <far-> smaller than the eqvivalent in asm?
05:03:12 <Robert> I guess that's not impossible, in some cases.
05:03:14 <Robert> Try
05:03:41 <far-> so the smaller size i can ever get will be in asm program?
05:03:53 <far-> are you sure?
05:04:10 <Robert> No.
05:04:21 <far-> ?
05:04:26 <Robert> Why do you need the smallest possible program anyway?
05:04:40 <far-> for research
05:04:55 <segher> of course.  in the worst case, you can write a custom forth inner interpreter in asm.  but that is a lot more work...  the interesting question is, how do you make a small program with *reasonable* effort.
05:05:29 <far-> i dont care to write 10000 lines for a program that i can write 10 lines in c
05:05:46 <far-> the compiled file-size is important to me
05:06:08 <far-> so you know what is the programming language im looking for?
05:06:42 <segher> asm.
05:07:03 <segher> anything you can do in any other language, you can by definition also do in machine code.
05:07:39 <far-> ok and is there a way to compress a asm binary data inside the memory? so it will take less space?
05:07:58 <Robert> Sure.
05:08:18 <Robert> Just compress the program, and have a small decompressing routine in the beginning.
05:08:28 <far-> nice
05:08:35 <Robert> But you'll have to have rather large programs for that to pay off
05:08:50 <far-> the compressing concept is only for asm or for anything in memory?
05:09:00 <far-> my programs are small- 100bytes
05:09:01 <Robert> Anything.
05:09:17 <Robert> 100 bytes? Then I'd go with an assembly language program.
05:09:48 <Robert> If you spend a number of hours, you can usually chop off a number of bytes 
05:10:23 <far-> by better coding?
05:10:38 <segher> by evil and/or clever tricks
05:11:03 <far-> you said i should use asm. with or without the compress? if its 100bytes
05:12:11 <segher> whatever is smallest
05:12:19 <segher> what's your target cpu?
05:12:54 <far-> x86
05:13:01 <far-> intel
05:13:25 <segher> a small decompressor might help, then
05:14:00 <far-> you know where i can read about memory data compressing/decompressing ?
05:16:22 <segher> get a good book, i'd say.  and/or study other people's code.  no, there is no easy or fast way to learn stuff like this.
05:19:34 <far-> ok thanks
05:20:50 --- quit: far- ()
05:22:25 --- quit: OrngeTide (Remote closed the connection)
05:27:48 --- join: OrngeTide (orange@rm-f.net) joined #forth
05:52:42 --- join: yeoh (~Yeoh@219.95.8.170) joined #forth
06:15:55 --- join: Downer (Downer@82.70.9.103) joined #forth
06:17:43 --- part: Downer left #forth
06:23:19 --- join: networm (~networm@L0658P25.dipool.highway.telekom.at) joined #forth
06:23:23 <networm> hello
06:23:46 <yeoh> hello networm.
06:28:11 --- quit: yeoh (Remote closed the connection)
06:34:23 <madwork> hi!
06:34:42 <Robert>  Hi
06:56:44 <madwork> So what the forth is going on today?
06:59:30 --- join: yeoh (~yeoh@219.95.8.170) joined #forth
07:12:07 --- join: Serg (~z@212.34.52.140) joined #forth
07:14:07 --- join: proteusguy (~proteusgu@workstations.pinnaclesports.com) joined #forth
07:22:38 --- quit: Serg ()
07:29:29 --- quit: yeoh ("Leaving")
07:43:08 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth
08:14:22 --- join: OrngeTid1 (orange@rm-f.net) joined #forth
08:14:22 --- quit: OrngeTide (Read error: 54 (Connection reset by peer))
08:48:10 --- join: Frek (~anvil@h33n2fls31o815.telia.com) joined #forth
08:48:52 --- quit: Frek (Client Quit)
08:49:28 --- quit: tathi ("leaving")
08:50:19 --- quit: networm (Read error: 113 (No route to host))
08:51:10 --- join: networm (~networm@L0632P25.dipool.highway.telekom.at) joined #forth
09:27:53 --- join: OrngeTide (orange@rm-f.net) joined #forth
09:32:09 --- quit: OrngeTid1 (Read error: 54 (Connection reset by peer))
10:14:26 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth
10:19:14 --- join: Herkamire (stjohns@24.91.225.238) joined #forth
10:34:05 --- join: thin (thin@bespin.org) joined #forth
10:43:32 --- join: kc5tja (~kc5tja@66-91-231-74.san.rr.com) joined #forth
10:43:39 --- mode: ChanServ set +o kc5tja
10:44:54 <kc5tja> Well, I have received a couple of datasheet views for the FTS1001 design, but nobody has responded, save one, who was apparently looking for a couple of chips to implement the asynchronous bus I had come up with.
10:45:47 <kc5tja> So now that my learning experience with the FTS1001 is finished (insofar as learning Verilog and how to use it to produce a working behavioral model), I suppose now I can work on some more sophisticated projects.
10:46:11 <kc5tja> At this time, the FTS1001 is likely to not make it to production.
10:46:25 <kc5tja> (Not as a discrete component, at least)
11:30:43 --- quit: tathi ("leaving")
12:16:33 --- join: wossname (wossname@HSE-MTL-ppp60965.qc.sympatico.ca) joined #forth
12:32:14 --- quit: segher ("Leaving")
13:44:49 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth
13:44:50 <slava> hey
13:58:12 <ianp> re sl4v4
13:59:12 <kc5tja> Hello.
13:59:18 <ianp> greets
14:00:22 <slava> ianp, what's with the 31337sp34k?
14:02:31 <ianp> slava: n07h1|\|
14:13:20 --- quit: wossname ("njoy")
15:01:06 --- join: Sonarman (1000@adsl-64-160-164-6.dsl.snfc21.pacbell.net) joined #forth
15:42:34 <Robert> Ah, _finally_ a QSO! And then I had to reply to a CQ DX from a neighbour country. ;)
16:01:18 --- quit: proteusguy ("Leaving")
16:25:11 <kc5tja> Band conditions are bad today?
16:30:44 <Robert> Not really. People just seem to avoid me.
16:31:37 <Robert> It's funny, this guy had the same rig as I have, IC730. His was on max power (100W), mine on min power (5W). I got a 599, he got a 579.
16:41:48 <kc5tja> Heh
16:42:10 <kc5tja> Well, I decided to start working on the Kestrel's custom bus interface chip design today.
16:42:56 <Robert> OK. :)
16:43:02 <Robert> Good night
16:43:11 <kc5tja> The UART didn't generate all that much interest.  I got more interest in my asynchronous bus design (which can be said to be equivalent to the Wishbone bus).
16:43:20 <kc5tja> 73 Robert 
16:43:22 --- quit: networm (Read error: 60 (Operation timed out))
16:50:25 <Robert> e e
16:51:42 <qFox> kc5tja> this might be a very silly question, but how do i call an XT while code is running?
16:51:57 <qFox> in that p24.... (althoug i guess it'll go for any cpu)
16:52:08 <qFox> i dont know how to call an address on the stack
16:52:08 <Sonarman> the std forth word is EXECUTE
16:52:17 <qFox> yea i know, but in machine forth
16:52:34 <qFox> i cant access the I and P registers directly with instructions
16:52:40 <kc5tja> Transfer it to the return stack, then "return".
16:52:42 <qFox> so i cant either change or save them
16:52:56 <qFox> kc5tja> yes but how will i return to my code?
16:53:09 <kc5tja> Make it a subroutine.
16:53:13 <kc5tja> : EXECUTE >R ;
16:53:25 <qFox> no i'm in machine language
16:53:28 <qFox> not in forth yet
16:53:32 <kc5tja> That *IS* machine language.  :)
16:53:36 <qFox> no i mean
16:53:42 <qFox> um like
16:53:49 <kc5tja> That is machine-Forth for EXECUTE.
16:54:04 <qFox> but, where will the ret go to?
16:54:07 <kc5tja> When execute is called, the return address for your code is already on the return stack.
16:54:26 <kc5tja> >R places the desired subroutine address on the return stack as well.
16:54:37 <kc5tja> The ; "returns" to the desired subroutine.
16:54:43 <qFox> yes, but what about P? the next instruction
16:54:44 <qFox> oh
16:54:47 <kc5tja> When that subroutine is finished, it'll return to your original code.
16:54:54 <qFox> well thats all nice, but how do i call P
16:55:02 <kc5tja> I JUST TOLD YOU!!
16:55:06 <kc5tja> ' p EXECUTE
16:55:08 <qFox> because i dont believe i can do that legimitly
16:55:18 <kc5tja> I challenge you to prove yourself wrong.
16:55:21 <qFox> kc5tja... there's no instruction to get P
16:55:24 <qFox> in the p24
16:55:29 <qFox> how do i get the register value?
16:55:30 <qFox> :\
16:55:35 <kc5tja> Dude, CALL places the return address on the return stack.
16:55:48 <qFox> oh
16:55:53 * qFox sees light
16:56:00 <kc5tja> How do you think RET works?  :)
16:56:06 <qFox> well lets see
16:56:12 <kc5tja> It's called a "return stack" for a reason.  :)
16:56:16 <qFox> yes i know
16:56:23 <qFox> but i didnt know how to get the P 
16:56:25 <qFox> thats all
16:56:33 <qFox> but didnt think of calling call that would ret
16:56:37 <kc5tja> By P do you mean the contents of the instruction pointer?
16:56:49 <qFox> contents of the next instruction pointer yes
16:56:52 <qFox> but i get it
16:56:55 <kc5tja> Or do you mean the addrses of the procedure?
16:57:10 <qFox> nonono look
16:57:19 <qFox> my eval.word will call (') 
16:57:26 <qFox> (ok just call it tick)
16:57:31 <qFox> tick will return the xt
16:57:43 <kc5tja> Ok
16:57:44 <qFox> then the code will execute that
16:57:49 <qFox> but i just didnt know how to get back to that code
16:57:50 <kc5tja> Ok.
16:57:57 * kc5tja slaps his forhead!
16:57:59 <qFox> but now i do :)
16:58:08 <kc5tja> Good.
16:58:10 <kc5tja> :)
16:58:15 <kc5tja> I was going to cry if you still didn't get it.  :D
16:58:46 <qFox> well i never thought about calling a function that pushed something on the return stack, then ret to it
16:59:12 <qFox> i just couldnt figure it out, i mean i knew i had to push ret, but that would not get me back to my original postition :)
16:59:23 * kc5tja nods
16:59:26 <qFox> ohwell
16:59:27 <qFox> sleep now
16:59:28 <qFox> nite
16:59:33 <qFox> (yes, "early" ;)
16:59:33 <kc5tja> Goodnight.
16:59:38 <kc5tja> Man everyone is going to bed here.  :)
16:59:54 <qFox> follow the lead :p
17:00:03 <kc5tja> It's mid-day here.
17:00:06 <kc5tja> 17:00.
17:00:08 <qFox> big deal
17:00:12 <qFox> its mid-night here :p
17:00:20 <qFox> almost
17:00:32 --- quit: qFox ("if at first you dont succeed, quit again")
17:00:43 <kc5tja> Bahh!
17:00:48 <kc5tja> Midnight SchmidMight!
17:00:58 <kc5tja> Try going to bed at 4AM, only to go back to work at 9AM.  :D
17:08:19 <Sonarman> : ENTER >R ; \ I already have it in my system *** ENTER already defined, cfa=4001810h(67115024)  ok[Dec]
17:08:22 <Sonarman> : 1-20 1 BEGIN DUP 21 < WHILE >R 2R@ SWAP ENTER R> 1+ REPEAT DROP RDROP ; ok[Dec]
17:08:24 <Sonarman> : foo 1-20 . ;
17:08:27 <Sonarman> foo
17:08:27 <Sonarman> 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  ok[Dec]
17:08:31 <Sonarman> that's cool :)
17:10:55 <slava> ?
17:12:18 <Sonarman> look at the definition of foo, and look at what it does
17:12:40 <Sonarman> (this is from comp.lang.forth)
17:13:46 <slava>     1] : foo 20 count . ;
17:13:47 <slava>     2] foo
17:13:47 <slava> [ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ]
17:44:04 --- join: yeoh (~yeoh@219.95.0.191) joined #forth
17:45:38 <Sonarman> hi yeoh
17:45:54 <yeoh> hi Sonarman. :)
18:15:29 <madgarden> Fear not, for I have R> !
18:17:41 <Sonarman> hehe... what?
18:18:50 <Sonarman> have you gone... mad? :) DIE SONARMAN
18:20:54 <kc5tja> OK, off cooking -- back in a bit.
18:24:01 <madgarden> :O
18:24:22 <madgarden> How do you pronounce R> ?
18:24:40 <Sonarman> First syllable: Arr
18:24:45 <Sonarman> Second syllable: >
18:25:20 <Sonarman> ['] CRASH R@ !
18:26:02 <madgarden> It's the closest thing to "returned" that I could think of.
18:27:12 <Sonarman> you could always just say that you have ; :-Q
18:27:34 <madgarden> I have semicolon?
18:27:42 <madgarden> No, I have a full colon, thank you!
18:27:56 <madgarden> Well, it's not full... 
18:28:01 <madgarden> Intact!
18:29:38 <Sonarman> Well then, I think that gives us a reason to PARRTAY! Here, have a party hat.  *twang!*  Haha!  Gotcha! I pulled the elastic string and it hit you!
18:30:09 --- join: ianp` (nobody@c-24-13-109-164.client.comcast.net) joined #forth
18:30:13 <Sonarman> Am I the most cleverisherest prankster ever!
18:31:07 <madgarden> Ow! Just don't hit my colon, dammit!
18:31:23 <madgarden> I imagine that would really smart.
18:32:29 <Sonarman> Not as much as resetting your return stack would!  Let me demonstrate... *ZHING!!!*
18:33:47 * madgarden empties colon on Sonarman.
18:33:50 <madgarden> Now you've done it!
18:34:03 <Sonarman> now... i wish i could figure out what the hell ALITERALs are on bigFORTH
18:34:08 <madgarden> How that "return stack" for you? ;)
18:34:33 <madgarden> Hmm, don't have bigForth here.
18:34:55 <Sonarman> it's fun :)
18:35:13 <madgarden> What's so big and fun about it?
18:36:12 <Sonarman> it's fun because there's so much to discover about it, and it's big because... because Bernd Paysan said so!
18:36:38 <madgarden> Bernd Paysan, that crazy guy!
18:36:55 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth
18:37:12 <madgarden> Does it let you create turnkey apps?
18:37:20 <Sonarman> apparently
18:37:38 <Sonarman> i haven't tried it, but it says so on the website
18:38:15 <madgarden> Hmm, I wonder how it compares to Win32Forth.
18:38:27 <Sonarman> http://www.jwdt.com/~paysan/ <-- hit the fancy-schmancy Enlightenment button labeled bigFORTH there
18:39:01 <tathi> :)
18:40:25 <madgarden> Oooh, it has OpenGL bindings?
18:40:42 <Sonarman> yes, and cool demos!
18:41:26 <madgarden> TEH!!1
18:42:10 <madgarden> Hey, I think I have actually installed this before at some point.
18:42:29 <Sonarman> in my experience, it works better on Linux than on Windows
18:43:21 <Sonarman> either way, it's still pretty quirky
18:44:11 <madgarden> Quirky how?
18:45:22 <Sonarman> subtle bugs, weirdnesses in the editor, strange window focus behavior on X, slight latency between click and response
18:46:38 <Sonarman> and very little documentation and comments :)
18:46:38 <madgarden> Huh. What's the deal when I run it? It drops me into what appears to be the interpreter, but it's a bit goofy.
18:46:55 <Sonarman> you running xbigforth on Linux?
18:47:07 <madgarden> No.
18:47:21 <Sonarman> don't run bigforth.exe, run xbigfalfskaf.asfxexe
18:47:32 --- join: blockhead (default@dialin-150-tnt.nyc.bestweb.net) joined #forth
18:47:37 <Sonarman> sorry, fingers got lazy. you know what I meant :)
18:48:24 <Sonarman> and the neat OpenGL dragon demo doesn't work right on Windows, at least in my experience
18:48:35 <madgarden> Run WHAT?
18:48:50 <Sonarman> xbigforth.exe :)
18:49:00 <madgarden> Don't have that.
18:49:03 <madgarden> There's MINOS though.
18:49:06 <madgarden> And THESEUS.
18:49:11 <madgarden> Whatever that is.
18:49:31 <Sonarman> you sure there's no xbigforth.exe? generic application icon?
18:50:07 <madgarden> Nay.
18:50:14 <madgarden> MINOS seems to work OK though.
18:50:20 <madgarden> Hmm, neat screen viewer.
18:50:41 <Sonarman> cool.  i guess there's been an update since the last time i tried it on Windows
18:51:02 --- quit: tathi ("leaving")
18:51:04 <madgarden> "29mar94py"
18:51:19 <madgarden> Is that some sort of Forthy convention?
18:51:19 <Sonarman> that's just that particular screen :)
18:51:33 <Sonarman> oh, it's a bigForth convention :)
18:51:41 <madgarden> I've seen it in other places.
18:51:50 <madgarden> Like, Blazin' Forth.
18:52:01 <Sonarman> oh yeah!
18:52:26 <Sonarman> well, it's a good timestamp format IMO
18:52:29 <madgarden> What's py?
18:52:38 <madgarden> Yes, I shall steal it.
18:52:58 <Sonarman> py is Bernd Paysan's ID
18:53:06 <madgarden> "ID"?
18:53:19 <madgarden> How is that his ID? 
18:53:26 <madgarden> In what was was it determined?
18:53:26 <Sonarman> he just chose it, i guess
18:53:51 <Sonarman> i suppose that it's usually one's initials, but bp was already taken by one of the volksFORTH authors
18:54:07 <madgarden> Hmm.
18:54:17 <madgarden> I wonder if there's a PP.
18:54:46 <Sonarman> or PU
18:54:56 <madgarden> I am not stinky.
18:55:04 <madgarden> Hmm, I don't like his Forth formatting style much.
18:55:21 <Sonarman> yeah, it's one of the "quirks" :)
18:55:44 <madgarden> Like, make up your mind, Bernd!
18:56:45 <Sonarman> you can't blame him-- he is crazy, after all
18:57:04 <madgarden> Seems that most -- if not all -- Forthers are.
18:57:40 <Sonarman> well, i was just joking
18:57:57 <madgarden> I wasn't!
18:58:58 <Sonarman> HEE HEE HO HO CRAZY HA HA HEAR THAT, ABE? HE SAYS WE'RE CRAZY!  NOW THE UNION NEEDS MORE FIREPOWER IF IT WANTS TO DEFEAT THE CONFEDS.
18:59:30 <yeoh> Wow, that was loud. I hear you.
19:00:00 <Sonarman> Alt
19:00:07 <Sonarman> whoops
19:00:56 <Sonarman> how annoying it is to accidentally press ^J when ^K was meant!
19:01:13 <madgarden> I want to make a screen editor with a fricking bigass font, like 10x10 pixel characters at 640x480.
19:01:42 <madgarden> ^K
19:01:56 <madgarden> : ^K ^J ;
19:02:00 <madgarden> problem solved!
19:02:33 <Sonarman> testing testing madgarden is a donkey butt
19:02:46 <slava> madgarden, the editor defines key bindings as forth words?
19:02:49 <madgarden> (  |  )
19:03:00 <Sonarman> hmm... seems that should have been: : ^J ^K ;
19:03:01 <madgarden> slava, what editor?
19:03:21 <madgarden> Sonarman, indeed you are correct.
19:03:23 <slava> i'm uncertain as to what i should add next to factor
19:03:39 <madgarden> slava, what else does it need?
19:03:49 <slava> nothing so far, that's the point :)
19:03:52 <slava> a debugger would be nice
19:05:19 <Sonarman> write a factor-style interface to SWT :)
19:05:20 <madgarden> Does it have a nice hyperlinky "syntax" colored editor?
19:06:27 <blockhead> does factor have real-time sound synthesis?  :D
19:08:14 <slava> the 'factor IDE' is planned
19:08:59 <madgarden> I have been thinking more about syntax hilighting for Forth code.
19:09:17 <madgarden> And *not* colorForth as a way of describing the source.
19:09:26 <Sonarman> what ideas have you come up with, mad?
19:09:35 <madgarden> Your stack flow, parsing words, immediate words, could all be described with colours.
19:09:53 <madgarden> The flow of the program would be more instantly recognizable.
19:10:22 <Sonarman> sounds good!
19:10:50 <madgarden> And, why does : EXECUTE >R ; give me an exception in FICL?
19:11:18 <madgarden> Sonarman, still pondering how to handle things like control flow words, especially if they take parameters.
19:11:43 <madgarden> But, I figure there'd be about 5 or 6 colours used... maybe more of you transition between them for certain combined cases.
19:11:46 <slava> here is my idea for a factor IDE
19:11:52 <slava> its based on the 'inspector'
19:11:59 <Sonarman> madgarden: maybe ficl scans for what seems like unbalanced return stack manipulation?
19:12:00 <slava> each object has an editor
19:12:19 <slava> the inspector calls up either a custom editor for an object, or just a list of slots where you can edit the slots
19:12:46 <slava> so the editor for a word is a structure editor, editor for a list is a list control, editor for a string is a text field, etc
19:12:50 <slava> and its all database-backed
19:13:15 <madgarden> Sonarman, could be. That bastard FICL!
19:13:52 <madgarden> slava, "slots?"
19:14:05 <slava> madgarden, think fields in a structure
19:14:10 <madgarden> Ahh.
19:14:15 <madgarden> CELLS. :P
19:14:27 <slava> also the inspector will be UI-independent
19:14:35 <slava> i have a basic inspector with command line and html UI already
19:14:54 <slava> also i want my editor to edit tokenized source -- a 'structure editor'.
19:15:11 <slava> so that its impossible to write code that doesn't parse
19:16:19 <madgarden> Hmm.
19:16:47 <madgarden> Sounds interesting.
19:17:33 <madgarden> And yes, a Forth editor could easily have checking for most words, I think.
19:17:46 <slava> in my case there are no parsing words
19:17:50 <madgarden> Here's what Minforth does:
19:17:50 <madgarden> 0 0 - : call >r ;  ok 
19:17:50 <madgarden> 0 0 - 10 ' dup  ok
19:17:50 <madgarden> 2 0 - call  ? Address alignment exception in CALL
19:18:02 <madgarden> slava, yours is 100% RPN, isn't it?
19:18:12 <slava> madgarden, with a few exceptions
19:18:30 <slava> madgarden, there are a handful of 'hand coded' parsing words, like :
19:19:18 <slava> it used to be no parsing words, but "square" [ dup * ] define looked ugly :)
19:20:09 <madgarden> Hmm.
19:20:27 <madgarden> There is no compile mode either, correct?
19:20:41 <Herkamire> 100% RPN with some exceptions :)
19:20:46 <Sonarman> does : foo 10  ['] dup  call ; work?
19:21:10 <madgarden> Win32Forth... gah!
19:21:10 <madgarden> : goto
19:21:10 <madgarden> >r ;  ok
19:21:10 <madgarden> 1 ' dup  ok..
19:21:10 <madgarden> goto
19:21:11 <madgarden> EXCEPTION: ACCESS_VIOLATION
19:21:49 <Sonarman> ' dinner call
19:22:12 <Herkamire> madgarden: ' probably gives you a pointer to the dictionary entry
19:22:29 <Herkamire> an XT is not nessesarily the CFA
19:22:59 <madgarden> Ahh.
19:23:37 <slava> madgarden, some words are sort-of evaluated at compile time, but its implicit.
19:24:12 <slava> the internals of the interpreter are more like lisp than forth actually.
19:24:23 <madgarden> Those could be coloured differently in the editor.
19:25:06 <madgarden> "implicit" hiding of programming features can make it hard to understand.
19:26:46 <slava> the 'compling words' are those that would behave the same if they're evaluated at compile or runtime.
19:27:37 <slava> for example, : twice ( code -- ) dup >r call r> call ;
19:27:43 <slava> [ "hello world" . ] twice
19:27:45 <slava> "hello world"
19:27:46 <slava> "hello world"
19:27:47 <slava> etc.
19:27:53 <slava> suppose you have:
19:28:03 <slava> : foo 1 2 3 [ + ] twice ;
19:28:14 <slava> it doesn't matter if 'twice' evaluates at runtime, or if it is expanded at compile-time into:
19:28:18 <slava> : foo 1 2 3 + + ;
19:29:02 <slava> in fact, while control structures are implemented by passing around linked lists, in compiled code there is no such thing, after all the transformations are done.
19:30:08 <madgarden> I see.
19:30:45 <madgarden> Why don't you have pure RPN syntax, and go for something like this?
19:30:46 <madgarden> "foo" : 1 2 3 [ + ] twice ;
19:31:08 <slava> the forth way looks nicer i guess.
19:31:30 <madgarden> Well... actually, the pure RPN way is more intuitive.
19:31:41 <madgarden> Especially once you know the simple syntax.
19:31:42 <slava> its just syntax sugar.
19:31:43 <slava>     0] ": foo bar baz ;" parse .
19:31:44 <slava> [ "foo" #<factor.FactorCompoundDefinition: foo> define ]
19:32:01 <madgarden> Syntactic sugar that's not needed, IMO.
19:32:22 <madgarden> Not sure what values you're trying to capture by doing it this way. it seems to detract from the system's purity.
19:35:45 <slava> i'm thinking of writing a little structure-aware line editor that can be used from the command line, as a neat hack :)
19:40:23 <madgarden> :)
19:41:08 <slava> i've got it--the debugger is just a continuation editor!
19:45:12 <kc5tja> Continuate THIS!
19:45:17 <kc5tja> Oh, wait . . . wrong movie.
19:46:08 <slava> i wonder if the stack effect of code that uses continuatiosn can be deduced, by looking at the program globally.
19:48:24 <Sonarman> i think i shall now try to understand the bigForth dynamic memory manager
19:48:27 <madgarden> Are continuations actually needed for anything, or are they just trendy and hip with the cool kids?
19:48:29 <Sonarman> key word: try
19:49:54 <slava> madgarden, i use them for co-operative multitasking
19:50:10 <slava> or loops that pass values between each other
19:50:13 <kc5tja> madgarden: That depends on your point of view.  It's a question that cannot be answered.  Like atoms in nature, continuations just "are".
19:50:51 <madgarden> Ok, but how useful "are" they?
19:51:12 <kc5tja> Well, continuations ultimately form the core of multitasking.
19:51:33 <madgarden> It's basically a saved stack frame, non?
19:52:00 <kc5tja> More generally, it's just the program state itself.
19:52:06 <kc5tja> (which necessarily includes the stack).
19:52:15 <slava> used sparingly, they simplify code.
19:53:07 <slava> for example, a blocking read word for a GUI console might save a continuation somewhere; when input is provided by the user, the event queue restores the continuation and the read word returns to its caller as if nothing happened.
19:53:25 <slava> so you can implement a read/eval loop 'spliced' into an event queue.
19:56:52 <madgarden> Seems to me that you could do this (very simply) in Forth by tossing a new return value on the stack in the middle of whatever's currently executing.
19:57:38 <kc5tja> madgarden: That return address is called a "partial continuation;" it differs from a complete continuation in that only the local state of the stack is affected, not the whole state of the program.
19:57:49 <kc5tja> And in most cases, partial continuations are all you really need.
19:58:38 <madgarden> A full continuation being an entire stack/register swap...
20:00:17 <madgarden> right?
20:00:24 <slava> yes.
20:00:33 <kc5tja> Stack, register swap, AND program state, yes.
20:00:51 <slava> in factor, a continuation is a datastack/callstack pair.
20:01:56 <slava> kc5tja, have you done any work on FTS/forth lately?
20:04:13 <madgarden> What is "program state" if not the stacks and registers?
20:04:42 <slava> in forth, the interpret/compile flag i guess
20:05:23 <madgarden> Right, any global variables, etc.
20:05:27 <slava> no
20:05:32 <slava> globals are not usually part of it
20:05:51 <madgarden> USER globals.
20:06:04 <slava> the current environment pointer is.
20:06:09 <slava> but variable values are not.
20:09:41 <madgarden> Hmm... I have this urge to "port" a Choose Your Own Adventure book to Forth.
20:10:30 <slava> YOU DECIDE TO BECOME A WORD. WHAT KIND OF WORD DO YOU BECOME:
20:10:37 <slava> - IMMEDIATE
20:10:39 <slava> - STATE-SMART
20:10:42 <slava> - COMPILE-ONLY
20:10:44 <slava> - NO-NAME
20:10:51 <madgarden> Heh!
20:11:14 <Sonarman> slava: you forgot to include what screen to load for each decision! :)
20:18:23 <madgarden> Yes!
20:18:25 <madgarden> I have...
20:18:34 <madgarden> "The Third Planet from Altair"
20:18:42 <madgarden> "Gorga, the Space Monster"
20:18:51 <madgarden> "Space and Beyond"
20:19:03 <madgarden> and "House of Danger"
20:20:49 <madgarden> Hmm, that's way more typing than I'm prepared to do.
20:21:08 <Sonarman> get a scanner and write an OCR system... in Forth@
20:21:11 <Sonarman> Forth!
20:21:26 <slava> that would be scary :)
20:21:39 <madgarden> That's also way more typing than I'm prepared to do. :)
20:22:04 <Sonarman> then write a speech-recognition system... in Forth!
20:22:34 <kc5tja> Hmmm...both of which rely extensively on neural nets.
20:22:45 <madgarden> In all honesty... the only thing that makes that *scary in Forth* is not using variables, and not having dynamic memory allocation (not that you cann't, but you know).
20:22:52 <madgarden> local variables, that is.
20:23:09 <slava> local variables are unnecessary, but memory management... :)
20:23:32 <slava> i don't consider ALLOT/FORGET to be on par with a generational incremental GC :)
20:23:49 <madgarden> local variables are unnecessary, as long as you're not trying to write some intensely convoluted algorithm.
20:24:12 <madgarden> Don't you mean ALLOT and - ALLOT? :)
20:24:21 <kc5tja> slava: That's because it's not.  It's more comparable to the memory management inferencing engine that ML uses to minimize its need to dynamically allocate memory in the first place.
20:25:46 <blockhead> 'nn all
20:25:56 --- quit: blockhead ("Client Exiting")
20:26:01 <madgarden> kc5tja, though I *am* going to make that text adventure, dammit.
20:26:12 <kc5tja> Hehe :D
20:26:15 <madgarden> I'll remake the first one I ever did. It was 3x3x3 rooms.
20:26:46 <madgarden> "Dungeonquest" I so imaginatively named it.
20:27:00 <kc5tja> OOOoooo....myst-EEERY-ous.  :D
20:27:11 <slava> bigforth seems to use a mix of blocks and files.
20:27:14 <kc5tja> (you'd understand the inside joke if you regularly read Sluggy Freelance.  :D)
20:27:31 <madgarden> Not so mysterious as the mysterious "Red room," "Green room," and "Blue room" that it contained!
20:28:35 <slava> a single-user wiki with support for styled text and embedded code...
20:28:41 <madgarden> I think I'll also add sound effects. :D They seemed to work nicely in some of the Infocom games such as Lurking Horror and Sherlock Holmes.
20:28:45 <kc5tja> Wait, let me guess -- your goal was to combine the rooms to produce the white room.  :)
20:28:50 <kc5tja> (with black curtains)
20:29:14 <slava> but to do that, you had to first deduce a topological homeomorphism between the rooms, in order for them to be combinable ;)
20:29:31 <madgarden> kc5tja, heh... no, nothing so obvious. Actually, you had to put on a magic amulet in one of the rooms, and there was know way you could know which room it was. Pick the wrong room, and you died! 
20:30:44 <kc5tja> Well, THAT'S no fun... :D
20:31:07 <madgarden> Heh no... 
20:31:43 <madgarden> But, at least you got to fight a bunch of trolls and orcs on the way down. You even had an elf and a wizard as buddies who wandered around doing stuff on their own.
20:32:12 <madgarden> I thought it was pretty damn cool when I wrote it, but I was about 11 or 12. :)
20:35:28 <kc5tja> :P
20:35:59 <slava> hsa anybody used bigforth before?
20:36:01 <madgarden> It's a perfect project for my current Forth abilities. ;)
20:36:08 <madgarden> slava, I used it about an hour ago.
20:36:26 <madgarden> I though the screen viewer was neat. Then I exited it. :P
20:36:49 <slava> its sort of like what i want to achieve
20:37:07 <madgarden> Forth with a GUI library?
20:37:25 <madgarden> And OpenGL bindings?
20:37:39 <slava> well, an integrated development environment
20:37:48 <slava> it won't be a GUI, but a styled-text-CLI-with-hyperlinks ;)
20:38:25 <madgarden> I guess I've not seen the IDE aspects of bigForth.
20:39:08 <slava> well, the browser, the editor, its all integrated in some sense :)
20:40:09 <madgarden> Heh.
20:40:11 <madgarden> Debugger?
20:40:43 <madgarden> I'd love to see a Forth with an MSVC-like debugger.
20:40:50 <slava> msvc?
20:40:58 <madgarden> M$ Visual C.
20:41:08 <slava> yea
20:41:13 <slava> whats so good about its debugger?
20:41:20 <madgarden> It's the environment I'm most familiar with for development.
20:41:35 <slava> sheesh... kids these days
20:41:50 <thin> i read sluggy!
20:42:28 --- part: thin left #forth
20:42:28 <madgarden> Well, it pretty much has everything you need, and it works well. Breakpoints, various forms of stepping, watch window, memory dump, stack trace, etc.
20:42:41 <madgarden> disassemble
20:43:28 <kc5tja> I actually kind of like MSVC for C development.  It's not at all bad.  One of the few products from Microsoft that (a) actually works, and (b) is actually half-way convenient to use.
20:43:37 <madgarden> And implementing all of this on a Forth system should be so much more straightforward than C/C++.
20:43:42 <kc5tja> Though I have to admit, I still prefer VI/ctags over it.
20:43:50 <madgarden> ctags?
20:44:06 <slava> ctags rocks.
20:44:08 <slava> i adopted it to forth
20:44:24 <kc5tja> A program that understands the structure of a C program to extract the location of all the identifiers and (optionally) type information, so that you can use Emacs or VI to *rapidly* jump to certain symbols.
20:44:33 <kc5tja> slava: Nice.  :)
20:44:41 <kc5tja> For Forth, I normally just use search though.
20:44:49 <kc5tja> I haven't had a reason to use ctags for Forth yet.
20:44:53 <slava> > cat ~/.ctags 
20:44:53 <slava> --langdef=forth
20:44:53 <slava> --regex-forth=/^:[ \t]*([!-~]+)[ \t]*/\1/d,definition/
20:44:53 <slava> --langmap=forth:.factor.lsd
20:45:03 <slava> except the --langmap will need changes :)
20:45:33 <madgarden> Hmm. Well, sounds neat, but I wouldn't call that an IDE. ;)
20:45:56 <madgarden> The MSVC debugger is king for me... it's saved my butt many many times.
20:46:34 <kc5tja> Ironically, that's the one thing I feel needs the most improvement.
20:46:35 <madgarden> Then again, I have been using MSVC for the past 7 years.
20:46:57 <madgarden> How so?
20:47:32 <kc5tja> The MSVC debugger is nice overall, but it sucks baboon balls when it comes to debugging assembly language code.  :(  And the memory "watch" functionality and memory dump functionality often get out of sync with the code you're stepping through.
20:47:47 <kc5tja> Even for C code.
20:48:44 <madgarden> Which version of MSVC were you using?
20:48:56 <madgarden> And what do you mean by "out of sync?"
20:48:56 <kc5tja> MSVC 6.2 or some such.
20:49:18 <kc5tja> A variable's address will be 0xDEADBEEF, but monitoring or dumping 0xDEADBEEF *won't* display the variable's contents.
20:49:43 <madgarden> I've never witnessed this effect.
20:50:00 <madgarden> It's been a rather consistent tool for me.
20:50:05 <kc5tja> I have.  It was a repeated occurance when I was working at Hifn to debug potential problems with chips (which turned out more often than not to be software issues)
20:51:16 <madgarden> Anyway, I think Forth could gain some major acceptance if such an IDE Forth was available, with a cool hyperlinking editor to boot.
20:51:51 <madgarden> That is, if it was free. :)
20:52:14 <madgarden> I imagine Swiftforth is something like this, though I haven't used it personally.
20:54:33 <kc5tja> I don't know.
20:54:46 <yeoh> Win32Forth has an IDE.
20:55:50 <madgarden> What IDE?
20:56:09 <madgarden> text debugger? 
20:58:54 <yeoh>   Having installed Win32Forth, and gotten to this point, you are probably
20:58:55 <yeoh>   starting to become familiar with the hyper-text editor/browser (and now
20:58:55 <yeoh>   also debugger) integrated into Win32Forth.
20:58:55 <yeoh> It's FAQ says that is has a hyper-text editor/browser and debugger. And WinEd, the editor.
20:59:26 <yeoh> Oop there earlier three lines were quotes from its help file.
21:09:20 <slava> here is something you guys might like
21:09:28 <slava> consider the stack effect of a function to be a pair (in,out)
21:09:40 <slava> define x*y to be the stack effect of first executing x, then y
21:09:49 <slava> this forms a very strange algebra indeed
21:10:09 <slava> however, if you define alpha((in,out))=out-in,
21:10:16 <slava> then alpha(x*y)=alpha(x)+alpha(y)
21:10:40 <slava> i'm trying to think of other such mappings from stack effects into simpler algebras...
21:10:56 <kc5tja> slava: I'm sorry, but you horrifyingly complicated the first case.
21:11:02 <kc5tja> x*y is simply the act of concatenation.
21:11:11 <kc5tja> What the *heck* does alpha(x) actually *do*?
21:11:19 <slava> yes but x and y are not programs here, they're pairs of numbers.
21:11:28 <kc5tja> You didn't define that.
21:11:30 <slava> [1,2]*[2,1] == [1,1]
21:12:10 <kc5tja> So x and y are the *number* of items on the stack, not the stack effects themselves.
21:12:17 <slava> yes
21:12:41 <kc5tja> So alpha(f) is the logical inverse of f.
21:14:25 <slava> if you define beta(in,out) = min(in,out), then there is a 1-1 relation from (in,out) to (alpha,beta) pairs, but i cannot think of a simple formula for beta(x*y) in terms of beta(x) and beta(y).
21:14:30 <slava> this isn't going anywhere :)
21:18:28 <slava> night.
21:18:30 --- quit: slava ("Leaving")
21:23:14 <kc5tja> Oh...kay...
21:23:37 <kc5tja> If that's meant to simplify things, then I'm sorely inadequate for programming.  :)
21:28:07 <kc5tja> Anyway, I'm going to go back to work on my FTS1002 chip design.
21:28:14 <kc5tja> This thing is going to be sweet when it's done.
22:08:36 --- join: Serg (~z@212.34.52.140) joined #forth
22:27:39 --- quit: yeoh ("User pushed the X - because it's Xtra, baby")
22:41:03 --- quit: Sonarman ("leaving")
23:26:39 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK")
23:29:33 --- quit: Serg ()
23:59:59 --- log: ended forth/04.04.08