00:00:00 --- log: started retro/12.11.08
05:47:40 --- join: impomatic (~digital_w@94.66.112.87.dyn.plus.net) joined #retro
06:14:27 --- join: kumul (~kumul@173.215.130.73) joined #retro
06:53:21 --- quit: impomatic (Quit: impomatic)
08:07:52 --- quit: kumul (Quit: WeeChat 0.3.9)
10:47:48 --- join: impomatic (~digital_w@94.66.112.87.dyn.plus.net) joined #retro
11:17:54 --- nick: harrison -> axolotl_hoffenas
11:18:16 --- nick: axolotl_hoffenas -> axolotlhoofenass
11:18:34 --- nick: axolotlhoofenass -> xltlhoofnssr
11:18:43 --- nick: xltlhoofnssr -> harrison
11:26:39 --- nick: harrison -> hp666
11:27:22 --- nick: hp666 -> satanicircnick
11:28:10 --- nick: satanicircnick -> haiku2012
11:28:48 --- join: Mat2 (~claude@91-65-144-133-dynip.superkabel.de) joined #retro
11:28:52 <Mat2> hello
11:29:44 <tangentstorm> wb
11:30:01 <tangentstorm> lots of new people in here today... hello quiet people :)
11:30:55 <Mat2> hi tangentstorm
11:31:43 <Mat2> haiku2012 maybe seem to be a friend of the haiku os ?
11:31:54 <impomatic> Hi all :-)
11:31:57 <tangentstorm> Mat2: what do you do for a living? I always get the impression you have like a 2-hour work day :D
11:32:12 <Mat2> hi impomatic
11:32:57 <tangentstorm> hey impomatic :) ... have we seen the end of itsyforth? :)
11:33:13 <Mat2> tangentstorm: I am a freelancer
11:34:10 <impomatic> tagentstorm: Not yet. I just need the time to tidy everything up.  I think most of Forth is implemented.
11:36:01 <Mat2> what is itsyforth ?
11:36:04 <impomatic> Also I've just bought an MSP430 board. I'm planning to put Itsy on there.  I'm thinking about making Itsy Forthlike at some point so I can tidy things up.
11:36:22 <Mat2> hmm, reads like camel forth
11:36:25 <tangentstorm> impomatic: i have to admit the assembler code kind of scared me off, but i still love the outer interpreter, and want to turn it into a lesson
11:36:28 <impomatic> Mat2: it's a minimal Forth in 8086 assembly language.
11:36:43 <Mat2> cool :)
11:37:33 <Mat2> what assembler do you use ?
11:37:37 <impomatic> I'd also like to build a decent Android Forth...
11:37:42 <impomatic> Mat2: nasm
11:38:03 <tangentstorm> Mat2: more importantly, it's tiny, and impomatic wrote each little piece up in a blog post http://www.retroprogramming.com/2012/03/itsy-forth-1k-tiny-compiler.html
11:38:18 <tangentstorm> that's the first one. you can see the others in the sidebar on the right
11:39:42 <tangentstorm> in the latest installment, someone else went through and added "what the heck is going on?" comments for practically every line : http://www.retroprogramming.com/2012/09/itsy-documenting-bit-twiddling-voodoo.html
11:41:49 <tangentstorm> which i think is a really good approach: you need someone who's already got it all worked out in their head to write the code, and then someone smart who doesn't understand it YET to point out what's confusing to newbies :)
11:43:13 <impomatic> Now I feel the need to comment my code better :-(
11:44:16 <tangentstorm> impomatic: no, what i mean is you can't
11:44:39 <tangentstorm> well, i mean obviously you can, but you're the worst person for the job :)
11:45:25 <impomatic> LOL. I try to imagine that I don't already know the code when I'm writing the blog posts...
11:45:36 <Mat2> try harder
11:45:43 <tangentstorm> you can't know how to explain something to someone else until you find out what they don't understand
11:46:22 <impomatic> It's the same if I try to write a tutorial for Corewar...  It really needs someone to write a tutorial as they're learning to play.
11:46:55 <tangentstorm> that's actually a really good approach. i mean you knew what were the important bits to put into that outer interpreter
11:47:16 <tangentstorm> and you narrowed it down to basically one screen of code
11:47:33 <impomatic> That was actually about the 5th attempt to write an outer interpreter!
11:47:45 <tangentstorm> :)
11:50:34 <tangentstorm> impomatic: It's really really good. I'm just saying learning/teaching is a feedback loop, and no matter how good you write, there's *always* going to be a mismatch with something in the student's head.
11:51:39 <tangentstorm> like for me... i just plain didn't know forth... I rewrote that code by hand, and i went through the ANS standard and looked up every single word you used.
11:52:51 <tangentstorm> by hand = on paper. i copied it to paper in black pen, then wrote the definition for each word in some other color.... i went over it like 4-5 times
11:53:50 <tangentstorm> ugh. every time i try to talk about this i feel like i'm insulting someone's work :/
11:54:13 <impomatic> No, it makes sense.
11:54:58 <impomatic> And now I know I wrote the outer interpreter exactly as it should be.  Only using standard words.
11:55:08 <tangentstorm> :)
11:55:13 <tangentstorm> it was actually a good exercise
11:55:39 <impomatic> Most of the other outer interpreters I've seen use quite a few non-standard words. They're a pain to understand :-(
11:57:51 <tangentstorm> hrm.
11:58:49 <Mat2> impomatic: Have you take a look at starting forth from leo brodie ?
11:58:52 <tangentstorm> if you're writing it for an audience that knows forth, then i agree 100%
11:59:35 <tangentstorm> as someone who knew almost 0 forth to begin with, i didn't care so much that the words were standard... what was important is that i could look them up.
11:59:55 <tangentstorm> the ansi standard just provided one way to do that
12:02:41 <tangentstorm> ... i'm just musing today. I just watched the gang in #lpmc adopt and drop a perfectly good js library for a game they're making, mostly overy issues in learnability and the installation process
12:03:54 <Mat2> impomatic: the goal of the ansi standard was providing a specific base on which an implementation can be build. Most implementation found more effective ways to do that
12:05:41 <Mat2> (bewqare I am not a big fan of the ansi standard)
12:08:21 <Mat2> so the ansi standard do not define a complete implementation
12:08:48 <Mat2> and in my opinion, that is the waek point of it
12:08:54 <Mat2> sorry weak
12:09:02 <tangentstorm> Mat2: i still don't know enough to have an opinion either way. i mostly just meant that there was a human-readable dictionary that made it easy to look words up. the html version of the ansi docs filled that role for me when i was studying his code
12:10:08 <tangentstorm> retro's docs weren't as easy to access at random... but now that we have the "describe" word in 11.5, it's much easier
12:15:35 <Mat2> tangentsorm: you can take a look at pfe, which support all forth standards from fig to forth 79, forth 83 and the actual ansi standard. The sources are understandable except the ansi extensions
12:16:34 <tangentstorm> http://pfe.sourceforge.net/ ?
12:16:37 <Mat2> yes
12:19:41 <Mat2> by the way, this is the famous forth version, which is used in some medicinical apperatures (and some satellites)
12:20:54 <Mat2> and old version of retro was used for digital image processing of the HRSC mars express mission
12:21:32 <Mat2> I mean a part of the image processing
12:21:43 <tangentstorm> interesting
12:22:10 <Mat2> that was a programming project of me and retro fitted well
12:22:27 <tangentstorm> :D awesome!
12:23:25 <Mat2> specifely it was part of the post processing for publishing digital images
12:28:58 <Mat2> sorry, the ng.ports.pas file has no interface. Does this mean all definations are public ?
12:29:24 <tangentstorm> no
12:29:49 <tangentstorm> ng.pas has the interface for ng.*.pas
12:30:04 <Mat2> thanks
12:30:21 <tangentstorm> actually... i renamed them to ng_*.pas so i could compile on ms-dos but i guess i forgot to commit that working copy
12:32:04 <Mat2> let's see how well the SDL code work
12:32:26 <impomatic> Mat2: I have "Starting Forth", but haven't read it yet.  I have about 20+ Forth books actually.  I'm a big fan of "The Forth Encyclopedia" and "Object Oriented Forth".  "All About Forth" is also pretty good.
12:32:58 <tangentstorm> Mat2: will it mess you up if i push the rename?
12:34:13 <tangentstorm> impomatic: You got me interested in that mouse book, and the recursion in pascal one too.
12:34:51 <impomatic> I'm still looking for a copy of these books... Very tricky to find: Solosky "FORTH: Elements of Programming Style", Wesseling "Forth: Implementation and Application", Ting "Inside F83", Koopman "Forth Floating Point"
12:37:52 <impomatic> I'm not a fan of the ANSI standard. It feels bloated. E.g. how many different types of division?
12:37:52 <Mat2> tangentstorm: no, otherwise you changed the code significally beside renaming
12:39:41 <Mat2> sorry for the scandinavian english :)
12:40:20 <Mat2> impomatic: I think your feel is right
12:47:54 --- nick: haiku2012 -> hp666
12:48:00 --- nick: hp666 -> satanicircnick
12:48:04 --- nick: satanicircnick -> haiku2012
12:48:21 * tangentstorm pushes the renamed files to github
12:49:11 <haiku2012> hp666/satanicircnick/haiku2012 is a haiku
12:49:43 <tangentstorm> i voted for limerick. no hard feelings?
12:51:28 --- nick: haiku2012 -> harrison
12:52:49 <Mat2> haiku212 is internic geschäftlich *lol*
12:54:27 <harrison> impomatic: http://www.latrobe.edu.au/phimvt/joy.html
12:54:53 <harrison> if you haven't read them you are in for a rare diversionary treat
12:55:20 <tangentstorm> a book on joy from a professor of logic who taught his courses in pascal :)
12:55:57 <tangentstorm> i mean a page about joy, the language he crated :)
12:55:59 <harrison> not joy as in de vivre
12:58:50 <Mat2> his logical foundations are important because there prove stack oriented languages are well suited for functional programming
12:59:18 <Mat2> there only differ in perspective
13:09:41 <Mat2> tangentstorm: ^H : write( ^H, ' ', ^H );
13:10:16 <Mat2> the ^H, what control code is that, a tabulature ?
13:11:16 <harrison> i worte a joyish language in oberon -- "toy"
13:11:18 <harrison> long ago
13:11:24 <harrison> it was fun
13:11:40 <harrison> i used a chain of boxes -- no stack
13:12:01 <tangentstorm> Mat2: ^H is backspace
13:12:05 <harrison> and when i had done it suddenly i realized that i understood "continuations"
13:12:09 <Mat2> ah ok :)
13:13:09 <Mat2> harrison: with chain of boxes, do you mean a linked list of arrays ?
13:13:36 <tangentstorm> Mat2: turbo pascal added the ^ syntax for control characters. you could also say #8
13:15:10 <Mat2> that's sad because '^' as operator is also used for pointer references !
13:15:43 <tangentstorm> yeah, at the end of the token.
13:15:52 --- join: kumul (~kumul@173.215.130.73) joined #retro
13:16:03 --- quit: kumul (Remote host closed the connection)
13:16:20 --- join: kumul (~kumul@173.215.130.73) joined #retro
13:17:09 <harrison> arrays? no, boxed types: a linked list of boxes which contain integers, characters, linked lists of boxed types ... etc
13:17:14 <tangentstorm> turbo pascal uses prefix @ for adress-of, prefix ^ for pointer-to ( in types ) or control-code ( for literals )
13:18:25 <tangentstorm> I actually use control characters kind of a lot lately. There's a good reference here: http://www.cs.tut.fi/~jkorpela/chars/c0.html
13:18:29 <tangentstorm> also in the "mn" directory :)
13:19:53 <Mat2> harrison: i see, my oberon knowledge is limited
13:21:14 <harrison> The "chain-of-boxes" design and oberon are orthogonal
13:21:33 <harrison> i wish i had my old toy code 
13:21:47 <harrison> now i would have to reconstruct it if i wanted to show you
13:22:07 <harrison> and of course i am always busy with the tracer
13:22:16 <Mat2> tangentstorm: it is good that freepascal support adr(x) := '@x'
13:22:21 <harrison> if i have any code juice in me i use it there
13:22:28 <harrison> lazarus pascal?
13:22:36 <Mat2> yes
13:22:45 <harrison> that's a good project
13:22:47 <harrison> in that
13:23:04 <harrison> fnecati adapts code from there to oberon
13:23:08 <harrison> specifically 
13:23:47 <harrison> i think he used the opengl interface for our opengloberon modules
13:24:02 <harrison> but
13:24:17 <harrison> do you have a realtime raytracer?
13:24:51 <harrison> he asked innocently with a wan smile
13:25:08 <harrison> while knowing the answer
13:25:38 <harrison> http://www.youtube.com/watch?v=eYSfWcW8cYA&feature=youtu.be
13:26:13 <tangentstorm> http://www.freepascal.org/docs-html/ref/refsu60.html <- is this sort of like a boxed type?
13:26:31 <tangentstorm> i guess oberon doesn't have CASE in records though...
13:27:26 <harrison> no. that must be one of the baroquely complex bits wirth took out.
13:28:27 <harrison> A boxed type would be like: (the harrison prepares to riff)
13:30:18 <harrison> TYPE toybox=RECORD; VAR prev,next: toybox END;
13:31:07 <harrison> TYPE intbox=RECORD(toybox); VAR int: INTEGER END;
13:31:11 <harrison> ...
13:40:22 <tangentstorm> got it
13:41:05 <tangentstorm> actually, i think it's very close to the same idea under the hood. he replaced variant types with type inheritance.
13:47:18 <harrison> i do not wish to seem forward,
13:47:41 <harrison> but may i enquire, sir, as to the operating system you currently operate?
13:49:43 <tangentstorm> who, me?
13:50:26 <harrison> who else?
13:52:53 <tangentstorm> oh. i thought we talked about this the other day. i generally use xp or vista and telnet into a debian box in a full screen console
14:02:47 <harrison> excuse me 
14:03:13 <harrison> i do have difficulty keeping track of the dizzying multiplicity of irc
14:05:56 <tangentstorm> ah yeah. me too.. i just remembered you because of the oberon list :)
14:11:18 <harrison> there is a good aos for windows
14:11:25 --- quit: impomatic (Quit: http://corewar.co.uk)
14:12:12 <harrison> http://www.ocp.inf.ethz.ch/wiki/OCP/WinAos
14:14:19 <Mat2> tangentstorm: I have replaced the terminal code in ng.ports.pas with code suitable for my VDP core, debugging follows tomorrow
14:14:30 <Mat2> I need some sleep, ciao
14:15:10 --- quit: Mat2 (Quit: Verlassend)
14:20:09 <tangentstorm> harrison: yep! i've played around with it, but haven't really gotten proficient with it
14:23:09 <harrison> the compiler is much much faster now than it was earlier this year
14:23:17 <harrison> at least on unix
14:23:34 <harrison> very likely for winaos too since it is the same code
14:30:01 <harrison> tangentstorm: i would imagine that it is easier to learn than pascal would be
14:30:15 <harrison> due to its radical simplicity
14:30:33 <tangentstorm> harrison: yeah, i'm writing an oberon compiler. i just call it pascal because nobody knows what oberon is :)
14:30:58 <harrison> but why write a compiler rather than using aos?
14:31:20 <harrison> that is like writing a c compiler instead of using linux
14:31:22 <tangentstorm> it's not for me
14:31:39 <harrison> ok i sort of take htat back
14:31:50 <harrison> it is good to write a compiler no matter what
14:31:54 <harrison> what i mean is
14:32:02 <harrison> please use aos
14:32:04 <harrison> also
14:32:26 <tangentstorm> i can't use aos for what i'm doing :)
14:32:37 <tangentstorm> the core oberon system maybe
14:33:57 <harrison> are you making a raytracer?
14:34:03 <tangentstorm> no
14:35:19 <tangentstorm> i honestly don't know how to explain what i'm making. :/
14:35:39 <harrison> Then you have already succeeded.
14:35:54 <harrison> I already don;t understand it.
14:36:44 <tangentstorm> okay, so for now, i'm making a compiler for a language like oberon-07 to target the ngaro vm that retro uses.
15:36:36 <harrison> well, that is good news
15:36:56 <harrison> ngaro vm is simple right?
15:37:11 <harrison> you could prt the VM to aos first
15:37:16 <harrison> recursiony
15:42:11 <tangentstorm> harrison: i might just take you up on your offer to help with the AOS side of things if i do that. I just don't have time to do it right now.
15:48:34 <harrison> roger daltry that
17:40:09 --- join: hhm (6c11508f@gateway/web/freenode/ip.108.17.80.143) joined #retro
17:41:27 --- quit: hhm (Client Quit)
18:26:43 --- quit: tangentstorm (Ping timeout: 252 seconds)
18:26:59 --- join: tangentstorm (~michal@108-218-151-22.lightspeed.rcsntx.sbcglobal.net) joined #retro
19:40:04 --- quit: kumul (Quit: WeeChat 0.3.9)
23:59:59 --- log: ended retro/12.11.08