00:00:00 --- log: started retro/13.05.03
08:08:13 --- join: ncv (~quassel@92.114.78.17) joined #retro
08:08:13 --- quit: ncv (Changing host)
08:08:13 --- join: ncv (~quassel@unaffiliated/neceve) joined #retro
08:50:38 --- quit: ncv (Remote host closed the connection)
09:23:30 --- join: ncv (~quassel@92.114.78.17) joined #retro
09:23:30 --- quit: ncv (Changing host)
09:23:30 --- join: ncv (~quassel@unaffiliated/neceve) joined #retro
09:31:46 --- quit: ncv (Remote host closed the connection)
11:45:51 --- join: Mat2 (~claude@91-65-144-133-dynip.superkabel.de) joined #retro
11:46:06 <Mat2> hi @ all
11:46:25 <tangentstorm> heya Mat2 :)
11:47:09 <Mat2> hi tangentstorm !
11:48:33 * Mat2 implementing new dictionary concept for next Saiwa version (+ backport to retro)
11:48:45 <tangentstorm> oh yeah?
11:50:02 <Mat2> yes, the new algorithm imporves performance because of index and content seperation (specially fpr Intel cpu's)
11:50:38 <Mat2> ^imporves
11:50:43 <Mat2> sorry improves
11:51:29 <Mat2> does my CDP test program work for you ?
11:51:51 <tangentstorm> yes, I got it working finally!! :)
11:52:03 <tangentstorm> haven't done anything with it yet but i will. thanks again. :)
11:52:10 <Mat2> that are good news for me
11:53:01 <Mat2> I have a plan for implementing it in hardware, just wait for my Papilio One board
11:53:24 <tangentstorm> oh you ordered one? :D
11:53:32 <Mat2> yes, nice design
11:53:41 <tangentstorm> Maybe you can show me how to use it :)
11:53:50 <tangentstorm> to be fair i haven't even looked at the docs yet.
11:54:20 <Mat2> we can study FPGA programming together
11:54:51 <tangentstorm> cool :)
11:55:40 <tangentstorm> I have one little cpu i kind of understand... niklaus wirth's compiler construction ebook has a verilog implementation of his cpu.
11:55:50 <Mat2> I had looked at the J1 cpu and think it is not complicated to extend it in the way of my AVM design (in fact, it is already somehat similar)
11:55:55 <tangentstorm> i don't understand the circuitry part but i understand how the machine itself works.
11:56:33 <Mat2> ^ok, one book more to read :)
11:56:43 <tangentstorm> oh. j1 is new to me. http://www.excamera.com/sphinx/fpga-j1.html neet screenshots :)
11:57:00 <tangentstorm> wirth's device is a register machine, but i guess the alu ought to be pretty similar
11:57:21 <Mat2> the J1 is a very clever MISC design
11:57:44 <tangentstorm> but this looks more like what we'd want anyway, yeah...
11:57:47 <tangentstorm> huh
11:58:07 <Mat2> it is also the core for the gameodino 
11:58:53 <Mat2> take also a look at www.opencores.org, specially the video cores are interesting
11:59:52 <tangentstorm> oh yeah. i've heard of that but never really looked into it.
12:00:02 <tangentstorm> i just don't have time to mess with this stuff right now :/
12:00:26 <tangentstorm> i need to ship my project so i can start making a paycheck again :)
12:00:48 <Mat2> ah ok
12:03:20 <Mat2> I have an interessant which want using Saiwa for some very ressource intensive applications
12:03:36 <tangentstorm> oh?
12:04:31 <Mat2> and it would be nice for him combining this with a embedded hardware solution
12:04:37 <tangentstorm> i've tried to explain saiwa to a couple people but i don't think it'll be very easy to explain until you have your book/docs/examples up
12:05:07 <Mat2> that is something I'm working slowly on
12:05:34 <Mat2> at current the language evolves a bit
12:05:53 <tangentstorm> fair enough. no point teaching something that's going to change :)
12:07:16 <Mat2> I found a way combining type checking and parsing (and restricting it this way to the tokenizer) for example which eases the language implementation a lot
12:07:41 <tangentstorm> ?
12:09:42 <Mat2> I restrict the type system to the parser, automatical converting typed data to vectored integers at demand
12:10:38 <Mat2> this way all input and output parameters for each function can be checked against validity before code generation
12:10:48 <tangentstorm> so like... you typecheck each token as it comes in?
12:10:52 <Mat2> yes
12:11:15 <tangentstorm> makes sense
12:12:06 <tangentstorm> i was playing around with syntax highlighting retro a while back, and i was thinking about maybe making the stack comments an actual part of the dictionary.
12:12:34 <tangentstorm> like in some computer-understandable coding system
12:13:22 <Mat2> that has two advantages: First type errors can be detected at typing and before compilation, second: the compiler is able to optimize most calculations to constant values before function invocation
12:13:36 <tangentstorm> especially the '"' comments.. i don't know what to call them... the ones where it manipulates the input stream as well as the stack.
12:14:00 <tangentstorm> yeah
12:14:03 <tangentstorm> like a supercompiler
12:14:29 <tangentstorm> hrm. that's not the word i was thinking of.
12:14:38 <Mat2> In addition the interpreter is able to detect all possible run-time errors before code generation
12:14:53 <tangentstorm> partial evaluating compiler i think.
12:15:54 <Mat2> I do not know if there exist some term for this technique
12:16:48 <tangentstorm> yeah, i have that problem a lot. :) i'll think of something and i don't know if it's a new idea or not because i don't know what someone else might have called it. :)
12:18:10 <Mat2> Java evealuate the source-code before compilation into J-code but I think this is a static approach not able to precalculate possible runtime errors
12:18:44 <Mat2> ^evaluate
12:19:52 <tangentstorm> hrm. dunno. i know haskell's type system can predict all the runtime errors.
12:20:16 <tangentstorm> hindley-milner type checking in general
12:21:02 <Mat2> I also use this method for implementing auto-parallelisation (which is beyond the scope of either the Java VM and the NET envirionment)
12:21:05 <tangentstorm> http://www.andres-loeh.de/LambdaPi/ <- really cool type system but every time i look at my notes i forget what all the symbols mean and have to re-learn it :/
12:21:23 <tangentstorm> man, you've got some pretty cool stuff going on.
12:21:52 <Mat2> I hope so :)
12:21:54 <tangentstorm> you gotta write it up... maybe someone will make you a professor or something :D
12:24:14 <Mat2> that would be nice 
12:24:48 <tangentstorm> how are your lectures going?
12:25:41 <Mat2> very good, I have converted the whole class to concatentative programming and stack processors *g*
12:26:23 <Mat2> last week I presented the GA144 cpu from Greenarrays as example
12:27:34 <Mat2> sadly there exist no cheap evaluating board
12:29:33 <Mat2> so the Propeller MCU was more interessting fro the audiance
12:29:41 <Mat2> ^interessting
12:30:04 <tangentstorm> :)
12:30:17 <tangentstorm> Mat2: you should talk to ttmrichter in #forth
12:30:32 <Mat2> who is that ?
12:30:33 <tangentstorm> he's a canadian (?) living in china and works with electronics
12:30:50 <tangentstorm> and he's been debating whether or not to make a ga144 prototype/exploration board.
12:31:03 <tangentstorm> but he doesn't know if he can recoup his investment.
12:31:40 <Mat2> thanks. A prototype board would be relally great
12:31:41 <tangentstorm> i think if there were a course or something that used the ga144 then there'd be a demand for it.
12:32:04 <tangentstorm> you have a course, he has the plans for the board and ability to manufacture it. :)
12:32:51 <Mat2> seems a good combination. The most important aspect will be the price then
13:04:40 --- nick: Mat2 -> Mat2-coding-afk
14:28:25 --- nick: Mat2-coding-afk -> Mat2
14:28:31 <Mat2> get some sleep, ciao
14:28:35 --- quit: Mat2 (Quit: Verlassend)
17:46:08 --- join: kumul (~mool@c-76-26-237-95.hsd1.fl.comcast.net) joined #retro
19:21:39 --- quit: kumul (Quit: Leaving)
20:30:38 --- quit: tangentstorm (Ping timeout: 255 seconds)
20:32:35 --- join: tangentstorm (~michal@108-218-151-22.lightspeed.rcsntx.sbcglobal.net) joined #retro
23:59:59 --- log: ended retro/13.05.03