00:00:00 --- log: started forth/09.11.20
00:03:45 --- quit: qFxo (Read error: 104 (Connection reset by peer))
00:05:19 --- join: qFxo (n=C00K13S@5356B263.cable.casema.nl) joined #forth
00:05:57 --- quit: qFox (Read error: 110 (Connection timed out))
00:10:16 --- nick: qFxo -> qFox
00:45:20 --- quit: ygrek (Remote closed the connection)
01:14:33 --- join: ygrek (i=user@gateway/gpg-tor/key-0x708D5A0C) joined #forth
01:19:42 --- quit: proteusguy (Read error: 110 (Connection timed out))
01:20:40 --- join: proteusguy (n=proteusg@110.77.145.40) joined #forth
01:22:46 --- join: ASau (n=user@host195-231-msk.microtest.ru) joined #forth
02:00:50 --- log: started forth/09.11.20
02:00:50 --- join: clog_ (n=nef@bespin.org) joined #forth
02:00:50 --- topic: 'Welcome to #forth. We discuss the Forth programming language and a variety of technical subjects. Introduction and  Helpful Reading: http://forthfreak.net/index.cgi?FnFC | ANS/ISO Forth Standard doc: http://tinyurl.com/nx7dx | Gforth compiler:  http://tinyurl.com/s8uho | http://quartus.net/search | Paste: http://forth.pastebin.ca | http://www.isforth.com'
02:00:50 --- topic: set by I440r on [Fri Feb 13 08:39:24 2009]
02:00:50 --- names: list (clog_ ASau proteusguy ygrek qFox Snoopy_1611 nighty^ TR2N dinya_ maht_ I440r kleinjt gnomon mathrick segher malyn gogonkt clog nottwo Quartus` schme madgarden scj nighty___ probonono aguai zbrown aguaithegeek KipIngram uiu_ Frek madwork yiyus uiu cataska)
02:07:52 --- quit: nighty^ (Remote closed the connection)
02:14:03 --- quit: clog (Read error: 110 (Connection timed out))
02:14:03 --- nick: clog_ -> clog
03:07:28 --- join: nighty^ (n=nighty@x122091.ppp.asahi-net.or.jp) joined #forth
03:16:07 --- join: GeDaMo (n=gedamo@212.225.98.255) joined #forth
03:23:15 --- quit: dinya_ (Read error: 104 (Connection reset by peer))
03:24:01 --- join: dinya_ (n=Denis@188.18.83.251) joined #forth
03:28:56 --- quit: proteusguy ("Leaving")
04:31:42 --- quit: TR2N (Read error: 110 (Connection timed out))
04:39:34 --- quit: GeDaMo ("Leaving.")
05:25:55 --- quit: ASau (Remote closed the connection)
05:26:16 --- join: ASau (n=user@host195-231-msk.microtest.ru) joined #forth
05:51:50 --- join: tathi (n=josh@dsl-216-227-91-166.fairpoint.net) joined #forth
06:04:14 <KipIngram> Neat.  I didn't realize "Thinking Forth" was available as a pdf download.
06:23:50 --- join: DrunkTomato (n=DEDULO@217.18.135.36) joined #forth
07:09:47 --- join: dinya (i=Denis@92.255.146.156) joined #forth
07:34:15 --- join: GeDaMo (n=gedamo@212.225.98.255) joined #forth
07:36:47 --- join: PoppaVic (n=pops@99.150.139.232) joined #forth
07:55:55 --- join: TreyB (n=trey@adsl-76-254-113-201.dsl.hstntx.sbcglobal.net) joined #forth
08:05:39 --- quit: nighty^ (Remote closed the connection)
08:07:54 <KipIngram> I'm trying to invent a nice set of words for real-time I/O control.  I want to include reasonably minimal generic timing functionality in my FPGA, and hook it into the Forth engine.
08:08:24 <KipIngram> I haven't thought this all out too carefully yet, but in some sense this will capture the function of interrupts, timer-counters, and so on.
08:08:34 <ASau> Tried reviewing POSIX?
08:08:50 --- quit: ASau ("off")
08:09:36 <KipIngram> You mean to use, or to get ideas?
08:09:47 <KipIngram> I had in mind something like this:
08:10:01 <KipIngram> <fpga pin #> OUTPUT <name>
08:10:36 <KipIngram> to declare an output and tie it to an FPGA pin.  Then I could define a word to manage that signal and store it's CFA like so:
08:12:11 <PoppaVic> event loop
08:14:53 <PoppaVic> KipIngram:  I'd start by thinking of what you need to handle as input to perform on.. THe syntax will never really be satisfying ;-)
08:19:43 <KipIngram> Sorry; my netbook hung and I had to reboot.
08:20:05 <KipIngram> This is for interfacing stuff like serial ADCs, DACs, communication links, and so on.
08:21:05 <KipIngram> Anyway, my idea is to have a vector queue integrated into my processor hardware.  If a word "returns" and there's nothing on the return stack to return to, then the hardware will check the vector queue. 
08:21:27 <PoppaVic> KipIngram:  sounds complicated
08:21:32 <KipIngram> If it's empty as well then nothing happens until a vector shows up.  If / when something is in the vector queue the hardware will go there and run.
08:22:27 <KipIngram> Anyway, each of these outputs will have a handler, and when the timer for that output expires hardware will put the address of the handler in the execution queue.
08:23:25 <KipIngram> The alternative of course is to poll.
08:23:45 <KipIngram> But I might have a lot of I/O, any one of which is very simple, but the polling loop might get pretty involved.
08:24:19 <KipIngram> This is really just a timer interrupt facility.
08:24:37 <KipIngram> I rather like the idea that something wel-defined happens if the outermost word returns.  It just means there's no work to do.
08:25:33 <KipIngram> Some of the modern FPGAs also have what they call a "flash freeze" state (at least that's what Actel calls it).  In other applications where the interrupts were external instead of based on timers then I could consider flash freezing the FPGA when the outermost word returned, to save power.
08:25:38 <PoppaVic> still sounds weird. I was reexamining the 8080/z80 rst opcode awhile back. raise an int line, the byte in some port is the int# - handle the i/o
08:25:52 <KipIngram> As soon as something "happened" (say, a keystroke) the thing would wake up and execute from the proper place.
08:26:50 <KipIngram> Ok, in that case then I'd have code that read the port and executed something based on that.  The only difference I'm proposing is that instead of an int # the value that comes in is an actual service routine address, that you'd set during initialization.
08:27:27 <KipIngram> And if the processor wasn't doing anything else at the time your service code could start running within nanoseconds.
08:28:40 <PoppaVic> yeah, it's easier to use a "service-slot" - an array of those addresses - the circular-queue of signals are resolved in the loop, the slot is indexed by the "signal byte" - and you can follow that in the handler with whatever data they need to snag. I'm not a super chip-head, so at this point I;ll just kibitz ;-)
08:28:42 <KipIngram> Anyway, the service routine would specify the next value to be written to the output and the time at which that should occur.  It would write that information into the hardware and then return.
08:30:46 <KipIngram> For inputs the expiration of the timer would cause the hardware to clock the current value of the input into a sample register and queue up the handler address.
08:30:57 <PoppaVic> KipIngram:  you are speaking of the bios-level code anyway, right? or lower - the code to service the magic hw (microcode?)
08:31:05 <KipIngram> The handler would read the sampled value, do whatever needed to be done, and set the time of the next sampling eveng.
08:32:14 <KipIngram> I think that these hardware functions (queuing up execution vectors and clocking inputs and outputs on timer expiration) would correspond to microcode, yes.  The code that gets executed in the handlers, though, would be application level code.
08:32:31 <PoppaVic> yah
08:32:36 <KipIngram> You'd write the signal handlers specifically for your app.
08:32:55 <KipIngram> So anyway, when I crashed I was about to type this:
08:33:04 <KipIngram> <pin> output <name>
08:33:13 <KipIngram> ' <handler> <name> handler !
08:33:23 <KipIngram> <time value> <name> next !
08:33:24 <PoppaVic> Easier for me to envision bios-services for all the event/scheduled and the rerouting and "handling" is the program space or a driver, etc
08:33:29 <KipIngram> <val> <name> value !
08:34:21 <PoppaVic> clumsy voc, that - might consider a struct and whatnot.. regularize the syntax.
08:34:36 <KipIngram> Yes.  Like I said, just getting going with this idea.
08:35:01 <PoppaVic> KipIngram:  what did you see the openbios volken have employed?
08:35:31 <KipIngram> I decided that my interest in making this thing ultimately fast (while still something I want to do someday) is a bit too "hobby-ish" for me to ponder at work.  On the other hand, an extremely precise signal manager that ran at, say, 100 MHz would be of work-related interest.
08:35:48 <KipIngram> I don't know what you mean.
08:36:27 <PoppaVic> KipIngram:  the openbios folks used to have a freenode channel - that's the semi-forth stuff for scaled and portable bios?
08:36:48 <KipIngram> I don't want a bios.
08:37:15 <KipIngram> Maybe I just have this facility I've just described attached to a serial port.
08:37:26 <PoppaVic> umm..
08:37:30 <KipIngram> With handler code that would let me have "outer interpreter" functionality.
08:37:37 <KipIngram> Guess that would be a bios, but that would be it.
08:37:47 <PoppaVic> OK, well whatever.. You can make it as easy or tough as you like.
08:37:49 <KipIngram> Just enough to send source code intot he thing.
08:42:06 <KipIngram> This thing is going to live on boards we design for our downhole tools.  The serial port would give me "programming access."  I'd define the rest of the pins however they needed to be defined to interact with the rest of the board.
08:42:30 <PoppaVic> yah
08:42:35 --- join: pgas (n=user@pdpc/supporter/active/pgas) joined #forth
08:42:54 <PoppaVic> I figured this was all not basement-boogie ;-)
08:42:59 <KipIngram> What I'm trying to do here is get away from having to write custom FPGA patterns in VHDL for every job.  Most of what we do just requires manipulating and monitoring signals with very precise timing; that's what drives us to an FPGA instead of a processor.
08:43:21 <PoppaVic> "very precise"?
08:43:23 <KipIngram> So if I can pull something like this off I get that functionality into the realm of a "processor."
08:44:04 <KipIngram> I intend for this thing to run at 100 MHz, and it will change outputs on precisely whatever edge of that clock I tell it to.  The output won't change when the processor writes the value register; that's just a holding place.
08:44:15 <KipIngram> That's the value that gets clocked to the output the next time the timer counts down.
08:44:27 <KipIngram> As long as I get it ready before it's needed it's fine.
08:44:59 <KipIngram> Most of our work involves sampling audio signals at relatively low sample rates, so there's *lots* of time between edges to run the handlers.
08:44:59 <PoppaVic> it's just a latched "reg"
08:45:08 <KipIngram> Yes.
08:45:30 <KipIngram> Same with inputs; they get sampled on precisely the clock edge I choose, and then the sample waits for me to come and read it.
08:45:50 <PoppaVic> You might just as easily used mem-mapped I/O as well.
08:46:18 <KipIngram> So that replaces pretty much all of our existing FPGA functionality, and yet I still have a "processor" that could take on the general purpose stuff that we use our other processors for now.
08:46:59 <PoppaVic> I've no experience with FPGA, other than the one shop were i was a pin-tester/tester/repair-bot.
08:47:44 <PoppaVic> from what I know, it's just a solid-state version of the ol' pinboard and jumpers I had in school ;-)
08:47:52 <KipIngram> And the serial port and the access it provides to the interpreter would give us a hook into the system for early debugging.  We could poke stuff on the board to see if it worked, and so on, and then gradually build up the primary functions.
08:48:11 <KipIngram> Yeah, it's just logic in a can.
08:48:18 <KipIngram> Lots of logic in a can.
08:48:26 <KipIngram> Lots of *fast* logic in a can.
08:48:49 <PoppaVic> yah.. design the logic blocks in one unit, burn it - the board of support surrounds it.
08:49:55 <KipIngram> You're right, though, my syntax is busted.  I want one FPGA pattern that never changes; I want the Forth code to do the customization.
08:50:14 <KipIngram> That means I don't get to decide at runtime whether a pin is an output, an input, an i/o, or whatever.
08:50:25 <PoppaVic> that's why I said BIOS - that was the entire point to it.
08:50:31 <KipIngram> And I don't get to just decide to create a new countdown timer; that's fixed at chip burn time.
08:51:29 <KipIngram> So I just need a flock of them that I can either use or not.
08:51:57 <KipIngram> Probably best to configure all the pins as I/O; I can choose at runtime whether to enable the output driver or not.
08:52:57 <KipIngram> Anyway, exactly how to wire up the pins has nothing to do with Forth, so I'll hush.
08:53:57 <KipIngram> And yes, the pin input and output values themselves will wind up memory mapped, PoppaVic.
08:54:10 <PoppaVic> ok
08:54:40 <PoppaVic> which of course means the timers as well. Since they are merely "devices" the fpga implements.
09:02:24 --- quit: dinya (Read error: 131 (Connection reset by peer))
09:03:23 --- join: TR2N (i=email@89-180-140-82.net.novis.pt) joined #forth
09:09:31 <PoppaVic> KipIngram:  yeah.. In your case, since you are designing a true chip, I am at a loss as to where to "code" what.. In what.. ;-)
09:12:50 --- join: kar8nga (n=kar8nga@jol13-1-82-66-176-74.fbx.proxad.net) joined #forth
09:33:13 --- join: ElmerJFudd (i=1000@c-68-56-142-41.hsd1.fl.comcast.net) joined #forth
09:36:51 --- quit: Quartus` (Read error: 110 (Connection timed out))
09:42:34 --- join: ASau (n=user@83.69.227.32) joined #forth
12:25:53 --- log: started forth/09.11.20
12:25:53 --- join: clog (n=nef@bespin.org) joined #forth
12:25:53 --- topic: 'Welcome to #forth. We discuss the Forth programming language and a variety of technical subjects. Introduction and  Helpful Reading: http://forthfreak.net/index.cgi?FnFC | ANS/ISO Forth Standard doc: http://tinyurl.com/nx7dx | Gforth compiler:  http://tinyurl.com/s8uho | http://quartus.net/search | Paste: http://forth.pastebin.ca | http://www.isforth.com'
12:25:53 --- topic: set by I440r on [Fri Feb 13 08:39:24 2009]
12:25:53 --- names: list (clog Maki Judofyr ASau Quartus` ben_m ElmerJFudd TR2N pgas TreyB PoppaVic GeDaMo tathi dinya_ ygrek qFox Snoopy_1611 maht_ I440r kleinjt gnomon mathrick segher malyn gogonkt nottwo schme madgarden cataska uiu nighty___ scj probonono aguai zbrown aguaithegeek KipIngram uiu_ Frek madwork yiyus)
12:26:05 <PoppaVic> amen
12:26:32 <KipIngram> Sorry for that split infinitive.
12:26:43 <ASau> And contrary to expectations, the loss of parentheses (like in Forth) does make text harder to parse.
12:26:50 <ASau> Even technically.
12:26:51 <KipIngram> My mom would frown at me (elementary school teacher before she retired).
12:27:32 <KipIngram> Only if your sentences are so long that you can't keep them straight any other way.  Good Forth is all about small definitions that you can wrap your brain around and *see* the correctness of.
12:28:01 <KipIngram> It's related to that thing about how we can keep seven or so things in short term memory at once.  A definitions that's seven-ish words or less long is tractable in a way that a longer one isn't.
12:28:50 <ASau> KipIngram: this is another language mismatch.
12:29:02 <KipIngram> Mismatch?
12:29:05 <ASau> English is more analytic language.
12:29:14 <ASau> "impedance mismatch"
12:29:21 <PoppaVic> infix math is pestiferous. You want it? writer a lexer and a parser. Don't need it, rpn works As Advertised.
12:29:37 <KipIngram> Yep.
12:30:04 <KipIngram> Besides, when you give a kid a calculator and tell them to calculate this:
12:30:14 <KipIngram> (42.7 + 16.3) * (19.8 - 6.4)
12:30:36 <ASau> Good calculator allows you entering algebraic forms.
12:30:43 <KipIngram> what do they do?  They do one of the sums and write it down.  Then they do the other sum, probably see that they don't need to write it down, punch the first sume back in, and multiply.
12:30:54 <KipIngram> They *wrote it down*.
12:31:04 <ASau> You live in past.
12:31:07 <KipIngram> Yes, and my kids still do what I just said. 
12:31:12 <KipIngram> I've watched them.
12:31:23 <PoppaVic> kids don't need calculators - let them do the longhand, learn to use slipsticks, learn to rtfm of logs and whatnot.
12:31:26 <KipIngram> With RPN the stack holds those for you, thank you very much.
12:31:27 <ASau> Good calculatos allow you to enter algebraic expressions.
12:32:00 <ASau> It's like teaching how to use slide rule.
12:32:11 <ASau> Archaic.
12:32:19 <PoppaVic> bullshit
12:32:29 <KipIngram> But the point, ASau, is *why* even start out with infix?  It's the lesser of the two forms.  It's only reason for existing is that we got started with it.  If I could push a button and replace it in everyone's head world wide with RPN, there would be *no* inclination to go back.
12:32:48 <KipIngram> Learning to use a slide rule helps solidify understanding of logarithms.
12:32:49 <ASau> Because infix is natural for associative operators.
12:33:04 <ASau> Neither prefix nor postfix are.
12:33:18 <PoppaVic> KipIngram:  let alone ratio & proportions, "floating point" and "rounding"
12:33:31 <KipIngram> Yeah, I believe in slide rules.
12:33:57 <KipIngram> A kid can hold it in his or her hands, and suddenly the concepts mean something.  They provide a tactile route for those concepts into the mind.
12:34:08 <PoppaVic> KipIngram:  I believe I inherited about 15, from dad.. to which I add my flight-comp, his, and my circ-sliderule
12:34:28 <KipIngram> ASau, I've seen, somewhere, a treatment of all the algebraic properties (commutative, associative, etc.) in RPN.  It can be done.
12:34:43 <KipIngram> I don't know if I can put my hands on it, though.  Maybe in one of Ball's books.
12:35:16 <ASau> KipIngram: sure, but it is much more complex.
12:35:23 <KipIngram> There's a threaded slide rule that gets like 5-digit accuracy on display in the UT Austin engineering library.
12:35:34 <ASau> You have to recreate a kind of AST in your memory.
12:35:50 <ASau> Infix notation allows you to bypass that step.
12:36:02 <KipIngram> Even if I grant that point, there's no reason to cling to learning concepts for your entire life of execution.
12:36:20 <KipIngram> So use infix to explain the properties, but not to teach manipulation.
12:36:22 <ASau> Associative operations are basics.
12:36:30 <ASau> Half-groups are everywhere.
12:36:31 <KipIngram> Being able to do the arithmetic fast and effectively is important too.
12:36:32 <PoppaVic> write the lexer-parser <shrug> Stop pointing at the machine.
12:38:07 <KipIngram> Look, here's the thing.  I used infix in high school.  I even used the ( ) keys.  Then when I discovered that HP calculator in college I flipped as fast as I could and never, *ever* had the slightest temptation to go back.
12:38:15 <KipIngram> There has to be a reason for that.
12:38:39 <KipIngram> Don't compare the number of infix users with the number of RPN users.
12:39:06 <KipIngram> Compare the number of people that have switched fron infix to RPN to the number that has switched the other way.  They're the only ones that are fully informed; they've tried both.
12:39:09 <I440r> no. thats like saying C is better than forth because more people use C
12:39:16 <I440r> i would object to that argument :)
12:39:22 <KipIngram> I guarantee you that the infix -> RPN crowd will *dwarf* the RPN -> infix crowd.
12:40:33 <KipIngram> I440r: right, only even more so.  I do still use C occasionally, but never infix unless I don't have my RPN calculator with me.  And then I hate every second of it and grouse about not having a real calculator.
12:43:27 <PoppaVic> everything about infix complicates the lexing/parsing.
12:43:40 <KipIngram> a(b+c) == (ab + ac) -------> a b c + * == a b * a c * + 
12:44:00 <KipIngram> Not sure I see the teaching advantage of one over the other.  You just need the one that corresponds to what you're familiar with.
12:44:49 <ASau> KipIngram: try (a+b+c)(d*e+f*g).
12:44:58 <KipIngram> (b + c) == (c + b) ------------------> b c + == c b + 
12:45:01 <ASau> And you'll see how infix is much more POWERFUL.
12:45:13 * PoppaVic sighs
12:45:38 <I440r> d e * f g * + a b c + + *
12:45:41 <KipIngram> But you don't teach it that way.  
12:45:55 <KipIngram> You teach it using the expression I wrote down, as a *concept*.
12:46:03 <I440r> infix is more powerful for visualizing the problem
12:46:12 <ASau> I440r: now show rules to expand it.
12:46:12 <I440r> but NOT for performing it
12:46:29 <PoppaVic> I440r:  more machine like, too - lower down and dirty and simple.
12:46:36 <KipIngram> You spend more of your life using these techniques than learning them.  Hopefully at least.
12:47:17 <KipIngram> And down and dirty == FUN.
12:47:20 <PoppaVic> KipIngram:  well, lightly or heavily
12:47:24 <ASau> Yes, I use algebra rather frequently.
12:47:41 <PoppaVic> KipIngram:  down and dirty does tend to show "how do I do this on any machine?"
12:48:04 <KipIngram> But using algebra frequently (which I do as well - I've manipulated equations a *lot* in my day) is an advanced thing, whereas doing arithmetic is a day-to-day thing for everyone.
12:48:18 <ASau> KipIngram: how so?
12:48:18 <KipIngram> Learn the alternative approach for the advanced thinking if you need to.
12:48:29 <KipIngram> How so what?
12:48:32 <PoppaVic> ind you, some friends the other week spanked me - I forgotted that the basic plus/minus/times, etc were all "gates" ;-) I do NOT want my head down that deep ;-))
12:49:00 <PoppaVic> KipIngram:  yeah, check-book/building maths.
12:49:03 <ASau> KipIngram: in real life arithmetic is less frequently done by humans than algebra.
12:49:33 <KipIngram> Bull.  Every time you make change, or confirm that you got the right change, or figure out how long it is until your TV show comes on, or whatever you're doing arithmetic.
12:49:43 <PoppaVic> ASau:  you are one confuzzled dude.
12:49:48 <ASau> I don't have TV set.
12:49:53 <KipIngram> How many more pounds do you need to lose?  
12:49:54 <ASau> How do I check without it?
12:49:59 <KipIngram> Then you're not representative of the mainstream.
12:50:21 <KipIngram> You don't even notice all the ways you use arithmetic.
12:50:25 * PoppaVic returns to SG:atlantis ala' hulu and his own interests
12:50:31 <ASau> When do I do it myself?
12:51:09 <KipIngram> Well, since you contentd that none of us are individuals then your own habits don't matter, do they? ;-)
12:51:15 <KipIngram> That's how we met, you'll recall.
12:51:22 <KipIngram> Couple of weeks ago.
12:51:37 * PoppaVic chuckles
12:52:01 <ASau> KipIngram: this isn't only myself.
12:52:26 <ASau> In my profession(s) we don't do arithmetic,
12:52:31 <KipIngram> Well look, if infix was simpler than rpn then computers would be built to use infix.  They aren't.
12:52:32 <ASau> in most cases it is algebra.
12:52:47 <ASau> They aren't built to use postfix either.
12:53:09 <KipIngram> It was the same for me in earlier parts of my career.  Electromagnetics research at a university research center.  Now I design circuitry for downhole, and have to actually crank the numbers.
12:53:26 <ASau> Except some special hardware like CADR or other Lisp machines
12:53:36 <KipIngram> But compiler build the expressions in postfix.  It's simpler.  Just count keystrokes.
12:53:41 <ASau> and some other special hardware like what you design.
12:54:08 <ASau> KipIngram: compiler builds expressions as AST, which is neither prefix nor postfix,
12:54:14 <ASau> or in SSA, which isn't either,
12:54:36 <KipIngram> (1 + 2) * (4 + 5) = 12 keystrokes
12:54:37 <ASau> or in _PREFIX_, which is what GCC outputs when it crashes.
12:54:56 <ASau> Key strokes are not measurement.
12:54:57 <KipIngram> 1 e 2 + 4 e 5 + * ---> 9 keystrokes
12:55:05 <KipIngram> e == ENTER
12:55:25 <ASau> How do you encode unary minus?
12:55:29 <KipIngram> 25% more efficient.
12:55:38 <KipIngram> +/- key.
12:55:41 <KipIngram> one keystroke.
12:55:51 <ASau> This conflicts with binary one.
12:55:52 <KipIngram> Same for algebraic calculators, right?
12:56:21 <ASau> Sure, but algebraic ones don't have problem with unary/binary conflict.
12:56:53 <KipIngram> Explain what you mean by that; I don't follow.
12:57:06 <ASau> "-1" is unary "-"
12:57:12 <ASau> "1-2" is binary "-"
12:57:28 <PoppaVic> KipIngram:  he's talking about the "operator" being unary/binary
12:57:32 <ASau> "-1+2" is valid, "-1-2" is valid, with no problems.
12:57:54 <ASau> "1 -" is invalid because it is underflow. :p
12:58:14 <PoppaVic> 1 negate ...
12:58:52 <ASau> Thus you have to introduce special notation for operators
12:58:52 <ASau> who have binary and unary usual meaning.
12:59:11 <ASau> 2x more symbols.
12:59:21 <KipIngram> That's a corner case.
12:59:50 <KipIngram> But forcing - to do the job of both is a lot like forcing ." to do the job of ...   Right?
12:59:51 <ASau> That's _not_ the corner case.
13:00:10 <ASau> Unary minus is regular thing.
13:00:16 <KipIngram> Negation and subtraction are completely separate concepts.
13:00:22 <KipIngram> Or at least separate operations.
13:00:30 <KipIngram> They're just not the same thing.
13:00:46 <KipIngram> Negation is a unary operator, like sin, or square root.
13:01:18 <KipIngram> Subtraction is a binary operator.  I argue that using one key for both is the strange thing.  It just happens to work *if* you happen to have 0 in your accumulator.
13:01:43 <qFox> KipIngram> - and ~
13:02:25 <KipIngram> ~ is a bitwise inversion isn't it?
13:02:28 <qFox> jseval [-1, ~1, -(-1), ~(-1)]
13:02:57 <KipIngram> Anyway, negation and subtraction are different, and if your approach lets them share a key on the calculator that's just a fluke.
13:08:37 <ASau> Subtraction is relative to negation, and there's no need to introduce separate notation.
13:08:37 <ASau> This reveals limitation of your approach.
13:08:37 <KipIngram> 25% keystroke savings in a typical calculation.  I'll take one more key on my keyboard, thanks very much.  Oh, wait, I saved the two keys for ( and ), so I'm a key ahead.
13:08:37 <KipIngram> In no way.
13:08:37 <ASau> Number key strokes doesn't matter much.
13:08:37 <KipIngram> It absolutely does - that's what determines how fast you can fly the machine.  That's what determines your productivity.
13:08:37 <qFox> oh. haha. sorry KipIngram, i was looking at the wrong chan
13:08:37 <KipIngram> For exactly such reasons emacs runs circles around other editors *once you learn to use it*.
13:08:37 <ASau> Can you read Huffman encoded English?
13:08:37 <KipIngram> We all know the answer to that, so go ahead and make your point.
13:08:37 <qFox> you do? cool!
13:08:37 <KipIngram> :-)
13:08:37 <qFox> oh and yes, ~ is a bitwise not. whereas - merely inverts the sign
13:08:37 <ASau> That English doesn't use hieroglyphs, refutes your keystroke argument.
13:08:37 <qFox> but that's in javascript, so i dont know how that applies to this discussion ;)
13:08:37 <KipIngram> I certainly do not deny that another form of written language that's more concise and efficient might exist.  I'm just not fortunate enough to know it.
13:08:37 <ASau> You can recall Mark Twain's article on improving English.
13:08:37 <KipIngram> Shorthand?
13:08:37 <KipIngram> I could write my own notes in shorthand, and then I would write faster and be able to read my own notes.
13:08:38 <ASau> Can you read them after several years?
13:08:38 <KipIngram> I don't know that I've had the good fortune of reading that one.
13:08:38 <ASau> Can another person read them as fluently?
13:08:38 <KipIngram> I wouldn't know - I don't know shortlhand.
13:08:40 <KipIngram> I'm sure not, but then again another person might not be able to read my RPN.  But that doesn't mean it's not more efficient than infix.
13:08:51 <KipIngram> It also wouldn't mean that the shorthand wasn't more efficient than English.
13:09:10 <KipIngram> My purpose is to do the arithmetic *right now* and get the answer.  I'm not logging it all for future consumption.
13:09:55 <ASau> You seem to think that particular effectivity is valued over general effectivity.
13:10:19 <ASau> Some cats run faster than dogs.
13:10:26 <ASau> But only for few seconds.
13:10:34 <ASau> In general they are much slower.
13:10:40 <KipIngram> I'm interested in my own effectivity.  As any good individual should be. ;-)
13:10:49 <KipIngram> ASau, you should read Ayn Rand.
13:11:04 <ASau> I don't care of her.
13:11:13 <KipIngram> No, I imagine not.
13:11:36 <ASau> IMO, her writings are dull and quite not interesting.
13:12:10 <KipIngram> Well, I certainly don't consider her a sparkly writer.  But her point of view on the individual is great.
13:12:18 <ASau> Your approach is flawed in the area of reusing others' works.
13:12:30 --- quit: pgas ("/quit")
13:12:59 <KipIngram> Yes, you have that point of view well entrenced in the logs from our first conversation.
13:13:21 <ASau> In many cases you can save great amount of time,
13:13:21 <ASau> if you know person to ask questions outside your domain of competence.
13:13:50 <KipIngram> I also buy my food at a grocery store instead of growing it myself.
13:13:56 <KipIngram> I don't sew my own clothes either.
13:14:20 <KipIngram> I depend on others for my survival to a postively frightening degree.
13:14:33 <KipIngram> That has nothing to do with individualism.
13:14:57 <ASau> If I come to my fellow mathematician and show her equations in my own notation,
13:14:57 <ASau> she will not look. And that's fair.
13:15:46 <ASau> When I ask it in common notation, which is INFIX, I have many chances to get answer in minutes.
13:15:53 <KipIngram> But if your notation makes you more productive (and why else would you use it) then she's missing out by not at least giving your approach a review.
13:17:01 <KipIngram> Maybe she could learn a trick or two from you and improve her own productivity.
13:17:04 <ASau> If my notation makes me more productive, I use it internally and convert to common one at I/O.
13:17:36 <ASau> She is expert in math, I'm expert in physics (let put it so).
13:17:53 <ASau> She may not understand how bozons differ from classic particles,
13:18:02 <ASau> that's my domain of expertise.
13:20:32 --- quit: Judofyr (Remote closed the connection)
13:38:56 <KipIngram> Sorry; guys came to my office.
13:40:25 <PoppaVic> LaTeX ;->
13:57:47 --- quit: Quartus` (Read error: 110 (Connection timed out))
14:23:20 --- quit: Maki ("Leaving")
14:39:16 --- quit: ygrek (Remote closed the connection)
14:48:27 --- join: kar8nga (n=kar8nga@jol13-1-82-66-176-74.fbx.proxad.net) joined #forth
14:49:40 --- quit: qFox ("Time for cookies!")
15:21:38 --- quit: GeDaMo ("Leaving.")
15:38:19 --- quit: kar8nga (Remote closed the connection)
16:31:32 --- part: ElmerJFudd left #forth
16:57:02 --- quit: ASau (Read error: 54 (Connection reset by peer))
17:04:59 --- join: krainbolt (n=krainbol@adsl-065-013-148-105.sip.msy.bellsouth.net) joined #forth
17:07:48 --- join: ASau (n=user@83.69.227.32) joined #forth
17:08:09 --- join: bogen (n=bogen@cpe-76-186-22-145.tx.res.rr.com) joined #forth
17:12:49 <ben_m> I'm trying to write a loop that uppercases every character I type in
17:12:52 <ben_m> until I press escape
17:12:56 <ben_m> : uppercase.loop begin key 27 <> while 32 - emit repeat ;
17:12:58 <ben_m> This is what I have
17:13:15 <ben_m> But when I press a key I get a stack underflow
17:13:16 <PoppaVic> uhh.. that don't sound too wise.
17:13:47 <PoppaVic> I'm pretty sure key would see an escape and wait around
17:13:47 <ben_m> I forgot a dup!
17:14:17 <ben_m> Works now :)
17:14:21 <PoppaVic> ok
17:14:41 <ben_m> What did you mean?
17:15:03 <PoppaVic> nm.. I keep thinking of scancode-hell, and - well, nm
17:15:48 <ben_m> No idea what that is
17:15:51 <ben_m> I'll nevermind :P
17:17:39 <PoppaVic> ben_m:  I'm used to assorted layers stealing the damned keys on me, and even have (painful) memories of having to read the raw scancodes from kbds. sorry.
17:17:50 <ben_m> Oh I understand
17:18:00 <ben_m> I was surprised that key is unbuffered at first
17:18:04 <ben_m> Pleasant surprise :)
17:18:20 <PoppaVic> yeah, in forth it'd not be much good if it wasn't unbuffered.
17:18:31 <PoppaVic> Interactive and other chewy-goodness ;-)
17:30:33 --- quit: krainbolt (Remote closed the connection)
17:30:45 --- join: krainbolt (n=krainbol@adsl-065-013-148-105.sip.msy.bellsouth.net) joined #forth
18:03:43 --- join: |dinya_| (n=Denis@188.18.83.251) joined #forth
18:05:59 --- quit: tathi (leguin.freenode.net irc.freenode.net)
18:05:59 --- quit: dinya_ (leguin.freenode.net irc.freenode.net)
18:05:59 --- quit: maht_ (leguin.freenode.net irc.freenode.net)
18:08:19 --- join: tathi (n=josh@dsl-216-227-91-166.fairpoint.net) joined #forth
18:09:00 --- join: dinya_ (n=Denis@188.18.83.251) joined #forth
18:09:00 --- join: maht_ (n=maht__@85.189.31.174.proweb.managedbroadband.co.uk) joined #forth
18:09:01 --- quit: dinya_ (Success)
18:50:20 --- join: ElmerJFudd (i=1000@c-68-56-142-41.hsd1.fl.comcast.net) joined #forth
18:52:15 <ElmerJFudd> forth and forth, what is forth?  it is incomplete, partially specificed, facade of semi-integrated, structure going no where fast.  the lack of support code and scaffolding prevents it from rightfully competing and beating all the scripting languages and most if not all compiled languages.
18:53:07 <ElmerJFudd> the "in-fighting" and arguing and endless discussions about saving space, saving resources are most detrimental and completely irrelevant where today's hardware and operating systems are concerned.
18:54:09 <ElmerJFudd> when are Forthers going to abandon these detriments and build up Forth to where it should be and finally put an end squabbling about irrelevant subjects?
18:54:09 <PoppaVic> it's never irrelevant.
18:54:36 <ElmerJFudd> and that response is why Forth continues to be a curiousity to the computers at large.
18:54:44 <ElmerJFudd> if they even know about it.
18:54:50 * PoppaVic shrugs
18:55:17 <ElmerJFudd> again, that too keeps forth in the dungeon, self-imposed.
18:55:52 <ElmerJFudd> did you ever finish your project to write a completely platform independent forth?
18:55:55 <PoppaVic> I'm certainly not stopping you from writing The Ultimate Forth, nor the std-commies.
18:55:56 <ElmerJFudd> or did you abandon it?
18:56:22 <PoppaVic> it wasn't really a forth, and I totally gave it up, except as a mine of routines/structs and notes.
18:56:29 <ElmerJFudd> figers
18:56:49 <ElmerJFudd> you never wanted to compromise and just write it.
18:57:04 <ElmerJFudd> you got stuck in minutia that you refused to surpass.
18:57:08 <PoppaVic> Got tired of the arguing; decided there was no point.
18:57:15 <ElmerJFudd> self-defeating.
18:57:52 <ElmerJFudd> i'm not a system writer; but i am an observer; and i've been observing this phenomena for years. 
18:58:50 <ElmerJFudd> ANS-forth spec is a detriment to being able to accomplish; people/coders dont need a standard to write so-called portable code. agree on the basics and move on.
18:59:09 <PoppaVic> They can't. They won't.
18:59:25 <ElmerJFudd> agree on reflexive words for inspecting the host and move on.
18:59:41 <PoppaVic> They can't and won't <repeat enabled>
18:59:43 <ElmerJFudd> right, so why dont YOU buck the trend and do so?  it just takes one to make a difference
19:00:01 <PoppaVic> Look at the earlier discussion of infix vs rpn <shrug> So be it.
19:00:46 <ElmerJFudd> being stubborn is a pride. just write the system that can compete.
19:00:53 <PoppaVic> why?
19:00:57 <ElmerJFudd> forth deserve so much more than it has.
19:01:01 <PoppaVic> Feel free, I'll not stop you.
19:01:11 <ElmerJFudd> i told you, i'm not a system writer.
19:01:37 <PoppaVic> almost all platforms already have "a system"
19:02:03 <ElmerJFudd> i'm referring to perl, python, ruby, etal as "systems"
19:02:06 <ElmerJFudd> scripting languages.
19:02:20 <ElmerJFudd> forth can compete with them if it had the scaffolding required.
19:02:56 <PoppaVic> Yes? So..? What is "the scaffolding"? Even asking that is a waste of time.
19:03:13 <ElmerJFudd> with today's multi-gigabyte (ram,hd) none of the arguments, mental and vocal, matter to prevent 
19:03:27 <PoppaVic> good, then yer all set.
19:03:28 <ElmerJFudd> you are such and still a bastard; you prick.
19:03:34 <ElmerJFudd> get over yourself.
19:03:34 <PoppaVic> indeed
19:03:45 <ElmerJFudd> ignored. go sulk in your self-pity.
19:03:55 <ben_m> This channel is ultra friendly.
19:04:15 <ElmerJFudd> ben, he's always been that way
19:05:47 <PoppaVic> ben_m:  love and warm fuzzies for sure ;-)
19:05:56 <ben_m> Just how I love it
19:06:40 <PoppaVic> ben_m:  I prefer my Solyent Green nice and quiet.
19:09:14 <tathi> Ya, we get crazies like ElmerJFudd occasionally.
19:09:27 <tathi> On top of our usual opinionated regulars. :)
19:11:31 <ben_m> i've noticed :d
19:18:20 <tathi> Like I said, if you want constructive community, try Factor. :)
19:20:29 <tathi> Speaking of which, have you checked out #concatenative?
19:26:03 --- join: Snoopy_1711 (i=Snoopy_1@dslb-084-059-206-044.pools.arcor-ip.net) joined #forth
19:31:29 --- quit: maht_ (leguin.freenode.net irc.freenode.net)
19:35:49 --- join: maht_ (n=maht__@85.189.31.174.proweb.managedbroadband.co.uk) joined #forth
19:43:23 --- quit: Snoopy_1611 (Read error: 110 (Connection timed out))
19:54:57 --- quit: tathi ("leaving")
20:26:39 <ben_m> How would I stop ; from being called when using it in a definition itself?
20:27:27 <ben_m> I'm trying to write something that allows me to write :: 10 0 do foo loop ;;
20:27:40 <ben_m> Instead of :noname 10 0 do foo loop ; execute
20:31:54 <PoppaVic> ben_m:  perhaps you are looking at the problem poorly.
20:32:08 <ben_m> That's why I wrote what I want to achieve too :)
20:32:15 <ben_m> Because I thought that I am.
20:32:38 <PoppaVic> : swoop ( cfa upper lower --- ) ... ;
20:32:53 <ben_m> um?
20:33:06 <PoppaVic> as in ' foo 10 0 swoop
20:33:43 <ben_m> does not compute
20:33:44 <PoppaVic> or, swoop ( "name" upper lower -- )
20:33:50 <ben_m> What would that do?
20:34:17 <ben_m> I'm just trying to write something that defines and calls a noname definition
20:34:24 <PoppaVic> that's just a comment/signature for a func. leave the executable func outside, ticked - the upper and lower do limits.
20:34:30 <ben_m> So I can play with conditions and loops without verbosity
20:37:09 <ben_m> I want to remove the :noname ... ; execute verbosity
20:37:14 <ben_m> or boilerplate
20:37:18 <ben_m> Whatever you want to call it :)
20:37:57 --- join: foxes (i=flash@222.131.163.19) joined #forth
20:45:23 --- join: _dinya__ (n=Denis@188.17.90.6) joined #forth
20:49:28 <PoppaVic> ben_m:  defer FOO   : ben_b ( top start -- ) do foo loop ;
20:49:51 <PoppaVic> ' .s is foo
20:50:52 <PoppaVic> 10 0 ben_b
20:51:14 <ben_m> What does defer do?
20:51:37 <PoppaVic> gforth supports it, try it - 30 sec wasted?
20:51:41 --- quit: |dinya_| (Read error: 60 (Operation timed out))
20:51:42 <ben_m> I did try
20:51:46 <ben_m> Doesn't tell me anything.
20:52:01 --- join: forther (i=62d2faca@gateway/web/freenode/x-iklxmocyvvhjpkei) joined #forth
20:52:09 <ben_m> defer foo ok
20:52:09 <forther> hi all
20:52:14 <ben_m> wow.
20:52:17 <ben_m> forther: Hi :)
20:53:37 <forther> tomorrow is Forth Day
20:59:21 <PoppaVic> ben_m:  did you try foo? did you "see" defer OR foo? did you check their idiotic info file?
20:59:34 <ben_m> Reading the info file as we speak
20:59:40 <PoppaVic> cool
20:59:45 <ben_m> see doesn't help :D
20:59:49 <PoppaVic> other than a book, yer stuck with that
21:00:24 <PoppaVic> ben_m:  it does if you can guesstimate forth an a ton of historical (hysterical) arguments, debates and ideas.
21:00:33 <PoppaVic> an/and
21:01:54 * madgarden is away: aseepin'
21:02:43 --- quit: forther (Ping timeout: 180 seconds)
21:06:14 <PoppaVic> <0> <0> 
21:06:22 <PoppaVic> /exec -o /sw/bin/gforth -e "defer foo  : ben_m do foo loop ; ' .s is foo  2 0 ben_m  bye"
21:07:23 <ben_m> Got it
21:07:32 <ben_m> But that doesn't do what I wanted really :D
21:07:44 <PoppaVic> of course, it wouldn't be.
21:09:29 <PoppaVic> 3 0 [do] .s [loop] <0> <0> <0>  ok
21:09:39 <PoppaVic> see [do]
21:09:40 <ben_m> Oh, nice.
21:10:22 <PoppaVic> back to Sanctuary and ##C ;-)
21:21:02 <PoppaVic> ben_m:  that pm might be useful
21:21:43 <ben_m> reading
21:23:10 <PoppaVic> ben_m:  well, I wrote most of it because I do a lot of base-bashing for others and myself. the fail/warn shit goes well with the [if] and [do] crap.
21:23:33 <ben_m> It looks useful, but the implementation is too advanced for me to understand.
21:23:58 <PoppaVic> so'k, grab it, rename to something.fs and tinker as you will.
21:24:23 <ben_m> http://www.complang.tuwien.ac.at/forth/gforth/Docs-html/User_002ddefined-Defining-Words.html#User_002ddefined-Defining-Words
21:24:26 <ben_m> This looks interesting.
21:25:02 <PoppaVic> sure.. w/o defining words - compiler-words - it's pretty useless.
21:30:32 --- quit: segher ("This computer has gone to sleep")
21:37:12 <probonono> What benefit is derived from defining 'noname' words? Is there more to it than simply reducing verbosity?
21:37:49 <PoppaVic> hidden routines - like C 'static' funcs
21:38:19 <PoppaVic> beyond that? no idea why
21:49:48 --- join: |dinya_| (n=Denis@188.17.90.6) joined #forth
21:52:09 --- quit: _dinya__ (leguin.freenode.net irc.freenode.net)
21:52:09 --- quit: maht_ (leguin.freenode.net irc.freenode.net)
21:57:30 --- join: maht_ (n=maht__@85-189-31-174.proweb.managedbroadband.co.uk) joined #forth
22:02:10 <I440r> probonono, it helps keep the name space sane. if every word you define in a module is visible your going to have alot of cruft in your "words" output. it also means you have to search a lot more headers during compilation
22:02:34 <I440r> also.  you can make all words other than the ones you want exposed headerless in a module
22:02:45 <I440r> as poppavic suggested
22:03:01 <I440r> thats the least important reason however
22:04:02 <PoppaVic> I suspect it's mostly an issue of the headers/dict space
22:04:59 <I440r> well every visible header you define in a vocabulary increases the time it takes to search that vocabulary
22:05:09 <PoppaVic> yep
22:05:37 --- quit: maht_ (Killed by sagan.freenode.net (Nick collision))
22:05:40 --- join: _dinya__ (n=Denis@188.17.90.6) joined #forth
22:05:40 --- join: maht_ (n=maht__@85.189.31.174.proweb.managedbroadband.co.uk) joined #forth
22:05:41 --- quit: _dinya__ (Success)
22:06:00 --- join: maht (n=maht__@85-189-31-174.proweb.managedbroadband.co.uk) joined #forth
22:11:10 --- quit: maht_ (Connection timed out)
23:11:28 --- join: adu (n=ajr@pool-71-191-173-164.washdc.fios.verizon.net) joined #forth
23:24:41 --- join: ygrek (i=user@gateway/gpg-tor/key-0x708D5A0C) joined #forth
23:25:05 --- quit: kleinjt (Remote closed the connection)
23:40:05 --- join: kleinjt (n=kleinjt@tarsonis.dhcp.rose-hulman.edu) joined #forth
23:42:40 <PoppaVic> Disembodied Head for all.
23:43:24 --- join: qFox (n=C00K13S@5356B263.cable.casema.nl) joined #forth
23:56:12 --- join: nighty^ (n=nighty@x122091.ppp.asahi-net.or.jp) joined #forth
23:59:59 --- log: ended forth/09.11.20