00:36:03 moebious joined #tunes
00:36:08 <moebious> yo
03:36:15 pratap joined #tunes
03:43:13 [QUIT] pratap quit: ircII EPIC4-2000 -- Accept no limitations
04:13:45 [QUIT] karltk quit: Ping timeout for karltk[janus.prosalg.no]
04:15:01 karltk joined #tunes
04:15:54 [QUIT] ink[sleep] quit: Ping timeout for ink[sleep][ppp-207-214-214-136.sntc01.pacbell.net]
04:50:53 water joined #tunes
04:51:22 water is on for just a couple of minutes
05:01:08 [QUIT] water quit: The Tao went that-a-way!
06:41:37 Fare joined #tunes
07:13:56 [QUIT] JALH quit: [BX] Pretzel Boy uses BitchX. Shouldn't you?
07:17:58 amokeins joined #tunes
07:18:47 <amokeins> hey anybody out there ?
07:18:51 <Fare> kind of
07:19:13 <amokeins> kind of what ?
07:19:40 [QUIT] amokeins quit: amokeins has no reason
07:20:07 Rhamphoryncus joined #tunes
07:20:24 <Fare> Rham: hello
07:20:29 <Rhamphoryncus> hiya
07:20:33 <Fare> so you want to make a new language?
07:20:38 <Rhamphoryncus> yeah
07:21:01 <Fare> learning about existing ones is so as to copy what they did well and avoid what they did wrong
07:21:22 <Fare> in the case of pliant, the open compiler design is great.
07:21:24 <Rhamphoryncus> oh, I think I'm well beyond that point :)
07:21:30 <Fare> About what LISP should have been
07:21:39 <Fare> Rham: what point are you at?
07:22:09 <Rhamphoryncus> Fare: if I sat my ass down and typed stuff up, I could have the compiler running probably within a week.  (I'm too lazy though)
07:23:25 <Rhamphoryncus> ewwwwww!   "PLiant is a true procedural environment (plenty of side effects, no garbage collector) because it is expected to be even faster than C."
07:23:51 <Fare> what kind of general design are you using?
07:23:58 <Rhamphoryncus> I expect mine to be faster than C for exactly the opposite reasons.  no side effects. :)
07:24:08 <Rhamphoryncus> for which aspects?
07:24:30 <Fare> well, the fact you have a pure language is interesting, for instance.
07:24:47 <Fare> How are you different from, say, Mercury, Haskell, or SAC?
07:25:54 <Rhamphoryncus> don't know about mercury or SAC, but the fact that I have no syntax has to be worth something :)  (ok, so technically there is one, but...)
07:25:56 <Fare> what evaluation model are you using?
07:26:13 <Rhamphoryncus> ?
07:26:33 <Fare> well, for instance is it a lazy functional setting as with haskell?
07:26:44 <Rhamphoryncus> sorry, I have trouble with terms.  I dont' know the ones other people use, and I chose oens that confuse the hell out of other people looking at mine
07:27:52 <Rhamphoryncus> well, as I understand it lazy means that it only evaluates when it's required too.  my language will do this, but it will be hidden from the programer; as far as he's concerned it's not lazy.
07:28:10 <Rhamphoryncus> (that's just a part of how it'll do optimizations)
07:28:41 <Fare> btw, where are you from (physically, academically, hackerly) ?
07:29:18 <Rhamphoryncus> I'm from edmonton alberta, have about grade 7 in school, and hackerly... not sure what to put for that :)
07:30:20 <Fare> grade 7 ?
07:30:45 <Rhamphoryncus> dropped out a while ago.  (technically I was there longer, but I don't consider that since I didn't get anything done)
07:30:59 <Fare> (I'm no canadian, so I don't know what grades mean exactly)
07:31:12 <Rhamphoryncus> 7 years of schooling
07:31:18 <Rhamphoryncus> where are you from?
07:31:23 <Fare> Paris, France.
07:31:36 <Rhamphoryncus> hmm
07:31:41 <Fare> of schooling, from what state of knowledge/age/foo ?
07:32:05 <Rhamphoryncus> well I'm 17 now, and that was several years ago, if that gives any indication
07:32:50 <Fare> ok. Well, apparently, it doesn't prevent your communicating like a well bred student.
07:32:58 <Rhamphoryncus> *grin*
07:33:44 <Fare> anyway. One thing about the language is the programming model it offers, as visible to programmers
07:34:23 [QUIT] smkl quit: Ping timeout for smkl[glubimox.yok.utu.fi]
07:34:26 <Rhamphoryncus> well in some ways my writing skills are extremely bad (fine in others), so I've been trying for th elast week to get some reasonable documentation written, but just not getting anywhere.  (I'm sure any successful student can push something out, but I can't)
07:34:31 smkl joined #tunes
07:34:38 <Rhamphoryncus> well I should explain the syntax thing to you
07:35:12 <Fare> then, depending on whether you're doing whole-program optimization, modular compilation, or interactive development, the model has increasing pervasiveness in the structural representation of your language
07:35:25 <Rhamphoryncus> rather that try to create some powerful and flexible syntax, that may not actually be all that powerful, I skipped it.  I just provide an API.
07:36:02 eihrul joined #tunes
07:36:06 <Fare> Also, the syntax is largely decouplable from the evaluation model
07:36:15 <eihrul> 'lo
07:36:34 <Rhamphoryncus> you know, I MOSTLY understand what you said :)
07:36:42 <Fare> (for instance, Caml has had many changes in syntax, and there is a preprocessor-prettyprinter that helps going from one syntax to another)
07:36:58 <Fare> eih: hi'
07:37:54 <Rhamphoryncus> well the API is about as extreme as you get there.  you are expected to create a parser (which is a language in itself) to impliment the written form (and all the features)
07:39:58 <Rhamphoryncus> at this point the API will be limited to just create the tree structure in memory, and triggering compilation.
07:40:25 <Rhamphoryncus> also, I'm limiting the i/o just stdin, stdout, and stderr.  easier than coming up with some fancy library.
07:40:52 <Fare> that's the way to do it
07:41:15 <Fare> except that the API in your initial bootstrap step is unclear
07:41:32 <Fare> (what metalanguage?)
07:41:34 <Rhamphoryncus> yeah.  I'm far more concerned with making the optimization work than I am with making some fancy syntax
07:41:49 <Rhamphoryncus> meaning?
07:41:52 <Fare> then you make as well keep the syntax of your metalanguage
07:42:10 <Fare> meaning the language in which you write your initial compiler
07:42:36 <Rhamphoryncus> ahh
07:42:44 <Rhamphoryncus> I'm writing the bootstrap compiler in C
07:43:20 <Rhamphoryncus> It'll be an extremely simplistic and innefficeient thing, but no biggie :)
07:43:31 <Fare> ~`:
07:43:44 <Rhamphoryncus> wha?
07:44:33 <eihrul> Rhamphoryncus: eliminating side-effects from your language doesn't make it more efficient
07:44:44 <eihrul> it only means that the burden of introducing side-effects is moved to the compiler
07:44:52 <eihrul> and that side-effects become better controlled...
07:44:57 <Rhamphoryncus> eihrul: yup.  exactly :)
07:45:01 <eihrul> inevitably, when you optimize, you have to deal with them
07:45:06 <Rhamphoryncus> sure
07:45:13 <eihrul> whether the programmer puts them there or the compiler
07:45:18 <Rhamphoryncus> it just gives you ALOT more flexibily on how to deal with them
07:45:24 <eihrul> but not allowing side-effects also leads to other interesting problems
07:45:49 <eihrul> like expressing input/output (which are side-effects) and some other things
07:46:08 <Fare> Rham: have you seen how Haskell deals with expressing side-effects?
07:46:14 <Rhamphoryncus> Fare: nope
07:46:14 <Fare> (Mercury does the same)
07:46:23 <Rhamphoryncus> well I think I have, but I don't remember
07:46:30 <eihrul> monads, eeew :)
07:46:35 eihrul likes linearity.
07:46:43 <Rhamphoryncus> same
07:46:48 <Rhamphoryncus> I absolutely hated haskel
07:46:54 <eihrul> haskell is lovely
07:47:04 <Fare> eih: linearity is essentially the same as monads
07:47:18 <Rhamphoryncus> no, my language is lovely.  disgusting too, but you can writing a parser to hide all that :)
07:47:18 <Fare> (or rather as concurrent monads)
07:47:19 <eihrul> Fare: i realize this, but linearity is a more natural way to express it :)
07:47:36 <Fare> eih: I'd say they are dual approaches
07:47:41 <eihrul> Rhampho: at this point, that's a claim, i haven't seen your language :)
07:47:54 <Fare> one (linearity) focusing on space structure, while the other (monads) focuses on time structure
07:48:57 <Fare> Rham: yes, if you see Haskell's monads, you might also see the linear Clean way.
07:49:50 <Fare> Rham: side-effect-freeness has also been used in designing SISAL (and now SAC), or in writing compilers (even C compilers)
07:51:12 <Rhamphoryncus> you could in theory write a parser for C for use with my language
07:51:14 <Fare> anyway, let that not prevent you from experimenting with things yourself
07:51:40 <Rhamphoryncus> )
07:52:31 <Fare> it's just that the day you want to step back and realize what is done, what remains to do, where to go, and where to not go, you ought to fit other people's past experiences in the picture.
07:52:45 Fare is much to way back
07:52:53 Fare needs to step forward
07:52:58 <Rhamphoryncus> well if I ever do get this to do some of the stuff I want it to, I'm sure it should cause a pretty huge splash :)
07:53:18 <Fare> what prompted you in the pure direction?
07:53:21 <Rhamphoryncus> (something about merging logic and formulas...)
07:53:33 <Rhamphoryncus> *shrug* just fit
07:53:36 <Fare> (functional language people tend to use "pure" for "side-effect free")
07:54:10 <Rhamphoryncus> I wanted it to be free of side effects so that the compiler could have the most possible liberty about how it would produce the output code
07:54:32 <Fare> sure. I mean, how did you come to realize this?
07:54:38 <Rhamphoryncus> hard to say
07:55:02 <Rhamphoryncus> you could research the message boards on i-lang.sourceforge.net, as that's where most of this spawned from
07:55:19 <Rhamphoryncus> that's a different lang though, and somewhat dead at the moment
07:56:01 <eihrul> the compiler doesn't have liberty past a point :)
07:56:21 <eihrul> sure, the ideal that your compiler never has to mess with side-effects is nice and pretty and appealing
07:56:38 <eihrul> but its not true... if you want to optimize for a side-effectual processor, you need to eventually optimize and deal with side-effects :)
07:57:27 <Rhamphoryncus> well *EVERYTHING* in the language proper is known.  when you do something, it's supposed to do that.  doesn't that make it free of side effects?
07:57:58 <eihrul> its almost as hard to decide where to introduce side-effects as it is to discover where they are in some code
07:58:53 <Rhamphoryncus> well if I get the Guarantees working it should cut down on problems too :)
07:59:11 <Fare> Rham: do you have a body of test applications in mind for your language?
07:59:23 <Fare> Also, what target language will you be using?
07:59:53 <Fare> (for instance, Stalin dumps asm-like C code from wholly-optimized Scheme programs)
08:00:46 <Rhamphoryncus> Fare: I'll mostly just write demo programs to test out features and optimization.  I've got no plans for any particular programs at this point
08:01:35 <Fare> do you have some goal in mind, like, optimizing better than a C compiler for such or such processor on your test programs?
08:01:42 <Rhamphoryncus> well it'll eventually put out assembly, but I'll be done some cheasy stuff with output C at first and compiling it into assembly to save me the effort of learning it properly :)
08:02:02 <Rhamphoryncus> "optimizing better than *"
08:02:07 <Rhamphoryncus> that about covers it :)
08:02:43 <Fare> the verb "to optimize" must have been propagated by a marketroid.
08:02:53 <Rhamphoryncus> hehehehe
08:03:06 <Rhamphoryncus> what's the word I'm looking for....
08:03:18 <Rhamphoryncus> canonical
08:03:19 <Rhamphoryncus> that's it :)
08:03:20 <eihrul> Fare: nah, "optimizing better than *" is propagated by marketroids
08:03:54 <Rhamphoryncus> I want to produce the canonical form of programs, meaning the absolute best possible.
08:03:54 <Fare> most everything is propagated by marketroids, when you think about it. Scary.
08:03:56 <eihrul> Fare: optimization in itself is a valid pursuit
08:04:09 <eihrul> but "optimizing better than *" doesn't make sense
08:04:22 <Rhamphoryncus> eihrul: true, but it let me copy and paste :)
08:04:24 <eihrul> as various kinds of optimizations are done for various "*"
08:04:34 <eihrul> which yield very different results
08:04:39 <Fare> Rham: the absolute canonical form is not computable in general, and it will depend a lot on the cost model.
08:04:56 <Rhamphoryncus> Fare: true.  but I oughta be able to get pretty close :)
08:05:29 <Rhamphoryncus> and given some hard parameters, it is computable
08:05:50 <Fare> what worries me most is the cost model.
08:05:57 <Rhamphoryncus> cost?
08:06:10 <Fare> You optimize wrt some notion of cost
08:06:16 <Rhamphoryncus> ahh
08:06:48 <Fare> And then comes the question of the adequacy of your notion of cost to the needs and expectations of users
08:06:53 <Rhamphoryncus> well at first it'll just be making optimizations, not worrying about tradeoffs.  then later I may add some stuff to do tradeoffs (may be much later though)
08:07:26 <Rhamphoryncus> yeah, well, I'm sure they'll appreciate a system that's been mathematically proven to not fail :)
08:07:45 <Fare> Beware: a mathematical proof is difficult
08:07:50 <Rhamphoryncus> true
08:07:58 <Fare> all the more when you want to prove not only the design, but the implementation
08:08:07 <Rhamphoryncus> but it's a worthy challenge
08:08:19 <Rhamphoryncus> well it depends on how you define "fail"
08:08:29 <Fare> all the more when the implementation relies on extremely complex hardware ISA (like the X86)
08:08:41 <Rhamphoryncus> if you can write a test for it, then it can prove it.
08:08:59 <Rhamphoryncus> hehe, as I just said :)
08:09:01 hcf joined #tunes
08:09:03 <Fare> don't overestimate the difficulty of proof
08:09:13 <Rhamphoryncus> I won't
08:09:58 <Rhamphoryncus> the proofs wouldn't extend to how the i/o operates.  that is external and may any result at all.
08:11:03 <Rhamphoryncus> but you can prove how each language is written internally, and the same techniques could be adapted to test the hardware itself
08:11:07 <Rhamphoryncus> oops
08:11:11 <Rhamphoryncus> s/language/program/
08:11:33 <Fare> anyway. I see you already have enough projects to fill up your whole life :)
08:11:49 <Rhamphoryncus> hehehe
08:11:53 <Fare> and it's refreshing to see youngsters who are not apathetic.
08:11:56 <Rhamphoryncus> I doubt it'll take that long
08:12:18 <Rhamphoryncus> yeah, most people just tell me to use C, or something equally lame.... *shakes head*
08:12:34 <Fare> Well, depends how deep results you're after. There will always be ways to deepen your results.
08:12:44 <Rhamphoryncus> (and I mean that telling me was lame, not C.  even though it is too ;) )
08:12:57 <Rhamphoryncus> I want to go all the way
08:13:11 <Fare> Then you'll have to go on after life :)
08:13:17 <Rhamphoryncus> hehehe
08:13:22 <Rhamphoryncus> that's another project ;)
08:13:38 <Fare> Well, nice to see you here, anyway.
08:13:47 <Fare> Keep us tuned.
08:14:32 <Rhamphoryncus> just find some way to "freeze" your brain, and maybe your whole body.  then you dissasemble it piece by piece, and load it into the molecular emulator running on your computer.  walla, you can now debug your brain :)
08:15:57 <Fare> You need a _big_ and _fast_ computer to make that run.
08:16:27 <Rhamphoryncus> nah.  you just won't notice the lag on your windows box :)
08:16:29 <Fare> or you may optimize by compiling the result into carbon, oxygen, hydrogen and nitrogen molecules
08:16:42 <Rhamphoryncus> hehehe
08:21:00 <eihrul> sure, but optimizing the mind would be especially tricky :P
08:21:07 <Rhamphoryncus> )
08:21:15 [QUIT] gREMLiNs quit: Ping timeout for gREMLiNs[as36-01-66-56.urc.bl.ac.yu]
08:21:17 <Rhamphoryncus> yeah, but by that point it should be a cake walk
08:21:40 <eihrul> well, you'd have to make several minds
08:21:46 <eihrul> all specialized for particular tasks
08:21:53 <eihrul> or else you'd have a mind that performs poorly at everything
08:22:37 <eihrul> or at the very least, an adaptive mind that reorganizes itself based on the situation :)
08:22:55 [QUIT] eihrul quit: Leaving
08:27:23 <Fare> my remark about compiling to atoms was ironical: the best way to execute your "frozen" image of a brain is to build a chemical brain that conforms to said frozen image.
08:27:45 <Rhamphoryncus> like I said, a molecular emulator
08:28:13 <Fare> hcf: oh, yes I should talk about this lisp + source-matching system I'm meant to be developing
08:28:19 <Fare> but then I ought to develop it
08:28:40 <Fare> Rham: yeah; molecules emulate themselves pretty well
08:29:08 <Rhamphoryncus> Fare: it's that external interference thing
08:31:14 <Fare> well, isolate your machine with other chemical devices. Like "clothes", "walls", and the like :)
08:31:22 <Rhamphoryncus> hehehe
08:31:56 <Fare> Oops. I gotta go.
08:32:11 <Fare> hcf: we'll talk about that when there's people interested about it in the channel.
08:32:33 [QUIT] Fare quit: Connection reset by pear
09:39:02 eihrul joined #tunes
10:00:28 [QUIT] JALH quit: brb!
10:27:08 hcf joined #tunes
10:37:09 [QUIT] kc5tja quit: BitchX: its magically delicious!
10:41:18 [QUIT] hcf quit: Ping timeout for hcf[207-172-225-147.s147.tnt1.pld.me.dialup.rcn.com]
10:42:40 cor[wk] joined #tunes
11:03:04 [QUIT] JALH quit: brb... AOL sux0rs
11:08:56 smkl joined #tunes
11:19:56 [QUIT] nate37 quit: brb -- moving computer upstairs (yay)
11:42:05 eihrul joined #tunes
11:42:27 nate37 joined #tunes
11:54:55 [QUIT] gREMLiNs quit: Ping timeout for gREMLiNs[195.252.64.36]
11:54:58 [QUIT] JALH quit: Ping timeout for JALH[host62-6-83-104.btinternet.com]
12:34:31 [QUIT] gREMLiNs quit: You must die to go to heaven
15:26:14 [QUIT] kintaro quit: Ping timeout for kintaro[dial-nas-nahr-2-243.inco.com.lb]
15:26:49 [QUIT] kintaro1 quit: Ping timeout for kintaro1[dial-nas-nahr-2-243.inco.com.lb]
15:32:07 [QUIT] kintaro quit: Ping timeout for kintaro[dial-nas-nahr-2-104.inco.com.lb]
15:34:29 eihrul joined #tunes
15:35:34 <nate37> re eihrul 
15:54:17 [QUIT] kintaro quit: Ping timeout for kintaro[194.126.19.15]
16:22:08 [QUIT] eihrul quit: Leaving
16:49:21 Fare joined #tunes
16:49:51 Fare is passing by
17:19:34 <Kyle> Who's here?
17:22:54 eihrul joined #tunes
17:23:07 <Kyle> re
17:23:17 <eihrul> er
17:23:44 <Kyle> I read the discussion with Rham.
17:24:08 <Kyle> Can you give me an example of a side-effect?
17:24:35 <eihrul> some effect that is visible outside of the scope of the thing causing it :)
17:24:43 <eihrul> i.e. a global variable getting set by a procedure
17:25:22 <Kyle> Is that everything that is not pure functional?
17:25:45 <eihrul> not necessarily
17:25:57 <eihrul> i.e. there could be side-effects within the function that are never visible outside of it
17:26:09 <eihrul> so from outside the function, it looks like there are no side-effects
17:26:13 <eihrul> within the scope of the function, there are though
17:26:21 <eihrul> is it still purely functional? and according to whom? :)
17:27:50 <Kyle> then it is pure functional, but non functional subcomponents can make a pure function.
17:29:13 <Kyle> did that make sense?
17:29:28 <Fare> Kyle: beware that this depends on the possible observations
17:30:03 <Fare> if some concurrent monitor can inspect the internals of the function, it may discover the impurity
17:30:28 <Fare> as long as the monitor is considered meta-, that's no problem
17:30:41 <Fare> but then it can be used at same level, and then you're in trouble
17:31:17 <Kyle> SO I just have to keep true to the definition of "visible outside".
17:31:33 <Fare> for instance, in fscking Scheme, anyone can replace a global function with something that intercepts and escapes continuations
17:32:42 <Fare> so that (define (foo x) ... (foo (next x))) is no more equivalent to (define (foo x) (do ...  (set! x (next x))  ...)
17:33:06 <Fare> i.e. escaping non-linear continuations can reveal internal impurity.
17:33:28 <Fare> Kyle: so you have to clearly define "outside"
17:35:29 <Kyle> Do either of you have a good suggestion of a language that handles side-effects, instead of trying to hide them?
17:35:45 <eihrul> most imperative languages don't try to hide side-effects :)
17:35:51 <eihrul> its just they don't deal with them elegantly
17:36:03 <eihrul> i will, however, plug Aardappel, as a languages that handles them nicely :)
17:36:56 <Fare> Kyle: Concurrent Haskell ?
17:37:09 <Fare> Clean ?
17:37:15 <Kyle> Is that differnt than Haskell?
17:37:22 <eihrul> abi: aardappel?
17:37:22 <abi> aardappel is a Concurrent Tree Space Transformation Language. at http://www.ecs.soton.ac.uk/~wvo96r/aardappel/index.html or Wouter van Oortmerssen's nickname (see wouter) or dutch for potato
17:37:31 <Fare> Some people have experimented with Side-effect typing, too.
17:37:45 <Fare> gotta go 3053
17:37:56 <eihrul> Kyle: look at that :)
17:38:19 <Kyle> thanks to you both
17:45:06 [NICK] lar[sleep] changed nick to: lar1
17:53:02 [QUIT] kintaro quit: Ping timeout for kintaro[194.126.19.156]
18:26:07 [QUIT] kikikiki quit: Ping timeout for kikikiki[dial-pm-nahr-1-4.inco.com.lb]
18:53:38 [QUIT] kintaro quit: Ping timeout for kintaro[dial-2-14.inco.com.lb]
19:10:34 [QUIT] kintaro quit: Ping timeout for kintaro[dial-pm-nahr-1-7.inco.com.lb]
19:23:00 [QUIT] kino1 quit: Ping timeout for kino1[dial-nas-nahr-2-54.inco.com.lb]
19:26:48 Rhamph-Sleeping starts to wonder if his definition of side-effects is screwed up
19:27:26 <eihrul> why, what is your definition of side-effect?
19:28:06 <Rhamph-Sleeping> basically stuff you didnt' expect it to do (which seems really vague)
19:28:23 [NICK] Rhamph-Sleeping changed nick to: Rhamphoryncus
19:28:44 <eihrul> what do you mean?
19:29:18 <Rhamphoryncus> well as I said, I think my defintion of side-effects is screwed up :)
19:30:16 <eihrul> well, what is your definition of side-effect?
19:30:34 <Rhamphoryncus>  <Rhamph-Sleeping> basically stuff you didnt' expect it to do (which seems really vague)
19:30:49 <eihrul> oh, i interpreted that wrong :)
19:30:56 <Rhamphoryncus> hehe
19:31:03 <eihrul> but if you look at your definition
19:31:09 <eihrul> what you expect it to do is inherently related to scope
19:31:16 <eihrul> and within what scope you expect it to be doing things
19:31:18 <Rhamphoryncus> well in hindsight, I think my language DOES allow side effects according to the normal definition
19:31:22 <Rhamphoryncus> yeah
19:52:40 Rhamphoryncus wonders if he should add #Tunes to his autojoin list
20:24:52 moebious left #tunes
21:17:41 [NICK] lar1 changed nick to: lar[dillydally]
21:51:41 [NICK] lar[dillydally] changed nick to: lar1
21:52:03 [NICK] ink[food] changed nick to: ink
22:37:34 [QUIT] eihrul quit: Leaving
22:39:01 ink joined #tunes
22:40:40 <nate37> splorot ink
22:50:26 <ink> ummmm
22:50:27 <ink> )
23:25:30 [NICK] cor[wk] changed nick to: cor[sl]
23:28:56 [QUIT] Kyle quit: Ping timeout for Kyle[cr168790-a.nmkt1.on.wave.home.com]
23:48:07 [NICK] ink changed nick to: ink[sleep]
23:54:31 [NICK] lar1 changed nick to: lar[sleep]
00:23:28 [QUIT] pratap quit: ircII EPIC4-2000 -- Accept no limitations