Arc Forumnew | comments | leaders | submit | suzuki's commentslogin
3 points by suzuki 299 days ago | link | parent | on: Semi-Arc with first-class continuations

This arc seems faster than rainbow in start-up, but slower in execution. On MacBook Air 2019, for N-queens puzzle (N = 9 and 11),

  $ time java -jar arc.jar ~/tmp/arc-in-java-0.2.0/9queens.arc 
  real	0m1.634s
  user	0m3.056s
  sys	0m0.159s

  $ time java -jar arc.jar ~/tmp/11queens.arc 

  real	0m27.865s
  user	0m29.582s
  sys	0m0.242s

  $ time java -jar rainbow.jar -q -f ~/tmp/arc-in-java-0.2.0/9queens.arc 
  *** redefining no
  *** redefining map1
  *** redefining pr
  *** redefining list
  real	0m2.279s
  user	0m6.492s
  sys	0m0.246s

  $ time java -jar rainbow.jar -q -f ~/tmp/11queens.arc 
  *** redefining no
  *** redefining map1
  *** redefining pr
  *** redefining list
  real	0m9.812s
  user	0m14.462s
  sys	0m0.360s

  $ java -version
  openjdk version "11.0.7" 2020-04-14
  OpenJDK Runtime Environment AdoptOpenJDK (build 11.0.7+10)
  OpenJDK 64-Bit Server VM AdoptOpenJDK (build 11.0.7+10, mixed mode)
As for the code size, this arc's arc folder has 4,626 lines of Java 11 while rainbow's src/java/rainbow folder has 14,520 lines (72,190 lines including auto-generated ones) of Java 5.

In short, this arc is a small-scale implementation of Arc.


2 points by rocketnia 298 days ago | link

It's great to see you again! Sometimes I've wondered about the status of Semi-Arc. Looks like you've been working on a number of other Lisp implementations in the meantime! Pretty exciting. Thanks for sharing this update with us.

As someone who ported Rainbow line by line to JavaScript, I have to say, the size of the codebase can be quite daunting. A smaller implementation sounds a lot easier to work on.


5 points by suzuki 298 days ago | link

It is my pleasure.

As I referred in, the implementation of this arc's continuations is based on If you are planning to port this arc to JavaScript, I suggest reading which implements the continuations in the same way in TypeScript.

And in the latter, the display function

  c('display', 1, x => {
      write(stringify(fst(x), false));
      return new Promise(resolve => {
          runOnNextLoop(() => resolve(None));
is applied asynchronously, in a sense, as follows:

  case ContOp.ApplyFun: // exp2 is a function.
      [exp, env] = applyFunction(exp2, args, k, env);
      if (exp instanceof Promise)
          exp = await exp;
This means the web page is still interactive during the evaluation effectively. Here is an example: Click the "Load" button twice and you will see two "yin-yang puzzle" threads run on the page. Click the "Stop at Writing" button twice to stop them.


4 points by suzuki 3792 days ago | link | parent | on: More Jarc and Rainbow bugs

My Semi-Arc 10.1 looks perfect;)

  $ java -jar semi_arc.jar 
  arc> (let ((o x 2)) nil x)
  arc> (let (a (o b 2)) '(1) b)
  arc> (let default 2 ((fn ((o x default)) x)))
  arc> (let default 2 (let ((o x default)) nil x))
  arc> (let default 2 (let (a (o b default)) '(1) b))
  arc> (let (x y) '(a b c) y)
  *** 2 args expected for #<fn:2::(#((a b c)))>: (a b c)
    0: (#<done> #<fn:2::(#((a b c)))> a b c)
  arc> (let (x y z nothing) '(a b c) nothing)
  *** 4 args expected for #<fn:4::(#((a b c)))>: (a b c)
    0: (#<done> #<fn:4::(#((a b c)))> a b c)


2 points by rocketnia 3791 days ago | link

I upvoted you 'cause Semi-Arc's doing what you want it to do (which is what really matters), but I'd like to mention that I actually kinda appreciate the way Arc accepts mismatched destructuring lengths.

I've seen other people's code use a (let (a b c) nil ...) idiom as a shortcut for (with (a nil b nil c nil) ...). As for me, I like being able to design multiple types so that they can be used in an ad-hoc polymorphic way like this:

    (def theta (point)
      (let (x y) rep.point
        (errsafe:atan y x)))
  #<procedure: theta>
    (let point (annotate 'point3d
                 '(-1 0 200))
Both these techniques could still be used in an Arc implementation with a stricter treatment of destructuring: The first could be done by defining a 'w/nil macro, and the second could be expressed as (let (x y . rest) rep.point ...). In your favor, the strict treatment makes it easier to write functions that do complain if the length is incorrect, encouraging fail-fast code. However, as far as portability is concerned, the "N args expected" errors are a quirk of Semi-Arc.

I'm going to try out Semi-Arc right now. ^_^ I'm sure I'll have some error reports for you in a moment. Hopefully not too many! :-p


1 point by akkartik 3792 days ago | link

Nice. Have you played around with multiple arc implementations in the process of making semi-arc? I'd love to read about your experience writing semi-arc, and the design decisions you made. (email in profile if you want to chat further)


4 points by suzuki 3791 days ago | link

No, I have not played the original Arc yet and had hardly known both implementations in Java until recently. My version 1 implementation was a little Lisp written in ISO Standard Pascal:

  Little Lambda Lisp in Standard Pascal
You find the notable(?) feature of Semi-Arc, macro expansion without free symbol capture, even in the version 2 of it:

  Little Lambda Lisp 2 in Standard Pascal
I had translated it to Ruby, Python, C# and Java. Recently I noticed that its design is near to the core of Arc. I made it up so that it may look Arc:-). The result is Semi-Arc.

As for argument destructuring, I implemented it in a simple, general and recursive (but not so efficient) way. Track the variable 'nestedArgs' in the method 'compile(Cell j, Cell env)' in the source:
Also, you can inspect the implementation interactively:

  arc> (def f ((a b)) (+ a b))
  arc> (inspect f)
  (#<fn> ((1)) (apply (#<fn> ((2) #<none>) (+ #0:0:a #0:1:b)) #0:0:$G453))