I've been playing around with CoffeeScript and docco a bit lately and I thought I'd share this little project. Thanks to Jeremy Ashkenas for creating these fantastic tools.

Nice code!

Docco always impresses too, eh?

I really like the use of:

this[key] = value for key, value of options

I'm going to use that for sure :)

Be sure to check out the annotated source - http://willbailey.name/conway/docs/conway.html

I really like how that is done. It would be great to be able to annotate source code like this within your IDE.

Looks like it was created using Docco - http://jashkenas.github.com/docco/

the newest version of vim has this seamlessly integrated :)

For real? If so, how do you access it?

!docco %

I love CoffeeScript, except that it makes me hate writing javascript code now :) Thanks @jashkenas for an amazing language.

I personally like code like this:

  countNeighbors: (cell) ->
    neighbors = 0
    neighbors += @isAlive cell.row+x, cell.col+y for x in [-1..1] when x || y for y in [-1..1]
    neighbors 
  isAlive: (row, col) -> if @world[row] and @world[row][col] and @world[row][col].live then 1 else 0

but you have to be careful, since putting a space before the "+x" will cause some bad stuff to happen. Maybe I shouldn't drop so many parenthesis.

@jashkenas can you fix [-10..10] (constant boundary) loops to not check for increasing/decreasing :) I can't think of an edge case that breaks it.

If you add a space after the + also, it’ll be fine (indeed, better).

But to be honest, I hate code that tries to do 5 bits of logic on one 100-character-long line, instead of just breaking into two more readable lines. [Edit: looks like that was fixed in the version at http://willbailey.name/conway/docs/conway.html]

Also, can’t your isAlive function look like:

  isAlive: (row, col) -> @world[row]?[col]?.live

@amirshim:

Yes, range comprehensions were recently optimized to remove direction-checking when possible. On CoffeeScript master ...

    x for x in [-10..10]

... compiles into:

    var x;
    for (x = -10; x <= 10; x++) {
      x;
    }

Even if the start/end values are variables, the check can be avoided by specifying the step as a number. It'll go out with the next release.

great!

Pretty clean code! A small suggestion:

  cell.live = false if count < 2 or count > 3
  cell.live = true  if count == 3

Could become:

  cell.live = count < 2 or count > 3
  cell.live = count == 3

since it's a bit redundant to write "true" since it's already a boolean expression. It's a little bit like saying:

  if a == true:
    return true
  else:
    return false

instead of:

  return a // Agreed that a could be a "truth" value without being the real "true". Still:

  return !!a // With a hack

Nope. The point of those if's is that the value of cell.live remains unchanged if the condition is false. Your code does not do that.

What should happen is that, if count is 3, then cell.live becomes true. If count is 2, then cell.live retains its previous value. For all other values of count, cell.live becomes false.

Or simply:

    cell.live = count is 3

no?

No. You aren't handling the case when count is 2 correctly. In that case, cell.live should remain unchanged, while you are setting it to false. (Note, however, that the code in the comment you are replying to is also incorrect. And does the same thing as your code. So you posted a correct transformation of incorrect code, that keeps it incorrect.)

The original version still contains redundancy. It would be:

    cell.live = count == 3 if count != 2

Yes, that would do it.

It's no APL... conway's game of life in a single line of code http://news.ycombinator.com/item?id=1041500

Thanks for the code review folks. I made the countNeighbors method a bit more concise and removed some unnecessary binding code from the travelWorld method.

And here I was thinking it must be about Ron Conway. Too much startup news.

heh, reminds me of my js version http://rapidpacket.com/