[Haskell-cafe] Elerea/GLFW Tetris

jean legrand kkwweett at yahoo.fr
Sat Aug 15 18:10:34 EDT 2009


> How come you started out with playing around with FRP
> libraries right
> away? It's a rather peculiar choice, I'd say.

I'm always curious about how the languages I study interact with OpenGL because I'm in the numerical simulation. When I came to Haskell on December, it was quite a new sensation at the functional level first (I knew Scheme but I never tried to find out how Scheme could interact with OpenGL!) and also about IO. But I liked it (and running). Then I studied Reactive but I had some problems installing it on Linux GHC6.6 so I was looking another library and I saw your post on March I think. But my knowledge about Haskell was too light at this time. Then I studied more and here I am!

> It's an interesting exercise, and quite a nice job from
> someone who
> considers themselves a beginner.

thanks, it took me a few days but as I said it was floating in my head for a few months.

> they can appear inside any expression. For instance, you
> could just say
> the following in the let declaration:
> 
> sfall = (uncurry . fall) <$> randomBehavior seed
> sid = pure id

yes, I saw sfall too late (however I haven't thought about sid, thanks)

> Another thing, which is really a matter of taste, is that
> you seem to
> like point-free style acrobatics. I don't think it's always
> the best
> choice, especially if you are to share code with others,
> since most
> people grok code with explicit parameters easier than magic
> involving
> flip and const and (un)curry and the like, except for the
> obvious cases
> of composing a chain of operations, where dropping
> arguments feels quite
> natural. For instance, the definition of f_a looks
> problematic to me
> because of these concerns.
> 

I agree with you in general (point-free is unreadable), but in the case of f_a particularly, it took me a long time (not as long as for sf_a, however) to design it (not the point-free version but the raw one) and I think the point-free version is the best way to dissuade anyone to  try to hack it and to persuade her to write a new function which fits her needs. 
It is actually the reason why I called f_a  like this : because its two arguments are a function and a non-function argument (the first argument of sf_a is of course a signal of a function). This means I don't want to know how the machine works when I write my program. But it's a matter of taste as you said.



      


More information about the Haskell-Cafe mailing list