From: gabriele renzi <surrender_it@...>
Date: 2005-10-01T23:16:46+09:00
Subject: Re: Lazy evaluaton (was Re: In your opinion....)

Robert Klemme ha scritto:
> Daniel Lewis <danieljohnlewis@gmail.com> wrote:
> 
>> oooh my thread has split :)
> 
> 
> Better your thread
> than your head.
> :-)
> 
>> I'm going to be noting all this down, Lazy Evaluation is a subject in
>> Functional Programming and AI.
> 
> 
> ... although it's by far not restricted to that.
> 
> Btw, recently I discovered (better: started using) an interesting idiom 
> that exploits the fact that both Proc and Hash implement #[].
> 
> FUN = config_memoized ?
>  Hash.new do |h,k|
>    # complex calculate val
>    h[k] = val
>  end :
>  lambda do |k|
>    # complex calculate val
>  end
> 
> # later
> 
> x = FUN[y]
> 
> Nothing really fancy I guess but it's nice how seemingly these 
> integrate. (Now, is this another thread split?)

could'nt you just use the default_proc in a Hash?
irb(main):001:0> fun= Hash.new do |h,k|
irb(main):002:1*  p 'calculations'
irb(main):003:1>  h[k]= rand
irb(main):004:1> end
=> {}
irb(main):005:0> fun[10]
"calculations"
=> 0.364681017817929
irb(main):006:0> fun[10]
=> 0.364681017817929

ok it is inverting the concept, I know but it seem to get the same 
result :)