James Edward Gray II wrote:

> On Dec 5, 2005, at 9:38 AM, Mike Harris wrote:
>
>>    each_divisor do |i|
>>      return false if sum-i == self
>>    end
>
>
> Hmm, does that work?
>
> I'm not trying to slander your code.  I'm actually asking because I  
> think it's very elegant, if it's also correct.
>
> You're code removes divisors one-by-one, but what if our divisors  
> looked something like:
>
> 1, ..., 50, ..., 100
>
> And the total sum was just 50 over the number we're testing?  The  
> order we remove divisors might be significant then.  Am I right?
>
> James Edward Gray II
>
>
>
ah, I screwed up and scrambled the problem description inbetween reading 
it and writing it.  My solution only considers subsets of size N-1, 
where N is the divisor count.