From: Dan Doel <djd15@...>
Date: 2003-08-12T07:04:15+09:00
Subject: Re: Ruby and OOP-design (question of an old "procedural person" ; )

Mills Thomas (app1tam) wrote:

>Here, Here!
>
>If someone has some worthwhile reasons they can share for returning the
>values, please do so.  Otherwise, I would prefer not giving someone data
>they might use that might not be there later.  I wrote method that returned
>a boolean-like value by returning !(!(something)) just to ensure exactly
>true or false.  The alternative is just to unnerving! :-)
>
>Drew
>
That all depends on what exactly you want the method to do. If #nonzero? 
is defined as "returns the
number when not zero, or false otherwise," then it's documented that the 
result can be used in that way,
which could possibly be useful in some cases.

If #nonzero? is defined as "returns a true value when not zero, and a 
false value otherwise," then there's
no compelling reason to return anything other than true or false 
exactly.  However, there also isn't much
of a compelling reason _to_ only return true or false exactly, except 
for rare cases like
1.nonzero? == 2.nonzero?, and the fact that it unnerves some.

However, many people don't have a problem with it, so perhaps it's just 
something you get used to
over time.  As long as you test adequately, it shouldn't make a 
difference one way or the other in practice.

- Dan

P.S.: The above can be written (1 != 0) == (2 != 0), which is shorter to 
type, even, so one could
argue that perhaps #nonzero? ought to provide different functionality, 
which it does, currently.