From: Matthew Smillie <M.B.Smillie@...>
Date: 2006-06-17T22:53:02+09:00
Subject: Re: self-similarity or closure in ruby

On Jun 17, 2006, at 13:17, Robert Dober wrote:

> On 6/17/06, uncutstone wu <uncutstone@sina.com> wrote:
>>
>> It seems
>>    In lisp, everything is expression
>>    In smalltalk, everything is object
>>    In Ruby,
>>        everything is an expression
>>        everything is an object
>>
>> This kind of characteristic is called closure.
>
>
> I am afraid that is not correct (c.f.
> http://en.wikipedia.org/wiki/Closure_%28computer_science%29 if you are
> interested)
> Cheers
> Robert
>
> It is also a kind of self-similarity.

Beware the Wikipedia, my son!
   The editors' bite, the facts which clash!
Beware the disambiguation page, and shun
   the frumious Bandersnatch!

(If I were more creative, I'd finish off the whole poem, but you get  
the point)

There's a more relevant meaning of 'closure' given here:
http://en.wikipedia.org/wiki/Closure_%28mathematics%29

This is more the sense which I think the original post was meant, and  
seems intuitively correct in that sense: in Ruby, any operation on an  
object results in an object.  This is not true in the general case  
for OO languages (consider int vs. Integer in Java, or pointers in C++).

I'm not as sure that Ruby is closed with respect to expressions in an  
interesting way.  The not-so-interesting case that I can imagine is  
that any valid expression in a language results in a (possibly  
trivial or empty) valid expression in that language, which seems to  
me like a necessary condition to have a useful programming language  
in the first place.  Maybe this definition of expression is too broad?

matthew smillie.