From: Leo Razoumov <see_signature@127.0.0.1>
Date: 2001-03-09T05:10:02+09:00
Subject: [ruby-talk:12297] Re: FEATURE REQUEST: 'my' local variables

>>>>> "BT" == Ben Tilly <ben_tilly@hotmail.com> writes:
    BT> And so I still absolutely think you are wrong that it is
    BT> bad to have the concept of being able to say that this,
    BT> here, is a private variable that shall forevermore be
    BT> private no matter what else is in scope.  An example why not
    BT> having it is a bad idea, when constructing strings or blocks
    BT> that will be evalled in someone else's context, it is good
    BT> for both you and them that you have the ability to within
    BT> this section declare certain variables as private to your
    BT> code and be able to limit how you will interact with their
    BT> code without having to examine it.

Thank you for emphasizing my point once again. Indeed, encapsulation properties
of a language should not depend on the way a programmer chooses the names for
his local variables. And we all beleive in the power of encapsulation in OO
languages, don't we!!

    BT> Absolutely none of the examples that you have presented has
    BT> shaken my belief that private temporary variables, a concept
    BT> that is available in most decent languages, would be an
    BT> unnecessary obfuscation in Ruby.  I use them religiously
    BT> when they are available.  It is true that 99% of the time
    BT> they accomplish absolutely nothing, but I am glad that I will
    BT> never find out which 1% of the time I saved myself a nasty
    BT> headache...
    BT> Cheers,
    BT> Ben

Correct! I am a bit surprised that my original posting caused so much a
discussion on this topic with a lot of people expressing their opposition to
this concept. I must admit my choice of 'my' brought negative associatins with
Perl which clearly hurt the discussion. Let's take Lisp or Scheme. These
languages have a scoping construct 

    (let (put_your_locals_here) body_of_the_block) 

which does exactly what I want and without Perl bugs.  By the way, even C has
this kind of behaviour (I can declare locals with no regards to other locals on
outer scopes).

I really do not see the reason to oppose this feature unless implementation is
very compicated (it should not be).

--Leo