From: Robert Klemme <bob.news@...>
Date: 2004-06-14T20:38:40+09:00
Subject: Re: undefine


"Claus Spitzer" <DocBoobenstein@gmail.com> schrieb im Newsbeitrag
news:a7c2a81904061400186bf233eb@mail.gmail.com...
> Again, this is where it's useful to consider the OO approach and think
> of variables as just the symbols you use to reference to objects. If
> the symbol you're using is not assigned to any object (a.k.a. it
> points to nil) then it it undefined.

I know I'm picky, but

1) You don't "assign a symbol to an object".  If at all it's the other way
round: you assign an object to a symbol.  But I'd rather say "you make the
variable point to an instance".  The symbol is just the representation of
the memory location that holds a reference to some instance.

2) 'nil' is in fact a valid instance.  It has methods that can be invoked
safely:

>> nil.to_s
=> ""
>> nil.to_i
=> 0
>> nil.id
=> 4

This is different from Java where you get a NullPointerException if you
try to invoke any methods on 'null'.

I think you wanted to say the right thing but it was a bit ill formulated
IMHO.

Kind regards

    robert

>
> On Mon, 14 Jun 2004 05:50:55 +0900, tony summerfelt
> <snowzone5@hotmail.com> wrote:
> >
> > On Sun, 13 Jun 2004 20:08:48 +0900, you wrote:
> >
> > in my particular case i'm thinking black box approach. i don't want
> > the variable to 'exist' any more.
> >
> > i don't know if that's how it's thought of in perl, but when i
> > undef($x), $x is no longer there for me to use unless i assign it
> > something, and i can test that it isn't with:
> >
> >  if ! defined($x) # do something.
> >
> > as long as ruby 'acts' the same way, it's all i need.
> >
> >
>
>