From: Daniel Berger <djberg96@...>
Date: 2005-09-20T12:55:30+09:00
Subject: Re: proposed attr writer patch

--- James Britt <ruby@jamesbritt.com> wrote:

> Daniel Berger wrote:
> > --- Mathieu Bouchard <matju@artengine.ca> wrote:
> > 
> > 
> >>On Sun, 18 Sep 2005, Christian Neukirchen wrote:
> >>
> >>>James Britt <ruby@jamesbritt.com> writes:
> >>>
> >>>>Daniel Berger wrote:
> >>>
> >>>IMHO, `foo?' makes sense (if @foo is meant to be
> >>
> >>boolean),
> >>
> >>IMHO, FOO$ also makes sense, if it is meant to be
> a
> >>string.
> > 
> >  
> > Oh, let's not go over the top, shall we?
> > 
> > To answer James' earlier question, there wouldn't
> be
> > much point to using this patch for bang methods,
> but
> > it would be handy for the boolean methods.
> > 
> > Instead of:
> > 
> > attr_writer :foo
> > def foo?
> >    @foo
> > end
> > 
> > I could just do:
> > 
> > attr_accessor :foo?
> 
> I see.
> 
> But does this, then, enforce the booleanitude of
> "foo?"?
> 
> In other words, if you're going to have a method
> foo?, should it not be 
> something like
> 
> def foo?
>    @foo ? true : false
> end
> 
> And should not foo=( val ) also enforce that @foo
> only takes on boolean 
> values?

Why should it?  The use of "?" and "!" in method names
is just a convention.  There's no enforcement when you
write one of these methods long-hand, so there
shouldn't be any enforcement when using a shortcut
method, either.

Regards,

Dan


		
__________________________________ 
Yahoo! Mail - PC Magazine Editors' Choice 2005 
http://mail.yahoo.com