From: Eric Mahurin <eric_mahurin@...>
Date: 2005-06-02T11:59:46+09:00
Subject: Re: ternary operator confusion

--- Gavin Kistner <gavin@refinery.com> wrote:

> On Jun 1, 2005, at 4:42 PM, Charles Hixson wrote:
> > Uniform access could have been achieved if only () was made
>  
> > optional.  (A good choice in my mind.)  OTOH, I'm not that
> happy  
> > with optional parenthesis where the parameters are
> required.  To me  
> > that feels like a bad design choice.  (And with some type
> faces the  
> > difference between (stuff) and {stuff} Yisn't all that
> obvious.  But  
> > that's a font problem...mainly.)
> 
> Would you have prefered to be forced to write:
>      my_object.foo=( the_value )
> versus
>      my_object.foo = the_value
> ?

Now you are bring operators into the picture.  Many (as above)
do get translated to method calls, but operator syntax is
completely independent of normal method call syntax.  Although
semantically the above is a method call, syntactically it is an
assignment.  If syntactically is was a method call, you would
be able to do this:

my_object.foo=(a,b)

but you cannot because it treats (a,b) as the RHS which is
invalid.  Also, consider that you can do the following which
get translated to 2 = methods.

my_object.foo,your_object.bar = a,b

Your argument is better suited for the ? methods that take an
argument (i.e. respond_to?).  I have to admit that I don't
always put () around its argument because it it usually just a
Symbol and not an expression.  But, these ? method calls do
seem to behave inconsistently:

respond_to? :test # valid
respond_to?:test  # syntax error, why?
respond_to?(:test) # valid
self.respond_to? :test # valid
self.respond_to?:test # valid
self.respond_to?(:test) # valid





		
__________________________________ 
Discover Yahoo! 
Find restaurants, movies, travel and more fun for the weekend. Check it out! 
http://discover.yahoo.com/weekend.html