From: Mark Hubbart <discordantus@...>
Date: 2005-05-12T08:41:45+09:00
Subject: Re: alternatives to ? : contruct

On 5/11/05, gwtmp01@mac.com <gwtmp01@mac.com> wrote:
> On May 11, 2005, at 4:07 PM, Jason Foreman wrote:
> >
> > I'd prefer it with less ;'s as well:
> >
> > a = if (y==z): b else c end
> 
> Is this really better than:
> 
>      a = (y==z) ? b : c
> 
> ?

A little, for readability. I like using actual words as operators
whenever possible. This is, of course, excepting symbolic operators
like + and -, which I learned back in first grade, and carry as much
meaning as actual words.

> The tertiary operator ?: can be used if you want
> something terse and "if then else end" if you want something
> more verbose.  I'm not sure I see the need for yet another
> variation.

First of all, it's the ternary operator :) As for the "if then else
end" statement, the dangling "end" at the... uh, *end*, is a bit of an
eyesore. It's what makes me shy away from using inline conditionals
like that.

I've thought a couple times that it would be nice to have something
like "otherwise" available. I wonder if it would be too difficult for
the parser to parse these two snippets the same:

----
a = if (y==z)
  b
else
  c
end
----
a = if (y==z) then b else c
----

That is, if else is on the same line as the code to execute if true,
with no semicolon, the "end" can be dropped. Or this, instead:

----
a = b if y==z else c
----

Can the parser definitively detect that the if...else is being used as
a statement modifier in this case?

Just some thoughts off the top of my head, here. I could easily waffle
on them if a good case is made against it.

cheers,
Mark