From: "David A. Black" <dblack@...>
Date: 2005-08-10T23:09:50+09:00
Subject: Re: new block notation (was: Re: ruby-dev summary 26468-26661)

Hi --

On Wed, 10 Aug 2005, Austin Ziegler wrote:

> On 8/9/05, Trans <transfire@gmail.com> wrote:
>> You could still use the brakets:
>
>>   double = do(x){ x*2 }
>
>> That would mean 'do' == '-> '.
>>
>> But if you ask me, it's all gotten pretty heinous. Why not try
>> somthing a little more common sensical:
>
>>   double(x) = { x*2 }
>
> Hmm. But that doesn't allow for clean, anonymous lambdas. I see
> where this is coming from, even though I don't think it's something
> that I'd use that much, and though I don't think that the various
> forms proposed so far are clean.

If I'm reading its intent correctly (create a lambda and assign it to
the variable "double"), it seems to me to depend on a kind of "lambda
context" for its meaning, very much like "array context" (etc.) in
Perl.  Normally, double(x) can't be an lvalue.  So it involves a kind
of scan-ahead to see that there's a code block on the rhs.  (Which of
course there isn't; I guess this is also predicated on having {} not
mean hash any more.)

I'd actually prefer to see a keyword or two, over the "context"-style
assignment scenario.  The fact that this would be the *only* context
in which it would work doesn't help.  It still can't be understood
without recourse to the rhs, which I don't like.


David

-- 
David A. Black
dblack@wobblini.net