From: David Alan Black <dblack@...>
Date: 2002-01-02T01:34:52+09:00
Subject: [ruby-talk:29964] Re: Proc.class vs yield

Hi --

On Tue, 1 Jan 2002, Michael Lucas-Smith wrote:

> I'm not against the iterative features. I just want them to be flexible.
> What if you need two blocks to effectively iterate? ever seen the method
> inject:into: in Smalltalk? how about hierarchy iteration. first block to
> tell it what method to call to iterate over the children, second block
> to perform the collect/select/reject.

OK, I think I see now that you're talking about perhaps having the
last block be *the* block, if it's greater than the number of &args.
As far as the desireability of this change goes, we've probably
reached the point of "agree to disagree" :-)  I'll just add that the
stuff you describe above is, certainly, feasible in Ruby; in other
words, even though the proc/lambda keyword may seem cumbersome, it can
be used to pass blocks around and "call" can be used to call them.

Also, some of what you're describing can, I think, be accomplished
with chained iterator calls.

Finally... this is interesting because it's sort of reminiscent of
discussions of parallel iterators but kind of turned sideways: rather
than having one iterator that iterates for two receivers, you're
talking about one receiver that takes multiple iterators.  (If I'm
getting it right.)

[re: {} idea:]

> It was just an idea - because someone mentioned that currently there
> might be a problem with the fact that {} is used for blocks and hashs.

Yeah, that was me, a few thousand words ago :-)


David

-- 
David Alan Black
home: dblack@candle.superlink.net
work: blackdav@shu.edu
Web:  http://pirate.shu.edu/~blackdav