From: Albert Wagner <alwagner@...>
Date: 2001-06-03T02:49:32+09:00
Subject: [ruby-talk:16171] Re: Block arguments vs method arguments


> There're TWO closely related but different problems.
>
>   (a) The semantics of block parameters are defined by multiple
>       assignment.  

I don't know what this means.  That yield calls a block multiple times?

>       This is because they are originally iteration  variables.  
>       But they are sometimes used/expected like formal
>       arguments (e.g. name shadowing, default value, etc.).

But Isn't such use a violation of the defined intent of a block?  I always 
thought of a block as pure code with any/all input supplied by the receiver. 
Therefore the receiver dictates the contents of |...| to the sender.  
Essentially, the receiver offers a contract: 
(a) with ".each" he says "I am iterating over a sequence; I will give your 
block one object from the sequence at a time; in your block, you tell me what 
to do with this one item, |a|, 
(b) with ".sort" he says "I am sorting a collection of objects; I will give 
your block two objects at a time;  in your block, you tell me how you  want 
them ordered; here are the two objects |a b|.
>
>       Plus, |a| accepts any number of values, where |a,| takes only
>       one value (this too is caused by multiple assignment semantics).

I guess I don't understand this either.

>
>   (b) blocks introduce new local variable scope, i.e. a local variable
>       that appears first in a block is not available out of the
>       block.  This is confusing sometimes.

Seems straightforward to me.  Why is it confusing? Isn't the block just a 
piece of helper code for the receiver?  Does the sender expect to renegotiate 
the offered contract on the fly?

I suppose I am too new to Ruby to even be commenting.  But explaining my 
understanding of a thing and then being corrected by those that know better, 
is usefull to my understanding.