From: Clemens Hintze <c.hintze@...>
Date: 2000-03-03T08:26:41+01:00
Subject: [ruby-talk:01701] Re: Conventions for our Ruby book

NAKAMURA, Hiroshi writes:
> Hi,

Hi too,

> 
> How can I read comp.lang.misc via Web? Do 'deja' or so help me?

Yes! But with a certain delay of some hours sometimes!

...

> > From: "Dave Thomas" <Dave@thomases.com>
> > Sent: Friday, March 03, 2000 2:38 AM
> 
> > Existing Ruby documentation gets around this by using the '#'
> > notation. It would say "see String#downcase" to refer to an instance
> > method in String.
> > 
> > Although its a good differentiator, it worries us slightly, because it 
> > isn't Ruby syntax. We though it might be confusing.

Dave! Here I have changed my mind. We could use the operator
'::'. Goto Kentaro has mentioned, he would see '::' as namespace
identification operator! And matz intended to use '::' for access of
class methods. Only because there is no big difference (internally)
between an instance and a class of an instance, the '::' operator can
also be used for instance method invocation!

I have already written a common remark about the following. Instead of
sending it too you via review comments I will propose it here, ok?

If you would decide to use '::' for indicating 'class'::'method'
syntax, it would have following advantages:

   - It is valid Ruby syntax
   - It is known by C++, iTcl, Tcl, CTalk, Perl and other programmers
   - Examples would be callable.
   - It is visually different from method invocation

I could not see any disadvantage at the moment. Only if we would have
method names beginning with caps letters, we would have to place '()'
after the method to convince the parser. No big deal, IMHO.

...

> I do not know about Smalltalk well, how about in Smalltalk world?

Hiroshi!In Smalltalk you would use '>>' normally. At least all
Smalltalk implementations I ever used, behave that way in the
backtrace window. So you would write:

   String>>downcase

> --
> NaHi, nakahiro, a.k.a. NAKAMURA, Hiroshi
> http://www.sarion.com  Sarion Systems Research
> 
> # I'm sorry for my poor English...
> 

\cle

-- 
Clemens Hintze  mailto: c.hintze@gmx.net