[ruby-talk:16881]

# -----Original Message-----
# From: Mathieu Bouchard [mailto:[e-mail]] 
# Sent: 26 ノタホム 2001 ヌ. 12:14
# To: Aleksei Guzev
# Cc: ruby-talk ML
# Subject: Re: [ruby-talk:16873] RE: Something strange with 
# Ruby's inheritance mechanism
# 
# 
# On Tue, 26 Jun 2001, Aleksei Guzev wrote:
# > Every strong-typed language implies a lot of text being not a code, 
# > but declarations.
# 
# Have you seen type-inference? Some languages actually analyze 
# your code for types and "templatize" your code according to 
# what you *don't* write. 
# You type code in the listener, and it answers with the 
# "declarations", which you don't have to put anywhere in your 
# files, as they are generated by compilation of the code.

Sometimes I think it would be great to define two methods of the same
name but different parameters ( C++ "overload").
But since Ruby does not support typing it cannot determine that method
should be called in particular. 

It would be interesting if Ruby stored types with variables and
determine the types while assigning VALUES. It should consider not the
classes which the object beliongs to but the methods it implements. Take
a look at parametred clusters in CLU. They require their arguments
implement particular methods.

Should term "interface" be added to the language? I think, no. Because
it implies some overheads. Parameterizing clusters in CLU is implemented
well.

Aleksei Guzev