From: Mike Furr <furr@...>
Date: 2009-02-03T07:27:41+09:00
Subject: [ruby-core:21768] Re: New documentation system!

Charles Oliver Nutter wrote:
>> I still hold out hope for a sufficient degree of type inference
>> to avoid this, though.
> 
> H/M type inference can't handle inheritance, so I think we're going to 
> have to accept that limited type hints are necessary for Ruby. Unless 
> you had some other type inference mechanism in mind?

This is getting slightly off-topic, but my research is exactly in this 
area and I have developed a static type inference algorithm that can 
type check many Ruby idioms.  You can read about the project here[1]. 
However, I don't want to divert this thread off-topic going into static 
typing :)

We currently only require annotations in two places: 1) for code written 
in C and 2) for introducing parametric classes and methods.  I would 
love to have access to an officially supported annotation language 
though.  As it stands now, I wrote my own Ruby parser that allows 
annotations before classes and methods (like RDoc), but also supports 
type annotations on individual expressions.  Since a major benefit of 
static checking is the *verified* documentation, a unified API for 
annotations would make this more accessible.

-Mike

1 - http://www.cs.umd.edu/projects/PL/druby/