PEP 245

[Alex Shindich]

On Sun, 1 Apr 2001, Carlos Alberto Reis Ribeiro wrote:
  
> I'm not the best person to answer you, but that's my take. Interfaces allow 
> for some very powerful forms of OO abstraction to be implemented. You 
> aren't forced to use inheritance for everything. This frees the class 
> hierarchies, making it possible to have a "mesh" of classes instead of a 
> very rigid tree structure. It also tends to minimize the use of multiple 
> inheritance, when all that you want is to support multiple interfaces. All 
> of this is possible even without static typing. As for your concern with 
> the semantics, no strongly typed language can enforce it - you have just to 
> hope that the function being called behaves as documented.
>
I do not want to restart the whole discussion on the difference between
base classes and interfaces, but I am of the opinion that they both
represent an is-a relationship. 
> >2. At what point will the compliance with the interface definition be
> >done, at compile time, or at run time?
> 
> Don't know what the dev guys are thinking. I would prefer *not* to 
> explicitly check interface compliance at any time. You said that you 
> supported a interface, it's your job to write the methods for it. If you 
> don't do so, you get an exception.
> 
Arguably, this behavior is already available through runtime method
binding.


>  But I think that is reasonable for 
> any language developer to borrow stuff from other languages. People has 
> been doing this for a long time anyway <wink>. 
I agree that borrowing best practices from other languages is a great
idea. In fact, interfaces make Java very powerful. I just don't think that
interfaces will be as useful for Python.
> Also note that, by 
> supporting a runtime (on-demand) approach to interface checking, we are 
> keeping with Python tradition.
This makes interfaces to be as good as doc strings...

-- 
Alex Shindich
mailto:alex at shindich.com
Visit http://www.shindich.com/