Coming from an Object Pascal/C++ background as I do, I am used to objects having different namespaces for their methods, so I could say 
"myWindow.draw(...)"
or
"myCardDeck.draw(...)".

In these languages there's no real reason for the method names to be different; as you say, it's pure coincidence that the method names happen to be the same.  I find this often happening with my own code, which coexists with a huge class library.

It seems that in Dylan we'll have to be careful to avoid methods to unrelated objects having the same name, and use the module renaming facility whenever conflicts arise.  This isn't a horrible thing, but it does seem like something of a pain.

Joseph N. Wilson

You're right about avoiding conceptually different methods having the same name in a single module. But it's more and less than that.

Dylan won't let one name have different types of bindings in a single context. That is, a name cannot refer to a generic function *and* some other kind of value. And because slots are provided with generic functions, one cannot have different behaviors when used as a slot vs. a regular-old generic function.

When taken in the light of other OO language models for slot access, the syntactic tricks for `.' and `:=' are somewhat confusing because of their interpretation in terms of generic functions.

Joseph N. Wilson

In such a setting, all such slots can be thought of as functions of one variable evaluating to a method.  The dylan cliche to give such a behavior is something like this:

define class <my-window> (<object>)
   slot id, init-keyword: id:;
 end class;

define method draw (window :: <my-window>)
   curry (my-window-do-draw, window);
 end method draw;

define constant my-window-do-draw =
    method (window, thing)
       print (window.id);
       print (thing);
    end method do-draw;

Obviously, printing is not how you draw something, but this example will execute with Marlais as follows:

? define variable w =
    make (<my-window>, id: "this-window");
 ? w.draw ("that string");
 "this-window"
 "that string"
 ?

I don't recommend that this cliche be put forth as anything other than a motivation for exploring the difference between the Dylan model and other OO models.