Armin Roehrl <armin / approximity.com> wrote in message news:<0111240116181H.28105 / linus>...
> > A couple of corrections:
> Thanks a lot for the fast reply. I corrected 2) already;
> Should I correct 1, too?
> >
> > 1) Smalltalk is dynamically typed
> Is this wrong?
> I typed dynamic/static here as in GNU Objective-C I think
> I can assign a variable a different type than id. Then the compiler
> can do a check as in C++.

No.  Smalltalk is most definitely purely DYNAMICALLY typed, just like
Ruby and Python.  Please correct this rather glaring error in your
table.

You cannot add type declarations in standard Smalltalk-80 code and
have the compiler statically check it, etc.

If this post is not sufficient proof for you, you'll note that the
various other tables mentioned in this thread all list Smalltalk as
being dynamically typed.

On another note, it might be good to put different values in the
Templates row for the dynamically-typed languages (Ruby, Smalltalk,
Python, CLOS).  Basically, templates are not necessary/irrelevant in
these languages because they already support genericity just by being
dynamically typed.  So I would change each of those to say "not
needed" or similar in the Templates row.

On the other hand, templates (or similar support for genericity) not
existing in a statically-typed language like Java *is* a problem, so I
would leave that one as "no".

- Doug Way
  dew_dropper / hotmail.com
  Whisker -- The stacking O-O browser for Squeak/Smalltalk:
  http://www.mindspring.com/~dway/smalltalk/whisker.html