declaration of variables?

[Mike Meyer]

mwilson at the-wire.com (Mel Wilson) writes:

>    My favorite poster child for not declaring things is:

I can't resist...

>     def number (s):
>         "convert something (probably a string) to a convenient numeric type"
>         try:
>             return int (s)
>         except ValueError:
>             try:
>                 return long (s)
>             except ValueError:
>                 return float (s)
> 
>    If I had to declare things, I'd have to declare the
> return type of `number`, and obviously I don't want to.

With a sane strict type system, you declare it as "number", a deferred
class of which int, long and float are subclasses. 

>    Since the parameter `s` has no declared type, `number` is
> safe to call with an object that's already a number, and it
> can be called successfully with any class instance that has
> a __int__, __long__ or __float__ method,

s, on the other hand, is harder to deal with, because you're relying
on the ability to dynamically test for a method, which is something
relatively rare in strictly typed languages.

BTW, I'd have added one more try-except clause, so that "j" would be
correctly converted to a complex :-).

        <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.