On Thu, 2009-02-05 at 10:04 -0800, bearophileh...@lycos.com wrote:
> This comes after a small discussion in another Python newsgroup.
> Haskell supports a where clause, that's syntactic sugar that allows
> you to define things like this:
>
> p = a / b
> where
> a = 20 / len(c)
> b = foo(d)
>
> That means:
>
> a = 20 / len(c)
> b = foo(d)
> p = a / b
>
> I don't know how much good this syntax can be in my Python programs,
> probably I have to use it some time to judge.
>
> In the Python shell you usally have to use a bottom-up style of
> programming, while a where may allow you a more top-down too. I can
> enjoy this.
>
I don't like it for the following reasons:
* Flat is better than nested.
* There should be one-- and preferably only one --obvious way to
do it.
One could imagine this getting "out of hand" e.g.
p = a / b
where
a = 20 / len(c)
where
c = p / b
try:
b = foo(d)
where
d = bar()
except:
b = 0
It also begs the question, should the except: clause be written to
handle an exception raised in foo() as well as bar()? or should one also
write a try/except around bar()?
Usually when I'm looking at an identifier (function, class, variable)
being used, I tend to look *up* to see where it is defined.
--
http://mail.python.org/mailman/listinfo/python-list
