From: Paul Prescod <paul@...>
Date: 2003-03-21T07:08:06+09:00
Subject: Re: Strong advantages over Python

Greg McIntyre wrote:
>...
> My answers currently include 'meta' programming (trapping undefined
> method calls, defining new class/module methods/attributes dynamically,
> etc.), refactoring control flow (Python has generators but they're more
> clumsy to use), invisible resource management (e.g. File#open { ...
> }), Array and Hash being extensible classes (Python has UserList and
> UserDictionary, but they're annoying and AFAIK do not work well with
> the existing control structures which expect primitive lists and
> dictionaries).

Please verify these publically before you teach this stuff to these 
students. I'm not going to get into an argument about whether Ruby is 
better than Python, but I do know for a fact that Python can do some 
things that you say it cannot. There may be nuances that are unique to 
Ruby but the overall features are there.

For instance:

~/p4/usr/pprescod> cat foo.py

class Foo:
     def __getattr__(self, attrname):
         return lambda : attrname

f =  Foo()
print f.a()
print f.b()
print f.c()

Foo.special = lambda(self) : "In dynamically assigned method"
f.special()

f.special2 = lambda : "In per-instance dynamically assigned method"

print f.special2()

for i in range(0, 10000):
     file = open("/tmp/test.txt")

class MyDict(dict):
     def foo(self):
         return "foo"

class MyList(list):
     def bar(self):
         return "bar"

d = MyDict()
d["a"] = 5
d[5] = "a"
print d.foo()

for i in d:
     print i, d[i]

m = MyList([1,2,3,4])
m[0]= "a"
print m.bar()

for i in m:
     print i

class MyInt(int):
     def doSomething(self):
         return "Did something"

x = MyInt(5)
print x.doSomething()
~/p4/usr/pprescod> python foo.py
a
b
c
In per-instance dynamically assigned method
foo
a 5
5 a
bar
a
2
3
4
Did something