[blixt]

I guess sum was only implemented to support numeric values. However you can easily roll your own:

    >>> def add(it):
    ...   return reduce(lambda x, y: x + y, it)
    ...
    >>> add(['hello ', 'world'])
    'hello world'
    >>> add(x for x in xrange(10))
    45

Edit: almost forgot a possibly even more Pythonic way:

    >>> import operator
    >>> def add(it):
    ...   return reduce(operator.add, it)

[chicago_wade]

    >>> import operator
    >>> def add(it):
    ...   return reduce(operator.add, it)

Ooh, I like this one. Thanks!

[blixt]

Looping back to why sum doesn't work with strings, it looks like they implemented it like this:

    >>> def sum(it):
    ...   return reduce(operator.add, it, 0)

Basically the first value is always added to 0. This has one important difference which can be a valid compromise if you expect to almost always sum numbers. If you try to call add([]) without the initial 0, you'll get an error because there's no way to know what the zero value is.

In a typed language you could use type inference and use the inferred type's zero value (if the language has such a concept for all types like for example Go does). In Python I guess you could fall back to None, but then you'd have code that doesn't behave consistently for all inputs.