[Python-Dev] Some comments on PEP 328 (absolute/relative imports)

[Barry Warsaw]

On Sat, 2004-04-03 at 13:24, Bernhard Herzog wrote:

> I don't think this is a good description of the actual problem.  The
> problem is that the modules in the package shadow global ones and
> there's no easy way to bypass the local package.  FWIW, though, I don't
> recall ever running into a problem with that.

It happens, and it makes for tricky choices.  I know to stay away from
internal modules like re and email, but as the standard library gets
fatter, there's more chance of collision.  The only remedy will be to
rename an internal module when I find that I suddenly want to use some
new standard module.  There should be more independence.

If all import paths were absolute, then I wouldn't have this problem,
and in fact, that's the rule my own code tries to adhere to.  See the
email package which uses absolute imports internally.  I know others
have use cases for relative imports, but I've never quite seen the need.

Ideally, IMO, all imports would be absolute.  If we have to throw a bone
to relative import proponents in order to get to absolute imports by
default, so be it.

-Barry