Best Method of Error Handling

[Kamilche]

I'm totally new to Python error handling. I see there are many, many
ways to handle errors in Python. What is the standard accepted 'best
practice'?

I will be writing unit tests for every module, and translating into
different languages. If I call everything an 'Exception', I can easily
load the messages from a text file... if I finely divide all the
errors up into separate error classes, I can't load from a text file
at run time. Or maybe I can, and I just don't see it.

Thanks for your time!

--Kamilche