The docs are still a bit fuzzy, but there are a couple of basic recipes

• use codecs.open
• use codecs.lookup, then..

• use wrappers around pre-opened files
• use explicit decode, encode

Things to remember

• string means a sequence of bytes (I like to say 8-bit string)
• unicode string is a sequence of unicode characters in Python's internal representation, which for the record may be either 16 or 32 bit.
• everybody else probably calls this next part obvious, but I find it reassuring to be explicit:

• encoding means translating from native to external format
• decoding means translating from external format to native
• encoded data is some kind of externally-meaningful binary data (in a buffer, in exernal files). It is not directly meaningful to python; it is encoded relative to Python native representation and must be decoded to get character data that python can operate on. I sometimes like to call coded data external data or foreign data.

###