I know it is almost certainly too late, but I think a lot of people will
be confused by str.decode() and bytes.encode() (or was that the other
way around)?

Calling the methods str.tobytes() and bytes.tostr() (or nicer,
str.to_bytes() and bytes.to_str()) would be hard to confuse and IMO will
lead to fewer bug reports and complaints once Python 3.0 final comes out.

And there is a kind of precedent in that threading.isDaemon() became
threading.is_daemon() between 30a5 and 30b1.

-1
encoding and decoding is generally accepted terminology. Besides then we
would be binding ourselves to returning bytes or a str; that's not
necessarily guaranteed.

Anyway, I think this hardly has a chance in the place we are (between
betas).

> encoding and decoding is generally accepted terminology

Yes, but I personally always pause a couple of seconds each time I have
to write "encode" or "decode".
Following Mark's idea, I think I will mentally use "en-bytes" and
"de-bytes" as a mnemonic...

There's nothing new to .encode() and .decode(). They have existed since
Python 1.6.