Interesting idea but there's the issue of similar sounding words when communicating orally. "gab, with a 'b' or gap, with a 'p'?". There's a number of these occurrences in this word list.

An alternative is Oren Tirosh's mnemonic encoding [0] project. It has 1633 words that were chosen specifically because they are short (4-7 characters), phonetically different, easy to understand over the phone, and recognizable internationally.

[0]: http://web.archive.org/web/20090918202746/http://tothink.com...

Here's a quick and dirty way to make this into a Python dictionary:

(echo 'IP_MNEMONICS = {'; w3m -dump http://gurno.com/adam/mne/ | grep -E '^[0-2][0-9][0-9] ' | while read quad word; do printf ' "%s": "%s",\n' "$quad" "$word"; done; echo '}';) > mnemonics.py