Moby Dictionary Interpretation
by Don Colton

On the plus side:
(1) this dictionary is dated 1998, suggesting it is more up to date
    than the 1994 CMU dictionary.
(2) It has 177K entries, versus only 110K in the CMU.
(3) The symbols look more nearly like WorldBet/IPA to begin with.
(4) The words are case-sensitive, whereas CMU is all upper-case.

On the minus side:
(1) I am still a little unclear on how this dictionary works.
(2) I have spent some time with it and I have noticed a few errors.  Rare
    two-word entries have a space between the words, while the rest
    have an underscore between.  These can be easily found and fixed.
(3) Some pronunciations look wrong, such as Aaron versus Aaronic.
(4) The @ and & symbols have the opposite meaning compared to WorldBet.
(5) It is unclear why some phonemes have // around them while others
    do not.  The // always seem to come in pairs, but do not always
    seem needed.
(6) There are odd phonemes such as "c" that seem to be accidents left
    over from the conversion process, which should have become a k or
    an s.

The dictionary feels to me that it was designed more for the creation
of pronunciations (text to speech) than for the recognition of
utterances (speech to text).  I am not sure.