The syllabic model (the Cowan meta-model), would be
straightforward, but constraining the rhyme
scheme requires more cleverness.
Hmm. If syllabic, we could define
the syllables as enumerations and
use a ref to insist this syllable be
the same. It would be a rather large
set of enumerations, so maybe fail the
"quickest test" (ok, OCR a dictionary
and tag it).
But the metric stresses? Oh, and
do they have to make sense and be funny?
That will be a lot harder. Yet another
HumanML app.
len
From: Mike Champion [mailto:]
OK, who can whip up a schema (broadly defined) that
could constrain future submissions to be valid
instances of the Limerick datatype the quickest.
I'll guess Examplatron is perfect for the job ...
