1991

	REALLY BAD (ANNOTATED) LIMERICK
	FOR PETER EKLUND'S DOCTORAL DEFENSE, LINKOEPING




	Once Dave'd defined in system			(
			(		Dave Touretzky defined inheritance
			(		nets in his 1986 CMU dissertation.
	said Erik, "some truths:  you've missed 'em;
			(		Erik Sandewall felt that there should
			(		be off-path preemption, which would yield
			(		more conclusions.
	look, Dave, the royal paths,"
			(		Sandewall's example was of royal
			(		elephants, a subclass of elephants
			(		that are typically white.
	"Look, Erik, my marines' 'ath's,"
			(		Touretzky's response was to re-label the
			(		net with an example having to do with
			(		marine chaplains.
	said Dave, "I propose that you kiss them."


	+ + +

	Now wait, H.T. - H allowed,
			(		Touretzky, J. Horty, and R. H. Thomason
			(		then embarked on a series of studies of
			(		various inheritance systems, publishing 
			(		under too many permutations of their names.
	72 times we three have rowed
			(		Horty quipped in frustration at the 
			(		St. Louis workshop that there were
			(		perhaps 72 systems worth looking at,
			(		the number being arbitrary, but in any
			(		case, too many for anyone to be happy.
	over ambiguity-blocking
			(		one major issue was whether ambiguity
			(		should block or propagate in a net.
	and Nixon's dove-hawking;
			(		Matt Ginsberg complained about the lack
			(		of expressiveness in the net which made
			(		reasoning about disjunctive conclusions
			(		difficult.  His example involved Nixon's
			(		"politically motivated" disposition,
			(		whether he was reasoned to be a dove because
			(		he was a Quaker, or a hawk, because he
			(		was Republican.
	imagine what clash in a crowd!
			(		The famous phrase of T, H, & T, was that
			(		there was a "clash of intuitions."

			(+ + +

	Next Stein doubled diamond again
			(		Lynn Stein's thesis considered the
			(		problem of "zombie" conclusions, where
			(		there was disjunctive meta-reasoning.  It
			(		occurred in nets where there were
			(		double-diamonds in the topology.
	and Schlechta found Makinson:
			(		An AI Journal paper by K. Schlechta and
			(		D. Makinson also reported on Stein's
			(		puzzle, and at the time was the most
			(		exciting development in knowledge
			(		representation.
	the zombie conclusions
			(		Their paper was about "zombie paths".
	saw quick disillusion
	of the claims that the skeptics forfend.
			(		Horty called ambiguity-propagating
			(		reasoners skeptical, and Stein called her 
			(		work a discovery of the impossibility of 
			(		skeptical reasoners.

			(	+ + +

	"Not good," our man Eklund decided
	And in Peter had Peter confided:
			(		Peter Eklund had consulted Peter
			(		Gaerdenfors for the main ideas of his
			(		thesis.
	"The belief sets they want
	could be found by savant";
			(		Eklund's approach was to resolve all
			(		ambiguity by querying the user during the
			(		knowledge representation phase.
	Dialogue had with update collided.
			(		Belief update, or belief revision, was
			(		what Gaerdenfors, Makinson, and C.
			(		Alchourron had made famous.

			(+ + +

	"Mode A, mode B, and mode C;
			(		Eklund's program could be run in any of
			(		three modes.
	No a priori assessment of domain ambiguity."
	It sounds like a thesis . . . ;
	with Loui, Aeachus,
			(		Aeachus, a judge from Hades, son of Zeus,
			(		father of Achilles, is a good model of the
			(		doctoral opponent in the Swedish system.
	Surely it'll be an early G & T.
			(		Eklund, being from Australia, would not
			(		miss this reference.  It was his aim to
			(		do the appropriate dereferencing to
			(		celebrate!
					)