COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	elabor[e88,jmc]		A research problem in elaboration tolerance, facts about visas
C00007 ENDMK
C⊗;
elabor[e88,jmc]		A research problem in elaboration tolerance, facts about visas

Facts about visas and passports seem to be a good domain for
studying formalization common sense, especially elaboration
tolerance.

Here are some aspects of the matter.

1. The domain involves no physical action, so physical skills
aren't involved.  It may be good for quarelling with the Dreyfus's
skill model, but that's a peripheral issue.

2. The knowledge about passports and visas is acquired incrementally
to general knowledge about travel.  This should be a requirement
for our formalism --- that the knowledge be addable incrementally
to knowledge not anticipating bureaucratic requirements.

3. The knowledge need not be based on experience.  The clerk in
the travel agency may not have travelled.

4. We can generalize to bureaucratic requirements for travel in
general.

5. Include visa requirements of various countries, Mexican tourist
cards, restrictions on visas, vaccination requirements.
The point isn't that the requirements are included in the initial
system but that the elaboration tolerance of the system should
permit their subsequent addition.

6. Nonmonotonicity at least to the extent of adding new facts
underneath a circumscription will be required.

need(go(SU),visa)

It should follow from the general formalism that it's a Soviet
visa that is required.

¬need(go(UK),visa)

That the info refers to an American should follow from context.
That there may be qualifications, e.g. w.r. length of stay
and employment, should follow from general knowledge.

Query: Can the knowledge be Horn clauses?  Presumably, we can
consider a system that doesn't generate new general knowledge - as
in the sterilization example.  However, it may need to infer
impossibility, e.g. of visiting the S.U. without a visa, and
not merely by negation as failure.

What is the next step in power beyond Horn clauses?

requires-for-entry(SU,visa)
¬requires-for-entry(UK,visa)
requires-for-entry(UK,passport)

has(traveller,passport) is ambiguous.  It is sometimes used as
owns(traveller,passport) and sometimes as
has-in-present-physical-possession(traveller,passport).  In either case,
it refers to the traveller's own passport.

A thought about contexts.  Maybe we can devise the formalism, so that
the statements of the people talking are expressed without formally
using the contexts.  However, their internal reasoning uses the
contexts explicitly, and someone talking about them may have to
refer to the contexts.