Issue: FUNCTION-TYPE-ARGUMENT-TYPE-SEMANTICS (Version 2)

[Kent M Pitman <KMP@STONY-BROOK.SCRC.Symbolics.COM>]

My notes from Fairfax meeting...

Cleanup meeting:

 Walter was willing to write this up, but had had trouble
 with net access and was missing some mail. KMP will call him
 and provide the relevant information for him to do the writing.

 It seemed reasonable to write a proposal for two things:
  - FTYPE, which allowed a total specification of a function's
    possible arguments and values.
  - PTYPE which allowed a partial specification of same.
    Multiple PTYPE specs would presumably be ok, but multiple
    FTYPE would presumably not. Proposal should say what FTYPE
    does when a PTYPE has been seen already and vice versa.