Date: Mon, 6 Jan 92 14:06:35 EST
Subject: KIF syntax
Len Schubert made the following suggestion, seconded by Bob MacGregor:
> (Aside: Bob prefers ?x where KIF uses $x; so do I.)
Let me add two more bits to the discussion:
1. The dollar sign is not acceptable in an international standard,
since it tends to get replaced with other symbols on keyboards
for languages other than American English.
2. The question mark is a very nice symbol to use as a trigger for
a backward-chaining style of inference, intended to find a value
for some variable.
In conceptual graphs, I use an asterisk, as in *x, to indicate variables,
but that might conflict with its use in multiplication. An underscore,
as in _x, might be a useful compromise for marking variables.
But syntax is one area where personal preferences can lead to
endless debate. For the ANSI X3H4 committee, we have adopted the
position that the standard will be defined in an abstract syntax,
but that two concrete notations will be used for the examples in the
technical report: traditional textbook-style predicate calculus
and conceptual graphs. Any other notation that can express the
equivalent semantics would be acceptable.
Using two distinct syntaxes for examples helps to insure that
design features are chosen on semantic grounds rather than
peculiarities of the notation. (Chris Date has pointed out
that some of the worst aspects of SQL semantics resulted from
design decisions that were motivated by its syntax.)