PDL2's representation language is the classical
attribute-value language. A pair p = (i, o) in I×O is
called an example. A precept is an example in which
some of the i-entries (inputs) are set to the special
value don't-care. An input whose value is not don'tcare
is said to be asserted. If i
j
is an input whose value
is don't-care in a pair p, then every pair p' obtained
from p by asserting i
j
to any one of its possible values
(in I
j
) and leaving p's other inputs unchanged has the
same output as p. Informally, this says that in the
context of p, the value of i
j
has no effect on the value
of the output. A pair containing don't-care inputs
represents as many examples as the product of the
sizes of the input domains of its don't-care inputs. The
number of examples a pair represents is its size. If a
pair has m of its n inputs asserted, its generality is
defined by n-m. Precepts are more general than
examples.