Manipulating higher order terms.

[Nick]

Hi,

I'm trying to implement part of an inductive logic programming algorithm
in mercury. This involves the construction of clauses, which are called to
see if they entail a set of examples. The obvious way to represent these
clauses is higher order terms.

However, I need to know some extra information about the clauses.
Specifically, I need to be able to extract the names of all of the
predicates. Is this possible?

The only alternative I can see is to use a ground representation of the
clause and pass the name around with the predicates, which I'd prefer not
to do for efficiency reasons.

Thanks,

Nick.

----------------------------------------------------------------
          'A one character regular expression is an
         regular expression that matches whatever the
       one-character regular expression matches' - ed(1)
----------------------------------------------------------------