Adjust obaptheory to allow some singletons in traps

[orcmid]

See #65 for what this is about.

The ob-exp a b c is now evaluated as a :: b :: c to provide a trap for there being no established applicative-interpretation for pure-lindy forms.

In an expanded form of the REPL output, such forms could be output in the applicative form as a simple variation on the pure CFob out;put case.

The adjustment to obaptheory will incorporate certain singleton cases (at least .NIL) in operand-position traps without changing the round-trip preservation of traps.