terms to tries.
* Changed answer depth metric to answer size metric. This improves
the speed of adding answers and lays the groundwork for
functor-specific (size) limits. Changed the names of the flags as
well (depth-> size)
* Made answer terms and lists use the same depth limit. This helps
unify code, and later we will be able to weigh list and non-list
functors differently.
* Redid size checks so that cycle checking is always done, even if
max_table_answer_size is not set. (By default, when the size
reaches 400). Also updated checks in trie_intern to do the same.
* Took out failure and warning actions, which weren't of much use.
* Added stubs for tripwires when an answer exceeds size limit.
* Fixed bug where lists in large answers were not being properly
abstracted.
* Changed lots of variable, macro, and function names to support
better maintenance. Updated some, but by no means all documentation
in the manual.