Tail recursion, etc.

[Guillermo J. Rozas <gjr@martigny.ai.mit.edu>]

I don't like the term "tail expression".  Of course, I don't like
"tail recursion" or "tail call" either.

MIT Scheme has used the terms "reduction" (and "subproblem" for
subexpressions that are not reductions).  The one advantage this term
has is that it is much nicer to say "x reduces to y" than "x tail
calls y".  I don't even know how I would rephrase "tail expression" to
become active.

One thing that the proposed formalizations do not seem to embody is
that the notions of "tail call", "tail expression", "reduction",
etc. are not absolute, but relative to some starting point.