tail recursion

[harrison@sp21.csrd.uiuc.edu (Luddy Harrison)]

>> I think that what you really don't like is the restriction, not the
>> way in which I phrased it.

Nonsense; don't kill the fun by making this personal.  I simply want for
you to tell me what it means without inadvertantly ruling out whole
classes of machines and implementations.

Consider my dataflow example again.  It may be that in order to
achieve speedup, such a machine uses a quantity of "resources" that is
not O(kX) where X is the consumption of M0.  Is it desirable to define
such an implementation to be in violation of the standard?  Many
implementation styles trade space for speed, parallel implementations
among them; your definition of proper tail-recursion appears to
disallow this.

-Luddy Harrison