[comp.theory] Naive question

matthew@cs.uq.oz.au (Matthew McDonald) (04/12/91)

	I pretty sure the answer to this is no, but I'd be interested
in knowing why exactly. If the answer is yes, even better. :-)
I'll collect replies and summarize to the net if there's any interest.
Sorry if the question is too stupid. 
	In general, is there a recursively enumerable set of functions which
are guaranteed to terminate for all input from a given domain, such that for
*any* function f, terminating for at least some input from that same input
domain, there is a member of the enumerated set whose input/output pairs are
a superset of the set of input/output pairs of f (Where the set of i/o pairs
for f includes only input for which f terminates.)
	I'd appreciate anything you got, thanks in anticipation,
		Matthew.

iam@Gang-of-Four.Stanford.EDU (Ian Mason) (04/13/91)

if by a recursively enumerable set of functions you 
mean a recursively enumerable set of indices of such
fuctions, then the rice shapiro theorem gives a strong
negative answer to you question. a reference to it is
nigel cutland's computability.


		-iam-