[net.ai] combinatory logic

srlm@ukc.UUCP (S.R.L.Meira) (06/28/84)

[-: kipple :-]

in the hope that many of you are also interested in combinatory logic...
please have a look at this and mail me any suggestions, references, etc.

------------------

[by a. pettorossi in notre dame j. form. logic 22 (4) 81]

define:
     marking

	is a function that assigns, for each combinator in a term (tree)
	the number of left choices (of path) that one has to make to go
	from the root to the combinator.
	ex.:
	marking SII = <S,2><I,1><I,0>

     the set right applied subterms of a combinator X is defined as:

	1) if X is a basic combinator or a variable ras(X) = {X)
	2) if X is (YZ) then ras(YZ) = union (ras(X)) Z

     a combinator X with reduction axiom X x1 x2 x3 ... xk -> Y
     has non-ascending property iff
	
	for all i, 1<=i<=k, if <xi,p> occurs in marking (X x1...xk)
	and <xi,q> occurs in marking Y, then p >= q.

     a combinator (X x1 x2 ... xk -> Y) has compositive effect iff

	a right applied subterm of Y is not a variable.

------------------
Theorem:
	given a subbase B={X1,...Xk} such that all Xi in B have non-ascending
	property and no compositive effect, every reduction strategy applied
	to any Y in B+ leads to normal form.

------------------
Open Problem:
	does the theorem hold if non-ascending property is the only condition?

------------------
My personal questions:
	
	if one specifies leftmost-outermost reduction only, would the Open
	Problem be any easier?

	how much of combinatory logic can we do with B?

	and with non-ascending property only?



	silvio lemos meira

	UUCP: 	...!{vax135,mcvax}!ukc!srlm
	Post:
		computing laboratory
		university of kent at canterbury
		canterbury ct2 7nf uk
	Phone:
		+44 227 66822 extension 568