Certification Problem

Input (COPS 329)

The rewrite relation of the following conditional TRS is considered.

le(0,x) → true
le(s(x),0) → false
le(s(x),s(y)) → le(x,y)
app(nil,x) → x
app(cons(x,xs),ys) → cons(x,app(xs,ys))
split(x,nil) → pair(nil,nil)
qsort(nil) → nil
split(x,cons(y,ys)) → pair(xs,cons(y,zs)) | split(x,ys) ≈ pair(xs,zs), le(x,y) ≈ true
split(x,cons(y,ys)) → pair(cons(y,xs),zs) | split(x,ys) ≈ pair(xs,zs), le(x,y) ≈ false
qsort(cons(x,xs)) → app(qsort(ys),cons(x,qsort(zs))) | split(x,xs) ≈ pair(ys,zs)

Property / Task

Prove or disprove confluence.

Answer / Result

Yes.

Proof (by ConCon @ CoCo 2020)

1 Almost-orthogonal modulo infeasibility

The given (extended) properly oriented, right-stable, oriented 3-CTRS is almost-orthogonal modulo infeasibility, since all its conditional critical pairs are infeasible.