It sounds like I should write it up as a SRFI, independently of the R6RS process. It seems like I might as well add Kent's graph isomorphism predicate to the SRFI. Is that okay with you, Kent? Will