Anyone with experience with JMP modeling reliabilitty growth with various development phases, various sample sizes and various test times and where not all failures have an engineering fix but are instead repaired.
I'd appreciate any white papers or comments.
Thanks