This brings back memories. I used xlispstat for my engineering thesis on pruning of neural networks in 1992. I enjoyed it very much. I still have Luke Tierney's book LISP-STAT. I also used Splus, the ancestor to R, which was also very good, but not open source.