The OCaml implementation is what we use internally to test implementations of complicated new data structures, and as such, it's quite unreadable to anyone else! We will probably talk at CUFP about our experiences of OCaml. It definitely has some upsides, but is not without some fairly major downsides (concurrency, serialization, lack of consistently-named library functions, ...) . As such, we're rewriting our basic data structures in Scala - this seems like it keeps most of the benefits but without some of the major problems. Now we know how to implement the structures, we might be able to do a clean implementation which can be released!