During the next few years (1987-1989), development of the system by its originators tapered off, and the gap thus created was filled by others who built a variety of interesting and useful systems that either used parts of its code or were inspired by p4. The version included in the book [4] had only rudimentary support for message passing in Fortran, so that was added by Dave Liebfritz at Argonne. Bob Beck at Sequent did a C++ version of the monitors part [1]. Rolf Hempel at GMD greatly improved the Fortran interface and added extensive functionality on top of it for grid-based computation. He borrowed the word PARMACS, and what is now called PARMACS is his system, widely used in Europe [3]. Robert Harrison of Argonne's Chemistry Division re-implemented the message-passing subsystem to provide more efficiency, better error handling, and a library of global operations in his system TCGMSG [13]. Many of his enhancements were later incorporated into p4. The SPLASH group at Stanford built their collection of shared-memory applications [25] on monmacs, adding instrumentation macros to support assorted research projects in parallel programming.