Adding threads to Standard ML

We have added multiple threads of control to the Standard ML programming language. Standard ML's support for first-class functions and automatic storage management influenced the design in a number of ways. We demonstrate how other concurrency and synchronization operations, such as cobegin/coend, futures, and events, can be implemented in terms of the thread interface. Finally, we describe three implementations of the thread interface: a coroutine version, a uniprocessor preemptive version, and a multiprocessor Mach-based version. This research was sponsored in part by the Defense Advanced Research Projects Agency, Information Science and Technology Office, under the title "Research on Parallel Computing," ARPA Order No. 7330, issued by DARPA/CMO under Contract MDA972-90-C-0035, and in part by a National Science Foundation Graduate Fellowship. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the U.S. Government