The ALPS programming language is based on a shared variable model of concurrency where the shared variables have built-in data synchronization. Two classes of shared variables that support fine-grain parallelism are provided, namely, cells and sets. A cell is a single-assignment variable which can be assigned only once, but can be read several times by parallel processes. A set is a distributed data structure that can be operated by parallel processes by inserting and removing items. The author describes the language notation for cells and sets and illustrates their use through examples.>

The synchronization and scheduling mechanisms used in an object-oriented concurrent programming language called ALPS are described. The main contributions of ALPS are the concept of a manager process and the concept of a hidden procedure array. An object can have a special high-priority process called a manager which intercepts entry calls and implements the necessary synchronization and scheduling for the object. The manager can be programmed to implement any form of preprocessing and postprocessing for the entry calls and to monitor the object. The concept of a hidden procedure array allows an entry procedure to be exported as a single procedure but implemented as a procedure array. When multiple calls to the entry procedure arrive at the object, each call gets attached to a different element of the hidden procedure array. This simplifies the programming of concurrency within an object. The language mechanisms are illustrated using several examples.>