Abstract
A language for describing finite and infinite networks of loosely coupled, concurrent, nondeterministic, communicating agents is introduced. To every program a finite or infinite graph ("network") is related representing graphically the communication structure of the described system. A denotational semantics is defined based on fixed point theory. Algebraic laws for the networks are studied that allow to transform them without changing their denotational meaning. Following the increasing complexity of denotational models for stream-processing networks a hierarchy of five languages is treated: first a language of finite, deterministic networks, then infinite (i.e., recursively defined), deterministic ones, then nondeterministic finite and nondeterministic infinite networks with free choice merge. Finally we study a language including fair, nonstrict merge.
| Original language | English |
|---|---|
| Pages (from-to) | 13-31 |
| Number of pages | 19 |
| Journal | Distributed Computing |
| Volume | 2 |
| Issue number | 1 |
| DOIs | |
| State | Published - Mar 1987 |
| Externally published | Yes |