Software/parallel processing-Linda on the networks

The use of the coordination language Linda to facilitate the division of software tasks among several processors in parallel programming is considered. Linda´s coordination facilities depend on a view of memory as object-oriented, associative, symmetric, and virtually shared, i.e., data are stored as simple data objects, with type signature and other properties, rather than as bits and bytes. To produce a parallel version of a language like C or Fortran, Linda adds just four major operations-in, out, rd, and eval.-to that language. The base language does the chores, such as arithmetic, loop control, I/O, and so on, for which it was designed, while Linda performs only coordination functions: process creation, interprocess communication and control.<>