The intriguing Linda language, developed at the Yale Department of Computer Science, is reportedly being used in, of all places, Bogata, Colombia to create a parallel processor out of a network of Apple Macintosh SE computers. Linda first surfaced with the news that researchers at the Sandia National Laboratory had used the parallel language to create a complex of 14 DEC VAXes at two sites over 1,000 miles apart that outperformed the lab’s Cray 1 supercomputer on a modelling problem (CI No 967). According to David Gelernter, who developed the language at Yale, Linda is both machine- and language-independent, and there are currently versions for use with programs written in C, Fortran, Lisp, and Modula II. According to the Newsbytes newswire Linda adds extensions to the language that speed up communications and co ordination in the program, using a concept called tuple space. Tuple space, says Gelernter, is more like a bulletin board than conventional communication within programs, which he describes as more like a telephone. You don’t have to dial a number or have an address, and you don’t have to synchronise with the recipient of the message. You can tack the data to the bulletin board whenever you want, and the reader can read it whenever he wants. Linda is to be marketed as a product by Scientific Computing Associates, New Haven, Connecticut.
This article is from the CBROnline archive: some formatting and images may not be present.
CBR Online legacy content.