David Turner, `Functional programming and communicating processes', PARLE, Parallel Architectures and Languages Europe, Eindhoven, The Netherlands, June 1987, pp. 54--74, Springer-Verlag, LNCS 259.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....an alternative rewrite semantics for the metalanguage. In Sect. 7.1 we will successively extend the metalanguage with constructions for data fields and bounds, including abstraction, and we extend M to cover the extended languages. 4. 2 Hyperstrictness Hyperstrictness was first defined by Turner [67]. His definition was informal. We give more stringent definitions of hyperstrictness and related concepts below. Definition 1 For any element d in an Eq cpo D and for any element d 0 in an appropriate cpo, we define the relation in by: d 0 in d iff: ffl d 0 = d, ffl d = d 1 ; d 2 ) ....

D. Turner. Functional programming and communicating processes. In Proc. PARLE'87 vol. 2, Volume 259 of Lecture Notes in Comput. Sci., pages 54--74, Berlin, 1987. Springer-Verlag.

....such a system can be structured. 3 PURE NON DETERMINISM 10 , V . kernel process Figure 4: Process Network for the kernel and one process 3.3. 1 System Structure Turner [21] demonstrates how to use non deterministic merge to write an operating system. The system is structured as a process network, as in section 2.1.2. One process is the sorting office, or kernel , that co ordinates communication between the other processes; the kernel is the only process that needs ....

David Turner, "Functional Programming and Communicating Processes," in LNCS, PARLE'87, Vol 2, pp. 54-74.

....is pushed to the sorting office, leaving processes referentially transparent, so equational reasoning can always be applied to each one separately. This makes the system non deterministic as a whole. This idea has been considered the best available compromise and has been used also by Turner [128] for the KAOS project. 5.2.3 Timestamping Another solution to the problem of merging streams of information would be to turn a non deterministic merge into a deterministic one by attaching timing information to items coming from the input streams. This approach has been used in the programming ....

David Turner. Functional Programming and Communicating Processes. In J.W. Bakker, A.J. Nijman, and P.C. Treleaven, editors, PARLE II, volume 259 of LNCS, pages 54--74, Eindhoven, The Netherlands, June 1987. Springer Verlag.

....typechecking altogether in numerous places. Recent developments in type systems do allow the specification and checking of existential types, but only to a limited degree [1] Turner points out that Message could be an abstract type, whose representation is known only to the runtime system [10]. He suggests that a series of wrapper functions be created, each of which casts a value of some type to a Message : type Wrapper omsg = omsg Message type Process imsg = imsg] Message] Each process in the system is associated with its own individual wrapper function, provided by the ....

David A Turner, Functional programming and communicating processes, pp.54-74, Proceedings of PARLE '87, Springer-Verlag LNCS 259, 1987.

.... an ingenious mechanism based on a standard type system that limits the type insecure constructions to the message delivery system itself, but processes are restricted to receiving messages of a fixed type which must be manifest in the type signature of the function implementing the process [7, 60, 61]. These restrictions effect the packaging and ease of use of the IPC mechanism, as each service used by the program would have to be incorporated into a single union type, the message type accepted by the process. As Cupitt s thesis bears out [7] and in contrast to our proposal, the task of ....

D. A. Turner. Functional programming and communicating processes. In J. W. de Bakker, A. J. Nijman, and P. C. Treleaven, editors, Parallel Architectures and Languages Europe (PARLE), pages 54--74. Springer-Verlag, June 1987. Volume 2 (LNCS 259). (p 64)

.... loss of referential transparency; an expression does not necessarily have the same value each time it is evaluated, and so equational reasoning cannot be used, as explained by Clinger [3] One way to reduce the effects of this is to use the sorting office model of Stoye [11] also used by Turner [12], where functional processes send messages to each other via a central delivery service called a sorting office. Each process is purely functional, and the non determinism is confined to the sorting office and communication channels. A different approach by Burton [1] is to record choices; every ....

D. Turner, "Functional programming and communicating processes" LNCS 259, 1987.

No context found.

David Turner, `Functional programming and communicating processes', PARLE, Parallel Architectures and Languages Europe, Eindhoven, The Netherlands, June 1987, pp. 54--74, Springer-Verlag, LNCS 259.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC