L. M. Pereira, L. Monteiro, J. Cunha, and J. N. Aparicio. Delta prolog: A distributed backtracking extension with events. In E. Y. Shapiro, editor, Proceedings of the Third International Conference on Logic Programming, volume 225 of Lecture Notes on Computer Science, pages 69--83. Springer-Verlag, 1986.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....on a blackboard. Similarly, initial goals are requested to share no variables. This design decision initially imposed by the purity of the model has turned out to be crucial for implementability purposes. Among others, problems caused by distributed backtracking (see e.g. Delta Prolog problems in [11]) are avoided. All blackboards created by the reduction of an application inherit the program of the application. This explains why the tellbb does not specify any program. Refinements could include the definition of the programs as sets of modules (as in [10] from which tellbb could extract ....

L.M. Pereira, L. Monteiro, J. Cunha, and J.N. Aparcio. Delta Prolog : A Distributed Backtracking Extension with Events. In G. Goos and J. Hartmanis, editors, Proceedings of the Third International Conference on Logic Programming, volume 225 of Lecture Notes in Computer Science, London, July 1986. Springer Verlag.

....the best of both worlds, achieving maximum parallelism but still allowing all solutions to be found. For several years, combining the two was deemed impractical. Recently, however, a number of algorithms have been published that combine stream AND parallelism and backtracking [Cleary et al. 1988, Pereira et al. 1986, Somogyi et al. 1988, Tebra 1987] Central to all of these algorithms (either implicitly or explicitly) is the notion of imposing a total ordering on the goals executed by the system. The natural temporal goal ordering of sequential Prolog is what allows it to backtrack successfully, but in a ....

....on the value of a binding, the binding made by the higher precedence goal is accepted; the lower precedence goal must retract its binding and recompute. When a lower precedence goal can find no solution, it may ask a higher precedence goal to recompute its bindings. Some systems, like those of [Pereira et al. 1986] and [Tebra 1987] retain the depth first ordering of sequential Prolog. Others, like that of [Somogyi et al. 1988] base the ordering on producer consumer relationships between goals. The least restrictive algorithm is that of [Cleary et al. 1988] any ordering will suffice. Our algorithm is based ....

L.M. Pereira, L. Monteiro, J. Cunha, and J.N. Aparcio. Delta Prolog: a distributed backtracking extension with events, In Proceedings of the Third International Conference on Logic Programming, pages 69--83, 1986. published as Lecture Notes in Computer Science 225 by Springer-Verlag.

....the run time overhead of marking. Comparisons with an implementation based on the B list algorithm [35] show favourable results, with the number of operations required for backtracking dramatically reduced under some circumstances. Various other intelligent backtracking schemes have been proposed [6, 34, 40, 50], which can reduce the amount of computation in an AND parallel system. These schemes usually rely on the information of the producer consumer relationships obtained at both compile time and run time, in order to determine a backtracking point. The algorithms used are typically very complicated. ....

Pereira, L.M. et al. Delta-Prolog: A Distributed Backtracking Extension with Events. In Proc. 3rd Int. Conf. on Logic Programming, July 1986, 67-83.

....learning (ML) techniques are to be used to guide the process of developing increasingly complex theories of behaviour. 2 The Eel Programming Language To deal with concurrency, the declarative programming community has developed programming languages with implicit [Gre87] as well as explicit [PMCA86, Gre97, Llo98] process communication. Eel [Dah98] takes the explicit approach and is built on a limited model of communication which only allows synchronous point to point communication and which doesn t have the concepts of sending and receiving messages but instead relies on message unification as the method ....

L. M. Pereira, L. Monteiro, J. Cunha, and J. N. Aparicio. Delta prolog: A distributed backtracking extension with events. In E. Y. Shapiro, editor, Proceedings of the Third International Conference on Logic Programming, volume 225 of Lecture Notes on Computer Science, pages 69--83. Springer-Verlag, 1986.

....parallel execution models. Gupta et al. [46] point out that there are many proposals for extending a logic programming language with explicit constructs for concurrency. They are largely put into three categories: 1. those that add explicit message passing primitives to Prolog e.g. Delta Prolog [72] and CS Prolog [41] Multiple processes are run in parallel that communicate with each other via messages. 2. those that add blackboard primitives to Prolog, e.g. De Bosschere et al. [26] and Shared Prolog [18] These primitives are used by multiple Prolog process running in parallel to communicate ....

L. M. PEREIRA, L. MONTEIRO, J. CUNHA, and J. APAR'iCIO. Delta Prolog: a distributed backtracking extension with events. In E. Shapiro, editor, Third International Conference on Logic Programming, LNCS 225, pages 69--83. Springer-Verlag, 1986.

....placed on a blackboard. Similarly, initial goals must not share variables. This design decision initially imposed by the purity of the model has turned out to be crucial for implementability purposes. Among others, problems caused by distributed backtracking (see e.g. Delta Prolog problems in [19]) are avoided. R definitions are used to define blackboards relations. Definition 3 1) In primitives are primitives of one of the following forms: in(bbn; O) int(bbn; t) inp(bbn; p) where bbn is a blackboard name, t is a term, p is a goal, O is either a term or a goal. Inc primitives are ....

L.M. Pereira, L. Monteiro, J. Cunha, and J.N. Apar'icio. Delta Prolog : A Distributed Backtracking Extension with Events. In E.Y. Shapiro, editor, Proc. Third International Conference on Logic Programming, volume 225 of Lecture Notes in Computer Science, London, July 1986. Springer Verlag.

....disjunction. In this case, the set of parallel processes must terminate successfully in order to make the conjunction or disjunction terminate successfully. Data shared between the two components of a conjunction or a disjunction require extra synchronization. The process creation of Delta Prolog [6, 14] is based on and parallelism. PMS Prolog [18] has a rather unusual mechanism of process creation: it forks a list of processes, and then waits until the child processes have all terminated before resuming. Forking one process is comparable to executing a remote procedure call. CS Prolog [9] ....

....for a set of messages. Each process needs at least two messages in order to proceed. Reading the messages sequentially without synchronization might give rise to a deadlocked system. 6 Languages mentioned Name type Refs. BlackLog blackboard [15] CS Prolog channel [9] Delta Prolog channel [6, 14] Prolog D Linda blackboard [16] MB Prolog channel [17] Log blackboard [11] Multi Prolog blackboard [8] PMS Prolog channel [18] Polis Prolog blackboard [5] Shared Prolog blackboard [3] Sicstus Prolog blackboard [1, 4] Polka blackboard [7] Quintus Prolog channel [13] List of process based parallel ....

L.M. Pereira, L. Monteiro, J. Cunha, and J.N. Apar'icio. Delta Prolog : A Distributed Backtracking Extension with Events. In G. Goos and J. Hartmanis, editors, Proceedings of the Third International Conference on Logic Programming, volume 225 of Lecture Notes in Computer Science, London, July 1986. Springer Verlag.

....have greater message numbers than their ancestors. An offspring event takes place along the same channel as its ancestors and must be preceded by them. Each instance of process synchronisation and information exchange is seen as an event. This approach has been taken before in Delta Prolog [10], which also had a complete backtracking mechanism for internal events. In Eel, communication with external processes is handled by the same abstraction. The algorithm that is used to find a set of real communication events that correspond to the restrictions given in the program, is inspired by ....

L. M. Pereira, L. Monteiro, J. Cunha, and J. N. Aparicio. Delta prolog: A distributed backtracking extension with events. In E. Y. Shapiro, editor, Proceedings of the Third International Conference on Logic Programming, volume 225 of Lecture Notes on Computer Science, pages 69--83. SpringerVerlag, 1986.

....users with the same abstraction, that of communication events. In addition, Eel extends the traditional, deductive Prolog proof procedure with abduction of terms formed using a set of reserved predicates. The idea of making process communication in Prolog explicit was introduced in Delta Prolog [11]. The idea of using events to synchronize concurrent processes and perform communication was also introduced in the Tempo language [4] where they are also used to reason about a systems safety properties. The IFF proof procedure [3] expanded the classical deductive proof proocedure with ....

L. M. Pereira, L. Monteiro, J. Cunha, and J. N. Aparicio. Delta prolog: A distributed backtracking extension with events. In E. Y. Shapiro, editor, Proceedings of the Third International Conference on Logic Programming, volume 225 of Lecture Notes on Computer Science, pages 69--83. Springer-Verlag, 1986.

....offered by CO OPN objects, the objective is to have a convenient way of abstracting away the behaviour of external components. This modelling facility is also supported by a solid formal basis [14] In distributed systems, however, synchrony is difficult to 1. To our knowledge, only Delta Prolog [20] and CS Prolog [13] offer built in distributed backtracking. 1. Here synchronous is not taken in the sense of blocking method calls. simulate, because of the lack of absolute time reference and the non negligible communication medium. Some work has been done to implement Esterel on local area ....

L.M. Pereira, L. Monteiro, J. Cunha, and J.N. Aparício, "Delta-Prolog: A Distributed Backtracking Extension with Events", in G. Goos and J. Hartmanis, editors, Proceedings of the Third International Conference on Logic Programming, volume 255 Lecture Notes in Computer Science, Springer Verlag, London, July 1986.

No context found.

Pereira, L.M.; Monteiro, L.; Cunha,J.C.; Apar'icio, J.N. Delta Prolog: a distributed backtracking extension with events, in "3rd Int. Logic Programming Conf., Lecture Notes in Computer Science 225 (E.Shapiro ed.), Springer

....(HCL) with constructs for the specification of distributed systems. As a distinctive feature, the DeltaP model relies on the programmer to specify the sequentiality constraints and the desirable parallelism existing in each problem, together with the corresponding communication schemes ( PP84] [PMCA86], PMCA88] CFP89b] CFP89a] The presentation is organized as follows. In section 2, we discuss the programming model and give an example. A discussion of the DeltaP sequential and parallel execution models is made in section 3. Then, in section 4, we discuss implementation issues of the ....

L.M. Pereira, L.F. Monteiro, J.C. Cunha, and J.N. Apar'icio. Delta-Prolog: a distributed backtracking extension with events. In 3rd Int. Conf. on Logic Programming, pages 69--83. Springer Verlag, 1986.

No context found.

L. M. Pereira, L. Monteiro, J. Cunha, and J. N. Aparicio. Delta prolog: A distributed backtracking extension with events. In E. Y. Shapiro, editor, Proceedings of the Third International Conference on Logic Programming, volume 225 of Lecture Notes on Computer Science, pages 69--83. Springer-Verlag, 1986.

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