Y. Moses and O. Kislev. Knowledge-oriented programming. In Proceeding of the 12th Annual ACM Symposium on Principles of Distributed Computing (PODC 93), pages 261-- 270, New York, USA, Aug. 1993. ACM Press.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... that support the analysis and verification of distributed systems in terms of knowledge, e.g. Rosenschein and Kaelbling, 1986 ] It has also been proposed that there should be knowledge based programming languages, containing prim2 itive constructs at the knowledge level [ Fagin et al. 1995a; Moses and Kislev, 1993; Shoham, 1993 ] Both applications require answers to the following question: given a description of a distributed system, how do we efficiently compute the answer to a query about the knowledge of the agents (processors) in a given state Closely related is the question of update: how should an ....

....upper bound of PSPACE. This type of analysis has shown the appropriateness of knowledge as a level of abstraction for the design of distributed systems. Consequently, the possibility of high level programming languages with explicit operators for knowledge is a current topic of investigation [ Moses and Kislev, 1993; Fagin et al. 1995a; Fagin et al. 1995b ] In order to permit the execution of knowledge based programs, such a language would have to support the compilation of knowledge formulae into standard functions of the state of the processor. Our work in this paper may be viewed as an initial foray ....

Y. Moses and O. Kislev. Knowledge oriented programming. In Proc. Symposium on Principles of Distributed Computing, 1993.

.... the effect it has on the mental states of the participants of the conversation [ Allen and Perrault, 1980; Cohen and Levesque, 1990; Cohen and Perrault, 1979; Perrault, 1990 ] Another potential application is the semantics of agent knowledge oriented programming languages [ Fagin et al. 1994; Moses and Kislev, 1993; Shoham, 1993 ] which include as primitive constructs modal operators for knowledge and belief. One can study the dynamics of belief under a variety of assumptions about the nature of agents and their environment. We work in this paper with a very specific set of assumptions. Prototypical of ....

Y. Moses and O. Kislev. Knowledge oriented programming. In Proc. Symp. on Principles of Distributed Computing, 1993.

....number of faulty processes. Our work proves that, when processors don t exhibit arbitrary failures, no such bound is necessary to develop a fast stopping TRB protocol. The application of knowledge theory [5] in distributed computing led to many interesting results related to the topic of this paper[4, 8, 9, 12, 10, 14]. From a knowledge theory perspective, we extended the concept of clean rounds [4] to systems with general omission failures and consider a weaker condition for cleanness, as a locally clean round provides as much knowledge as a systemwide clean one. More specifically, to deliver, a process uses ....

Y. Moses and O. Kislev, "Knowledge-Oriented Programming", In In Proc. of the 12th PODC, 261-270, 1993.

....with IO Regular Expressions for Data ow Networks 37 4. sequence: If a term is (s 1 t s 2 t : t s m ) t and 9k : s k = t 1 t 2 , then t 2 t. We call (1) 4) determinacy conditionsD It is known that a data ow network whose behavior is not decided uniquely has the undesired property [11, 12] of not being monotone. If an IO regular expression satis es the determinacy conditions, then the output is decided uniquely for the same present input, latest input and latest output. Suppose In(Out) is an input(output) channel set of process P, as follows: In = fin 1 ; in 2 ; in k g,Out ....

....semantics of object oriented programming languages. Operational semantics describes the meaning of a programming language by specifying how programs expressed using it execute on an abstract machine [16] The evaluation and execution relations are speci ed by rules in a way directed by the syntax [11]. Operational semantics itself has two di erent approaches, mainly natural semantics and structural operational semantics. Schmidt [13] discusses the di erence between the two approaches. The structural operational semantics is also known as small step semantics and the natural semantics is ....

[Article contains additional citation context not shown here]

Y. Moses and O. Kislev. Knowledge-oriented programming. In Proc. 12th ACM Symp. on Principles of Distributed Computing, pages 261-270, 1993.

....run that makes f true, if there is such a point. 4. Example: Two Phase Commit To illustrate our framework we consider a problem from the area of distributed databases. Twoand three phased commit protocols have previously been considered (though less formally) from a knowledge based perspective in [11, 22, 16]. Problem. Suppose n agents collaborate on a database transaction. After each of the agents has done its share of the transaction, they want to come to an agreement on whether to commit the transaction for later use of the result by other transactions. Each agent can be assumed to have formed an ....

Y. Moses and O. Kislev. Knowledge-oriented programming. In Proceeding of the 12th Annual ACM Symposium on Principles of Distributed Computing (PODC 93), pages 261-- 270, New York, USA, Aug. 1993. ACM Press.

....that enables some quite useful refinement rules. Finally, we will in future show how to extend the framework developed here to include quantification over local predicates [EvdMM98,EvdMM99] this provides a framework generalizing knowledge based [FHMV95] and knowledge oriented programs [MK93]. ....

Y. Moses and O. Kislev. Knowledge-oriented programming. In Proc. 12th ACM Symp. on Principles of Distributed Computing, pages 261--270, 1993.

....LAT E Xed on January 20, 2000 2 2 A Semantics for Reasoning about Knowledge and Time implementations. As a specification formalism, however, knowledge based programs are unbalanced, abstracting only the tests performed by agents, but providing no abstraction mechanism for their actions [11]. Action abstraction is handled much better in refinement calculi [1, 9, 10] also known as broad spectrum languages. Such calculi view programs and specifications as having the same semantic type, and support a formal methodology for the development of programs that are correct by design , ....

Y. Moses and O. Kislev. Knowledge-oriented programming. In Proceeding of the 12th Annual ACM Symposium on Principles of Distributed Computing (PODC 93), pages 261--270, New York, USA, Aug. 1993. ACM Press.

....in the full paper. Two extensions of the idea of knowledge based programs appear in the literature. One is the notion of programs based on algorithmic knowledge [HMV94] in which knowledge is assumed to be computed by a specific algorithm. Another is the notion of knowledge oriented programs [MK93]. Recall that the actions in knowledgebased programs are standard actions, just as in standard programs. Based on an earlier draft of this paper and of [FHMV95] Moses and Kislev use the notions of contexts and programs very much in the way we do here to define a formalism in which the meaning of ....

....way we do here to define a formalism in which the meaning of actions is also allowed to be knowledge based. One example of such an action is notify(j; which means that the acting agent must ensure that K j holds at some point in the past or future. Knowledge oriented programs are shown in [MK93] to provide transparent high level descriptions of protocols in a number of applications. This paper also introduced the notion of a context in which a program is executed. By doing so, we are able to reason about executing the same program, or the same knowledge based program, in different ....

[Article contains additional citation context not shown here]

Y. Moses and O. Kislev. Knowledge-oriented programming. In Proc. 12th ACM Symp. on Principles of Distributed Computing, pages 261--270, 1993.

....assertions that enables some quite useful refinement rules. Finally, we will in future show how to extend the framework developed here to include quantification over local predicates [EvdMM98] this provides a framework generalizing knowledge based [FHMV95] and knowledge oriented programs [MK93]. Acknowledgments Work supported by an Australian Research Council Large Grant. Thanks to Kai Engelhardt for helpful discussions on the topic of this paper. ....

Y. Moses and O. Kislev. Knowledge-oriented programming. In Proc. 12th ACM Symp. on Principles of Distributed Computing, pages 261--270, 1993.

....are based on the intuition that what an agent does depends on what it knows. They have already met with some degree of success, having been used in papers such as [Dwork and Moses 1990; Hadzilacos 1987; Halpern, Moses, and Waarts 1990; Halpern and Zuck 1992; Mazer and Lochovsky 1990; Mazer 1990; Moses and Kislev 1993; Moses and Tuttle 1988; Neiger and Bazzi 1992; Neiger and Toueg 1993] both to help in the design of new protocols and to clarify the understanding of existing protocols. However, as we show here, there are cases when, used naively, knowedge based programs may exhibit counterintuitive behavior. We ....

....with no epistemic operators in the scope of the counterfactual. It suffices for dealing with simple programs. The second is a generalization that works for arbitrary programs. The only other use of counterfactuals in distributed computing that we are aware of is in the work of Kislev and Moses [Moses and Kislev 1993; Kislev 1991] on knowledge oriented programs. They define a high level action called Notify(R; which is intended to ensure that R knows . 2 The implementation they had in mind implicitly involves counterfactuals; it says that the notifier S can stop sending messages when it knows that even ....

[Article contains additional citation context not shown here]

Moses, Y. and O. Kislev (1993). Knowledge-oriented programming. In Proc. 12th ACM Symp. on Principles of Distributed Computing, pp. 261--270.

....and semantics within which this approach can be represented. Knowledge based programs resemble ordinary programs in their syntax, but they are better viewed as constituting abstract specifications of the way a process concrete actions relate to its state of knowledge. Knowledge oriented programs [MK93] take the level of abstraction one step further, allowing abstract actions such as notify whose purpose is to effect changes in processes states of knowledge. We aim to generalize the frameworks of knowledge based and knowledgeoriented programs. Inasmuch as these are a hybrid of program and ....

Y. Moses and O. Kislev. Knowledge-oriented programming. In Proc. 12th ACM Symp. on Principles of Distributed Computing, pages 261--270, 1993.

....Algorithmic programs can be viewed as a halfway point between knowledgebased programs and standard programs, since, although they have tests for knowledge, these tests are, in a precise sense, guaranteed to be implementable. An extension of the framework of knowledge based programs is presented in [MK93]. Moses and Kislev argue that actions, as well as a program s internal tests, should be thought of at the knowledge level. The effect of sending a single message in a context with reliable communication can be considered similar to sending many messages in an unreliable context. As a result, they ....

Y. Moses and O. Kislev. Knowledge-oriented programming. In Proc. 12th ACM Symp. on Principles of Distributed Computing, pages 261--270, 1993.

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