Naqvi, S., Tsur, S.: A Logic Language for Data and Knowledge Bases. Computer Science Press, 1989.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Fifth International Conference on Database Theory (ICDT95) 14] Work partially supported by the Murst 40 projects Data X and D21. S. Greco et al. an operational semantics that is amenable to very efficient implementation as demonstrated by a number of prototypes of deductive database systems [25, 30, 22]. Unfortunately, the basic DATALOG language (without negation and function symbols) is severely limited in its expressive power and cannot express many of the queries of practical interest. While the exact expressive power of DATALOG has not been characterized completely, it has been shown that ....

....fashion: the power can be controlled by the user to select the desired level of complexity which may range from 7 p up to the entire query hierarchy which includes a large number of meaningful problems. 4 Choice and stratification in DATALOG queries The choice construct (supported in 7P [25] and, in some form, in Coral [30] is used to enforce functional dependencies (FD) constraints on rules of a logic program. A rule r with choice constructs, called a choice rule, has the following general format: r: A , B(Z) choice( X) Y) choice( Xk) Yk) where B(Z) denotes the ....

Naqvi, S., Tsur, S.: A Logic Language for Data and Knowledge Bases. Computer Science Press, 1989.

....to the de nition of a stable model semantics for choice. While the declarative semantics of choice is based on stable model semantics which is computationally intractable in general, choice is amenable to ecient implementations, and it is actually supported in the logic database language LDL [14] and its evolution LDL [2] On the other side, strati cation has been a crucial notion for the introduction of nonmonotonic reasoning in deductive databases. From the original idea in [1] of a static strati cation based on predicate dependencies, strati ed negation has been re ned to deal with ....

S. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1989.

....of such databases are fully understood and there has been a great deal of research dealing with implementation issues, particularly in query optimization in the presence of recursive rules. This research has culminated in various experimental systems such as NAIL [12] glue NAIL [15] LDL [1, 14], Aditi [19] EKS V1 [20] CORAL [16] and Starburst SQL [13] the utility of which have been successfully demonstrated. Therefore, it is not unreasonable to assume that within the next decade, commercial systems with deductive capabilities will become available. In the presence of a large number ....

Shamim Naqvi and Shalom Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, Rockville, MD, 1988.

....leading to the definition of a stable model semantics for choice. While the declarative semantics 2 of choice is based on stable model semantics which is intractable in general, choice is amenable to efficient implementations, and it is actually supported in the logic database language LDL [20] and its evolution LDL [2] On the other side, stratification has been a crucial notion for the introduction of nonmonotonic reasoning in deductive databases. From the original idea in [1] of a static stratification based on predicate dependencies, stratified negation has been refined to deal ....

S. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1989. 21

....conjunction of a query under the possibility semantics and a query under the certainty semantics. The result follows from the fact that possibility semantics captures NPand certainty semantics captures coNP , 2 4 Choice and Strati cation in DATALOG Queries The choice construct (supported in LDL [19] and, in some form, in Coral [23] is used to enforce functional constraints on rules of a logic program. Thus, a goal of the form, choice( X) Y) in a rule r denotes that the set of all consequences derived from r must respect the FD X Y . In general, X can be a vector of variables possibly ....

S. Naqvi and S. Tsur. A logic language for data and knowledge bases. Computer Science Press, 1989.

.... graph , whose edges are represented as g(X, Y, C) and g(Y, X, C) for some cost argument C, is constructed as follows: st(nil, a, 0) st(X, Y, C) st( X, g(X, Y, C) X 6= Y, choice(Y, X, C) 2 Similar to least most, the semantics of the meta level predicate choice, supported in LDL ([NT89]) and in CORAL ( RSS92] can be defined by rewriting into non stratified negation with stable model semantics ( SZ90, GZG93] Choice supports a don t care form of non determinism, hence it may be used to express greedy search strategies. On the other hand, it also carries semantics w.r.t. FDs, ....

S. Naqvi, S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1989.

....and so now fit easily in the memories that are becoming available on workstations. So in memory systems will become increasingly useful for certain kinds of applications. All this suggests that it is time to explore serious implementations of these ideas. Work 24 is being done in the DD area [NT89, LV89, PDR91, Ram90] It is appropriate for work to be done in the LP tradition. One advantage of this OLDT approach is its integration with Prolog evaluation techniques. Prolog s strategy, while very simple, can be very fast. Also, much work has been done in compiling and optimizing Prolog ....

S. A. Naqvi and Sholom Tsur. A logic language for data and knowledge bases. Computer Science Press, New York, NY, 1989.

....to the definition of a stable model semantics for choice. While the declarative semantics of choice is based on stable model semantics which is computationally intractable in general, choice is amenable to efficient implementations, and it is actually supported in the logic database language LDL [14] and its evolution LDL [2] On the other side, stratification has been a crucial notion for the introduction of nonmonotonic reasoning in deductive databases. From the original idea in [1] of a static stratification based on predicate dependencies, stratified negation has been refined to deal ....

S. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1989.

....in such visual environments involve pattern matching and finding and traversing paths of various kinds. Such operations are beyond the power of first order query languages such as relational algebra and SQL; they require additional expressive power, as provided by deductive databases such as LDL [12] or CORAL [13] In this paper we show how deductive databases can help support the growing demand for visible and visualizable data and for queries and operations on such data. We do so in the context of the Hy visualization system [7] and the database query language GraphLog [1, 6] As a ....

....implicitly over all variables appearing in the distinguished edge and its endpoints. In this paper we are only concerned with non recursive aggregation. For a review of recursive aggregation in GraphLog see [8] 4 Translating GraphLog This section describes the translation from GraphLog to LDL [12] and CORAL [13] The translation of a program resembles the logical function used to define the semantics of GraphLog in terms of Datalog programs [1, 6] An alternative optimized translation is described in [14] Since Hy is implemented in the Smalltalk object oriented programming language, ....

S. A. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1988.

.... can either give a singleton relation containing the tuple (marc; ohm) or that containing the tuple (marc; bell) 2 Static Choice In Datalog, eligible advisors can be computed by the following rule: elig adv(S; P) student(S; Major) professor(P; Major) Then, using LDL s static choice construct [32], the goal choice( S) P) can be added to the rule forcing the selection of a unique advisor for each student. The goal choice( S) P) denotes that the functional dependency S P must hold in actual adv: Example 2 Computation of unique advisor by choice rules. actual adv(S; P) student(S; ....

....construct for declaring and enforcing FDs in Datalog. Observe that Datalog :s , i.e. Datalog with stratified negation, cannot express a non deterministic query; e.g. it cannot select an arbitrary advisor for a student out of a set of several eligible advisors. The original definition of choice [30, 32] relies on the re writing of the program into one where the original choice rule is broken in three parts. For instance, the program of Example 2 is basically rewritten as follows (using a mixed algebra rules notation where chosen and extchoice denote the binary relations corresponding to the ....

[Article contains additional citation context not shown here]

S. Naqvi and S. Tsur, A logic language for data and knowledge bases, Computer Science Press, New York,1989.

....(Van Gelder [1986] Apt, Blair, and Walker [1988] Naqvi [1986] or well founded semantics (Van Gelder, Ross, and Schlipf [1991] Ross [1991] The systems of this class are generally less advanced toward commercialization, but there are a number of experimental systems under development. LDL (Naqvi and Tsur [1989]) is the first, and most advanced of this class, and can be considered ancestral to all. Other projects of this type include ffl CORAL (Ramakrishnan [1990] at the Univ. of Wisconsin, ffl Aditi (Vaghani et al. 1990] at the Univ. of Melbourne, ffl LOGRES (Cacace et al. 1990] at the Univ. of ....

Naqvi, S. A. and S. Tsur [1989]. A Logic Language for Data and Knowledge Bases, Computer Science Press, New York.

....of the deductive database system LDL [AOTZ93, AOZ93, AO93] to express situations that need spatio temporal reasoning in Sea monitoring. The temporal extension we used is LDLT [DM94] a temporal database proposal that integrates the expressiveness and deductive power of the logic language LDL [NT88] with the possibility to perform temporal reasoning based on the Interval Algebra by Allen [All83] The aim of our work is to realize an efficient integration between Deductive Databases and Geographical Information Systems (GIS) to study the advantages that the integration can bring to both of ....

S. Naqvi, S. Tsur. A logic language for data and knowledge bases. Computer Science Press, New York (1988).

....nodes, one can formulate questions like Is A connected to B and What is the cost of the shortest path between A and B . And in a database storing information about parts, one can express bill of material questions. This functionality has been proposed in the context of logical query languages [6, 17] and in the context of the relational algebra [1, 13] Independent of the context, at the implementation level one needs an algorithm to efficiently The research of Maurice Houtsma has been made possible by a fellowship of the Royal Netherlands Academy of Arts and Sciences; email: ....

Naqvi, S. and Tsur, S. A logic language for data and knowledge bases, CS Press, 1989.

....options to present summaries of the information carried by the hidden edges) 3. 2 Query Processing Query processing in Hy is performed by translating queries (and data, if necessary) into logic programs suitable for execution by one of two back ends: the logic programming language LDL of MCC [24] and the experimental deductive language CORAL of the University of Winsonsin [28] In this section, we give a more precise definition of GraphLog, and we describe how the query processing proceeds within the Hy system. In GraphLog, a term is one of a constant, a variable, an anonymous ....

S.A. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1988.

....optimization algorithm for recursive and nonrecursive queries written in Datalog or SQL. The magic sets transformation has been implemented in database systems for optimizing recursive and nonrecursive queries: Coral [RSS92] implements magic templates [Ram88] Aditi [VRK 90] EDS [FF93] LDL [NT88] NAIL [MUVG86, MNS 87] and Glue Nail [PDR91, DMP93] implement the supplementary magic sets transformation [BR91] and Starburst [MP94] implements magic conditions [MFPR90a] Performance experiments have shown the magic sets transformation to be a good optimization technique, both for ....

Shamim Naqvi and Shalom Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1988.

.... for rule goals as to ensure safe and optimized executions [33, 53, 60] Systems and Applications Many interesting applications of deductive databases have been reported, ranging from the rapid prototyping of information systems and scientific databases, to data cleaning and stock market analysis [30, 51, 37]. More recent applications include websearching [25] integration and mediation of heterogeneous information systems [32, 2] and GUI generation [17] Deductive database languages support rapid prototyping via rules that express both domain knowledge (either user encoded or discovered) and ....

S. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, New York, 1989.

....in such visual environments involve pattern matching and finding and traversing paths of various kinds. Such operations are beyond the power of first order query languages such as relational algebra and SQL; they require additional expressive power, as provided by deductive databases such as LDL [12] or CORAL [14] In this paper we show how deductive databases can help support the growing demand for visible and visualizable data and for queries and operations on such data. We do so in the context of the Hy visualization system [6] and the database specification and query language GraphLog ....

....non recursive aggregation. For a review of recursive aggregation in GraphLog see [5] GraphLog s non recursive aggregation is equivalent in expressive power to what Mumick et al. call stratified aggregation[11] 4. From GraphLog to LDL This section describes the translation from GraphLog to LDL [12]. The translation of a program essentially resembles the logical function used to define the semantics of GraphLog in terms of Datalog programs [4] 4.1. Classifying Terms A database schema is a description of the stored data and its organization in facts. The schema specifies among other things ....

S. A. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1988.

....optimization algorithm for recursive and nonrecursive queries written in Datalog or SQL. The magic sets transformation has been implemented in deductive database systems for optimizing recursive queries: Coral [RSS92] implements magic templates [Ram88] Aditi [VRK 90] EDS [FF93] LDL [NT88] NAIL [MNS 87] and Glue Nail [DMP93] implement the supplementary magic sets transformation [BR91] Part of the work of this author was done at IBM Almaden Research Center and Stanford University. However, the deductive implementations have ignored several aspects critical to commercial ....

....implementation, and this paper describes the results of our implementation effort. The magic sets implementation in Starburst has goals and addresses issues very different from the goals and issues addressed by previous and concurrent deductive database implementations (Aditi [VRK 90] LDL [NT88] NAIL [MNS 87] Glue Nail [DMP93] Coral [RSS92] and EDS [FF93] The Starburst implementation is done in the context of an extensible relational system, with extended SQL as the query language. With the exception of EDS [FF93] all other implementations of magic sets have been done for ....

[Article contains additional citation context not shown here]

S. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1988.

....or if it contains non ground updates, the transaction is aborted and no update in the set is performed. The notion of consistency is an important one, in that it prevents a set of updates containing both an insertion and a deletion of the same fact to be executed. By contrast in DLP [25] LDL [30] and DL [3] updates are executed as soon as they are evaluated. This approach leads to complex semantics and to computations performed in a sequence of states instead of in a single one. In the following we recall the language and its informal behavior (see [29] for a complete description) We ....

S. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1989.

.... in [Mai86] and refined in [KKS92, CKW93, KL89] Automatic creation and manipulation of object id s based on Skolem functors are considered in depth in [HY90] It is observed in [AK89] that object id based set formation (as provided by the object id based fusion) can replace explicit (LDL like [NT88] grouping operators. They also advocate a ptime sublanguage by prohibiting recursion through object creation. The architecture we use prevents such potentially dangerous expensive form of recursion: objects are created in a mediator based on the objects in lower level sources (that can ....

S. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1988.

.... for defining and efficiently evaluating recursive views, thus significantly extending the class of queries that can be expressed, by providing structured data in the form of Prolog terms and sometimes sets, and by using a logic programming language as the uniform query and application language [9, 12, 29, 36, 37]. Their most appealing property is that they preserve the firm logical foundation of relational systems, and much work has been done on their application to problems involving recursive or cyclic data, on the semantics of databases with increasingly complex combinations of recursion, negation and ....

S. Naqvi and S. Tsur. A Logic Language for Data and Knowledge Bases. Computer Science Press, 1989.

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