Ullman, J. Bottom-up beats top-down for Datalog. In Proceedings of the 8th ACM Symposium on Principles of Database Systems (1989) 140--149. 45  Home/Search   Document Not in Database   Summary   Related Articles   Check

....(calculus and algebra) are equivalent. Keywords: Logic Programming, Functional Logic Programming and Deductive Databases. 1 Introduction Deductive databases are database management systems based on logic programming and functional programming [18, 15] Many languages, such as DATALOG [21] and FDL [17] among others, have been developed around this research area. In most deductive database systems, the database model deals, as the classical relational model, with relations de ned on basic values (for instance, strings, integers, etc. wherein these relations are in rst normal ....

J. D. Ullman. Bottom-up Beats Top-down for Datalog. In Proc. of PODS, pages 140-149. ACM Press, 1989.

....shortest path and of minimal spanning tree. If one ignores the abstract notion of run time, our logic programming model is not particularly novel. The language is variant of bottom up logic programming. Bottom up logic programming has been widely studied in the context of deductive databases [13, 11, 6, 12]. Bottom up logic programming is closely related to memoing or tabling for Prolog programs [10, 9, 1] The bottom up language described here allows deletion. Our notion of deletion is superficially similar to widely studied notions of negation in logic programming such as well founded ....

J. Ullman. Bottom-up beats top-down for datalog. In Proceedings of the Eigth ACM SIGACT-SIGMOD-SIGART Symposium on the Principles of Database Systems, pages 140--149, March 1989.

....(for instance, support for non rst normal form and recursive relations) and they are suited for application in which a large number of data must be accessed and complex queries must be supported. With respect to the operational semantics, most deductive database systems (for instance, DATALOG [19], CORAL [16] ADITI [20] use bottom up evaluation instead of top down one like Prolog systems. The reason is that the bottomup approach allows to use a set at a time evaluation, i.e. it processes sets of goals, rather than proceeding one (sub) goal at a time, where operations like relational ....

J. D. Ullman. Bottom-up Beats Top-down for Datalog. In Procs. PODS'89, pages 140-149. ACM Press.

....computation purposes, such as a typical Prolog program, it is common to only need one answer to the query. The author has been partially supported by the Spanish CICYT (project TIC 980445 C03 02 TREND) One consequence of this is that most deductive database systems (for instance, DATALOG [25], CORAL [20] ADITI [26] use bottom up evaluation methods instead of top down one. Bottom up approach allows us to use set at a time evaluation, i.e. it processes sets of goals, rather than proceeding one (sub) goal at a time, where operations like relational joins can be made for disk resident ....

J. D. Ullman. Bottom-up Beats Top-down for Datalog. In Procs. of PODS'89. pp. 140-149. ACM Press.

....query, it is customary to want to compute all of the answers to the query. When writing a program for computation purposes, such as a typical Prolog program, it is common to only need one answer to the query. One consequence of this is that most deductive database systems (for instance, DATALOG [Ull89] CORAL [RSSS94] ADITI [VRK 94] use bottom up evaluation methods instead of top down one. Bottom up approach allows us to use set at a time evaluation, i.e. it processes sets of goals, rather than proceeding one (sub) goal at a time, where operations like relational joins can be made for ....

J. D. Ullman. Bottom-up Beats Top-down for Datalog. In Procs. of PODS. pp. 140-149. ACM Press, 1989.

....rather, they are used as variable length subrecords. For instance, suppose the following complex term address(Name,City)which can be used in the base relation person(DNI,address (Name, City) Job) in order to structure the stored information. Most deductive database systems (for instance, DATALOG [20], CORAL [14] ADITI [21] LOLA [23] allow to handle negative literals, increasing their expressive power as query languages. The introduction of negation in logic programming (see [4] for a survey) and thus the study of semantic models for logic programs have been widely studied in the past ....

J. D. Ullman. Bottom-up Beats Top-down for Datalog. In Proc. of PODS, pages 140-149. ACM Press, 1989.

.... value at compile time) expression simplification (replacing an expression by an equivalent but more e#cient one) and common subexpression evaluation (changing a block of code in such a way that a subexpression which is common to a number of expressions is computed only once) Aho, Sethi, and Ullman 1986; Cocke 1970). More sophisticated techniques include value range optimisations, where not the value of an expression is propagated, but rather a limited range of values the expression can evaluate to (Patterson 1995) Control flow optimisations Other techniques alter the control flow in a program, or ....

Ullman, J. D. (1989a). Bottom-up beats top-down for datalog. In Proceedings of the Eighth ACM SIGACT-SIGMOD-SIGART Symposium on Principles of Database Systems (PODS '89), Philadelphia, PA, USA, pp. 140--149. ACM Press.

....the total number of prefix firings is O(jGjn 3 ) which gives the running time of the algorithm. This is a logic program presentation of the Cocke Kasimi Younger (CKY) algorithm for context free parsing. Bottom up logic programming has been widely studied in the context of deductive databases [25, 24, 17]. Bottom up logic programming has been advocated as a formalism for expressing a variety of natural language parsing algorithms [23, 5] Bottom up logic programming has also been advocated for program analysis algorithms used by compilers [24, 21] The contribution of this paper lies in the two ....

.... studied in the context of deductive databases [25, 24, 17] Bottom up logic programming has been advocated as a formalism for expressing a variety of natural language parsing algorithms [23, 5] Bottom up logic programming has also been advocated for program analysis algorithms used by compilers [24, 21]. The contribution of this paper lies in the two meta complexity theorems which provide a simple characterization of the running time of logic programs. A characterization of running time seems essential if one is to view a logic program as an algorithm. 1.2 A Framework for Static Analysis This ....

J. Ullman. Bottom-up beats top-down for datalog. In Proceedings of the Eigth ACM SIGACT-SIGMOD-SIGART Symposium on the Principles of Database Systems, pages 140--149, March 1989.

....these facts. We present one such program in Example 7.2. For such programs Opt NGBU combines the best features of Prolog evaluation and bottom up evaluation. The question of how bottom up and top down methods compare is considered important, and has been under investigation by several researchers [20, 3, 9, 16]. Our result carries the comparison of top down and bottom up methods farther than earlier results in three important ways: a) it extends the class of programs considered from safe Datalog to full logic programs, b) it compares bottom up evaluation with a sophisticated model of Prolog ....

.... it extends the class of programs considered from safe Datalog to full logic programs, b) it compares bottom up evaluation with a sophisticated model of Prolog evaluation, which incorporates last call optimization, and (c) it takes all time costs into account (earlier results with the exception of [20] ignored the cost of unification, and only compared the number of operations such as inferences performed) Since we remove all these restrictions, we believe our work represents a major advance on earlier work. 8 Where each action of Prolog evaluation takes at least unit time. 7 Discussion ....

[Article contains additional citation context not shown here]

J. D. Ullman. Bottom-up beats top-down for Datalog. In Procs. of the Eighth ACM Symp. on Principles of Database Systems, pages 140--149, March 1989.

....one of his methods (see [13] The bottom up approach is complete and terminating over programs without function symbols, but used naively computes the entire deductive closure of a program in order to answer any question, so it is not usually efficient. Various methods based on magic sets (cf. [12]) have been proposed to guide bottom up deduction, but appear to sometimes need very large numbers of auxiliary magic rules (see [2] Pereira and Warren proposed a mixed bottom up and top down deduction method (in [11] which is based on an algorithm by Earley which parses context free ....

Ullman, J.D., Bottom-Up Beats Top-Down for DATALOG , Proc. 8 th ACM SIGACT-SIGMODSIGART Symp. on Principles of Database Systems, Philadelphia, Pennsylvania, 1989, pp. 140--149

....we remove some of the restrictions of Spocus transducers. These results suggest that Spocus transducers achieve an appealing balance between expressiveness and tractability. Related work Our work is motivated by electronic commerce (e.g. see [YA96] We use the framework of relational databases [AHV95, Ull89]. Our relational transducers can be viewed as active databases with immediate triggering [WC95, PV98] although the questions we consider are quite different. The logic background is relational calculus [AHV95] for the static aspect and temporal logic [Eme91] for the temporal aspect. Since in some ....

J. D. Ullman. Bottom-up beats top-down for datalog. In Proc. ACM Symp. on Principles of Database Systems, pages 140--149, 1989.

....are usually of a special form known as a datalog program. A datalog program is a set of first order Horn clauses that do not contain function symbols. Any data Polynomial time Computation via Local Inference Rules, July 13, 2000 2 log program defines a tractable inference relation [Ullman, 1988] [Ullman, 1989]. There has been interest in generalizing the inference rules used in deductive databases beyond the special case of datalog programs. In the general case, where function symbols are allowed in Horn clause inference rules, a set of inference rules can be viewed as a Prolog program. Considerable ....

J. Ullman. Bottom-up beats top-down for datalog. In Proceedings of the Eighth ACM SIGACT-SIGMOD-SIGART Symposium on the Principles of Database Systems, pages 140--149, March 1989.

....while ours provides a topdown one. However, chaotic iteration seems difficult to generalize to the class of problems considered in this paper. 8. 3 Deductive Data Bases Fixpoint computation algorithms have been extensively studied to solve efficiently queries in deductive data bases (e.g. [5, 27, 28]) Bottom up and top down approaches are both used and have been related by means of rewriting techniques (e.g. magic sets [1, 27] Our algorithm should be related to OLDT resolution [26] However, our algorithm is inefficient in this context because elements of A and A 0 denote sets of values ....

....in this paper. 8. 3 Deductive Data Bases Fixpoint computation algorithms have been extensively studied to solve efficiently queries in deductive data bases (e.g. 5, 27, 28] Bottom up and top down approaches are both used and have been related by means of rewriting techniques (e.g. magic sets [1, 27]) Our algorithm should be related to OLDT resolution [26] However, our algorithm is inefficient in this context because elements of A and A 0 denote sets of values (facts) instead of individual values. As a consequence, each iteration redoes all computations performed by the previous ones. ....

J.D. Ullman. Bottom-up Beats Top-Down for Datalog. In Proc. ACM SIGACT-SIGMODSIGART Symp. on Principles of Database Systems, pages 140--149, 1989.

.... [79] 49] 67] 98] 121] 109] 39] 9] 22] 51] 133] 99] 135] 101] 52] 42] 43] 54] 93] 107] 108] 31] XSB, YAP, Lola, and others Optimizations [36] 5 Overview ffl Magic Sets and Related Research (e.g. Alexander Method) Formulation [102] 7] 111] [124], 20] 87] 85] 10] 70] 91] 117] 47] 60] 96] 11] 114] 104] 84] References are a little dated. Implementation and Systems [8] 25] 125] 97] 113] 130] 131] 44] 59] 63] 86] LDL, Aditi, Coral and others. Optimizations [95] 88] 105] ....

J. Ullman. Bottom-up beats top-down for datalog. In Proc. of 8th PODS, pages 140--149. ACM, 1989.

....database that is relevant to the query, the computation of the answer is (hopefully) less expensive. In fact, it has been shown that a magic rewritten datalog program evaluated by the semi naive method is no less efficient and frequently more efficient than the original program evaluated top down [20]. However, it is not always clear whether there is a benefit from the magic rewriting given the same evaluation method such as the semi naive evaluation. This is so because the efficiency of the rewriting can be defeated if the search for relevant data is too expensive. This happens because the ....

J.D. Ullman, "Bottom-up beats top-down for datalog", in proceedings of ACM Symposium on Principles of Database Systems, Philadelphia, Pennsylvania, 1989, ACM, New York, pp. 140-149

....there is an ongoing debate about the question whether deductive query evaluation methods should be based on bottom up or on top down reasoning. One of the most prominent researchers in the field has recently published a paper which already in its title announces that bottom up beats top down [Ull89b]. This claim which pretends to provide an answer to the debate mentioned is based on a very special and restricted top down method, however. Therefore it is not particularly well suited for establishing a general statement about the one or the other paradigm. 11 Moreover, in a very recent ....

J. Ullman, "Bottom-up beats top-down for Datalog", in: Proc. 8th ACM Symp. on Principles of Database Systems (PODS), Philadelphia, 1989

.... [110] 23] 27] Implementation and Systems [6] 116] 64] 42] 2] 55] 82] 84] 106] 107] 93] 7] 21] 43] 44] 96] 121] 45] 83] 95] 92] Optimizations [30] 6 Motivation Magic Sets and Related Research (e.g. Alexander Method) Formulation [87] 5] 97] [111], 18] 69] 67] 8] 59] 73] 103] 39] 49] 79] 9] 101] 89] 66] Implementation and Systems [6] 24] 112] 80] 99] 118] 119] 37] 48] 52] 68] Optimizations [78] 70] 90] 91] 88] 61] 102] 51] 50] 104] 16] ffl Bibliography is incomplete: ....

....sg(Z1,Z) p(Y,Z) 2. sg(3,Y) 2. sg(4,Y) 2. sg(3,3) 2. sg(3,Y) p(3,Z1) 2. sg(4,4) 2. sg(4,Y) p(4,Z1) In particular, this is how a semi naive evaluation would evaluate a magic rewritten same generation program. 44 Definite Programs: Tabling and Magic Relations have been often studied: [111], 18] 103] 15] 106] 36] ffl Asymptotically Equivalent: A version of magic templates under naive evaluation is asymptotically equivalent to a version of QSQR tabling [97] ffl Iteration Equivalent: A version of magic templates under semi naive evaluation is asymptotically iteration ....

J. Ullman. Bottom-up beats top-down for datalog. In Proc. of 8th PODS, pages 140--149. ACM, 1989. 93

No context found.

Ullman, J. Bottom-up beats top-down for Datalog. In Proceedings of the 8th ACM Symposium on Principles of Database Systems (1989) 140--149. 45

No context found.

J. D. Ullman. Bottom-up beats top-down for datalog. In PODS '89: Proceedings of the Eighth ACM SIGACT-SIGMOD-SIGART Symposium on Principles of Database Systems, pages 140--149. ACM Press, 1989.

No context found.

J.D. Ullman. Bottom-up beats top-down for Datalog. In Proc. 8th ACM Symp. on Principles of Database Systems, pages 140--149, 1989.

No context found.

J. D. Ullman. Bottom-up beats top-down for datalog. In Proc. 8th ACM SIGACTSIGMOD -SIGART PODS'89, pages 140--149, 1989. ACM Press. 15

No context found.

J. D. Ullman, Bottom-up beats top-down for datalog, in: S. ACM SIGACT, SIGMOD (Ed.), Proceedings of the 8th ACM SIGACT-SIGMOD-SIGART Symposium on Principles of Database Systems (PODS'89), ACM Press, Philadelphia, USA, 1989, pp. 140-149.

No context found.

J. D. Ullman. Bottom-up beats top-down for datalog. In Proc. 8th ACM SIGACTSIGMOD -SIGART PODS'89, pages 140-149, 1989. ACM Press. 15

No context found.

J.D. Ullman, Bottom-Up Beats Top-Down for Datalog. In Proc. 8th ACM SIGMOD-SIGACT Symposium on Principles of Database Systems, 1989.

No context found.

J. D. Ullman, "Bottom-up Beats Top-down for Datalog." Proceedings of the 1989 ACM Symposium on Principles of Database Systems, pp 140-149.

First 50 documents

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