M. Codish and B. Demoen. Deriving Polymorphic Type Dependencies for Logic Programs Using Multiple Incarnations of Prop. In B. Le Charlier, editor, Proc. SAS, volume 864 of LNCS, pages 281-296. Springer-Verlag, 1994.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... of [66] and in the summary optimization of [67] Another example is abstract compilation where the equations and xpoint computation are compiled (often in the same language as the one to be analyzed so that program analysis amounts to the execution of an abstract compilation of program) see e.g. [1, 4, 9, 34, 60]. Since the local analysis phases of the program parts P i , which consist in computing the xpoints lfp are delayed until the global analysis phase, which consists in computing lfp #] not much time and memory resources are saved in this computation, even though the ....

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of Prop. In B. Le Charlier (ed), Proc. 1 Int. Symp. SAS '94, LNCS 864, 281296. Springer, 1994.

....element and is ordered by boolean implication. Groundness analysis corresponds to checking the rigidity 2 of program variables w.r.t. the termsize norm 3 and abstracts a unification such as X = Y jZ] by the boolean formula X Y Z. However POS can also be used with other semi linear norms [4]. In e.g. normalised programs, it only requires to redefine the abstraction of the unifications. For example, with the listlength norm 4 , unification of X = Y jZ] is abstracted as X Z, and a formula X means that the program variable X is bound to a term with a bounded listlength, i.e. either ....

M. Codish and B. Demoen. Deriving Polymorphic Type Dependencies for Logic Programs Using Multiple Incarnations of Prop. In B. Le Charlier, editor, Proceedings of the First International Symposium on Static Analysis, number 864 in LNCS, pages 281--297, Namur, Belgium, September 1994. Springer-Verlag.

....In [12] Codish and Demoen developed so called abstract compilation to perform groundness analysis of logic programs. Applying a transformation on the original program, they obtain an abstracted program. Groundness is then derived from a least model of the abstracted program. In later work [11], they used the same technique to perform other analyses. Boulanger and coauthors [7, 6] explored the use of pre interpretations to approximate the s semantics [5] of programs and to derive properties from this approximation. In [18] the approach of Codish and Demoen is generalised and presented ....

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of prop. In Proceedings SAS'94 (Static Analysis Symposium), LNCS 864, pages 281--296. Springer-Verlag, 1994.

....work. While the extraction axioms seem somewhat non standard [1] one would hope that they do not actually enter the algorithm provided one unifies two normal abstract terms. There is an implementation of the analysis we proposed in [15] In fact, this paper relates to [15] as [4, 5] relates to [3]. This is interesting because the authors mention that an implementation using ACI unification turned out to be much faster than the implementation in [3] In particular, abstract unification is not as bad in practice as it seems by the theoretical result that it is NP complete. To compute the ....

....normal abstract terms. There is an implementation of the analysis we proposed in [15] In fact, this paper relates to [15] as [4, 5] relates to [3] This is interesting because the authors mention that an implementation using ACI unification turned out to be much faster than the implementation in [3]. In particular, abstract unification is not as bad in practice as it seems by the theoretical result that it is NP complete. To compute the abstraction of a program, in [3, 5] the user must provide information about the particular type language used in a program (see paragraph after Def. 4.1) ....

[Article contains additional citation context not shown here]

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of P rop. In B. Le Charlier, editor, Proceedings of the 1st Static Analysis Symposium, volume 864 of LNCS, pages 281--296. Springer-Verlag, 1994.

....would not be suitable for comparing properties with respect to ordering false . Then, the domain of abstract programs can be assimilated with Datalog. 7 Related works We know of several other works with similar looking objects: Codish and Demoen s, and Codish and Lagoon s type dependencies [CD94, CL96] Smaus, Hill and King s typed analysis [SHK00] and Bossi et al. s semi linear norms [BCF91] The rst work and ours combine types and Pos, but the main di erence is that the rst work adopts the descriptive view of types, whereas we 31 stick to the prescriptive view. In Codish and ....

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of Prop. In First Symp. Static Analysis, volume 864 of LNCS, pages 281-297. Springer-Verlag, 1994.

.... Abstract compilation frameworks Our analysis blends the closure framework of [14] with the abstract compilation work of [5] and [11] Abstract compilation and abstract programs have been proposed for deriving modes [5, 7] encoding regular approximations [12] inferring directional types [4] and deducing inter argument relationships [15] In addition, 5, 16] 15] and [7] present abstract compilation schemes in which analysis is realised as the bottom up and top down evaluation of constraint logic programs. None of these approaches, however, traces synchronisation behaviour. 8 ....

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of prop. In SAS'94, pages 281--297. Springer-Verlag, 1994.

....comparison between our type system and the Mycroft O Keefe one. Other related work. Rouzaud and Nguyen Phong [50] describe a type system where types are sets of non ground terms and express directionality, but these sets must be tuple distributive. Our types need not be tuple distributive. In [22], Codish and Demoen infer type dependencies for logic programs. Their techniques (abstract compilation) are quite di erent from ours and the derived dependencies express all possible input output relationships. Probably the work of Heintze and Ja ar is the one that is most closely related to ours. ....

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of prop. In B. L. Charlier, editor, Proceedings of the First International Static Analysis Symposium, volume 864 of LNCS, pages 281-296. Springer-Verlag, 1994.

....only meaningful expressions are allowed. This enables most typographical errors and inconsistencies in the knowledge representation to be detected at compile time. An increasing number of applications using typed languages are being developed. Our notion of modes is, in contrast, a descriptive one [3, 7]: Modes characterise the degree to which program variables are instantiated at certain program points. This information can be used to underpin optimisations such as the specialisation of uni cation and the removal of backtracking, and to support determinacy analysis [9] When a mode analysis is ....

....are used in an abstract compilation [4] framework: A program is abstracted by replacing each uni cation with an abstract counterpart, and then the abstract program is evaluated by applying a standard operational semantics to it. 1 We believe that this work is the natural generalisation of [3, 5] and takes the idea presented there to its limits: Our abstract domains provide the highest degree of precision that a generic domain construction should provide. Not only can this work be used directly for the mode analysis of typed logic programs, but it could be used as a basis for constructing ....

[Article contains additional citation context not shown here]

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of Prop. In SAS'94, pages 281-297. SpringerVerlag, 1994.

....the domain Pos for groundness analysis. In [16] the authors define generic type domains with dependencies but, again, this is not a real generalisation of Pos , though it can be used for abstract compilation. The only paper which generalises the definition of the domain Pos to generic types is [3]. However, in this paper it is assumed that the usual properties of Pos hold for all the type domains. For instance, logical conjunction between formulas is used as conjunction operator and Schroder elimination as cylindrification operator. It is not obvious at all that these operators, which are ....

....every positive type domain, we have generalised the result contained in [6] for the case of groundness. Therefore, we can use many incarnations of positive type analyses, combining them with a reduced product operation and using the operations and diagonal elements of Definition 7 as done in [3], where, however, the authors did not provide any justification of the correctness of this approach. The following example shows that we can apply our theory to more complicated cases than those considered above, dealing with many types, dependencies among different types and polymorphism. ....

M. Codish and B. Demoen. Deriving Polymorphic Type Dependencies for Logic Programs Using Multiple Incarnations of Prop. In Proc. of the first International Symposium on Static Analysis, volume 864 of Lecture Notes in Computer Science, pages 281--296. Springer-Verlag, 1994.

....types, we have maybe given a new view of the results in [1] Other related work. Rouzaud and Nguyen Phong [28] describe a type system where types are sets of non ground terms and express directionality, but these sets must be tuple distributive. Our types need not be tuple distributive. In [14], Codish and Demoen infer type dependencies for logic programs. Their techniques (abstract compilation) are quite different from ours and the derived dependencies express all possible input output relationships. Probably the work of Heintze and Jaffar is the one that is most closely related to ....

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of prop. In B. L. Charlier, editor, Proceedings of the First International Static Analysis Symposium, Lecture Notes in Computer Science 864, pages 281--296. Springer Verlag, 1994.

....in the type language. This is a standard solution, used for instance in [3,13,19,6] The same solution is used in the framework of regular approximations of the success set in [11] However, the use of type variables does not allow one to express all type dependencies between argument positions. [5] is the only example of an analysis which explicitly expresses type dependencies. However, the underlying type language does not contain variables. Hence type dependencies arising from polymorphism can not be handled without an in nite set of dependencies. Moreover, 5] does not consider ....

....between argument positions. 5] is the only example of an analysis which explicitly expresses type dependencies. However, the underlying type language does not contain variables. Hence type dependencies arising from polymorphism can not be handled without an in nite set of dependencies. Moreover, [5] does not consider dependencies between di erent types, since the analysis is performed separately for every type. Finally, there is no approach which is able to express negative information on types. One of the contributions of this paper is to give a general technique for explicitly expressing ....

[Article contains additional citation context not shown here]

M. Codish and B. Demoen. Deriving Polymorphic Type Dependencies for Logic Programs Using Multiple Incarnations of Prop. In Proc. of the rst International Symposium on Static Analysis, volume 864 of Lecture Notes in Computer Science, pages 281-296. Springer-Verlag, 1994.

....Prolog with type systems in the forms of type checking systems [4, 18, 39] and type analysis systems. Type analysis systems infer types from the text of a program. Some type analysis systems infer a type for the program and the type is meant to denote a superset of the success set of the program [1, 2, 10, 16, 19, 20, 37, 42, 48, 49, 50]. Other type analysis systems infer types for procedure calls and successes [6, 22, 24, 27, 28] There are three main approaches to inferring types [20] The first approach obtains from a program P a system of formulae which describes set theoretic relationships between the atoms appearing in P. A ....

....The type description only describes part of the success set of the program. The atoms that are not well typed are not described by the type description. Neither are those atoms whose successful SLD resolutions with the program all select at least one atom that is not well typed. Codish and Demoen [10] apply abstract compilation to infer type dependencies of logic programs by associating each type with an incarnation of the abstract domain Prop [32] Abstract compilation [21] is an alternative to abstract interpretation [13] In abstract compilation, a program is analysed by applying a concrete ....

[Article contains additional citation context not shown here]

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of Prop. In Proceedings of the First Symposium on Static Analysis, volume 864 of Lecture Notes in Computer Science, pages 281--297. SpringerVerlag, September 1994.

....information does not allow to distinguish between e.g. bounded lists such as [1; X; 3] and open ended lists such as [1; 2jZ] because the open end can be situated at an arbitrary depth. Making such distinction requires to consider type information. The domain P os has been adapted to do so in [6] where each type was associated with an incarnation of P os. However, that analysis was for untyped programs and each incarnation was developed in a somewhat ad hoc fashion. The model based analysis of [14] is also related. Here, the models of the program, based on different pre interpretations ....

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of prop. In Proc. SAS'94, volume 864 of LNCS, pages 281--296. Springer, 1994.

....consequence [18, 7, 19] The abstract operations that arise in analyses are: Boolean conjunction, as a greatest lower bound, Boolean disjunction, as a least upper bound, and existential quanti cation, for projection. There is a substantial body of literature on this type of analysis, for example, [1, 4, 8]. In this section we describe two problems with the Pos domain. The multiplicity problem: The rst problem arises in groundness analysis of logic programs. For groundness analysis, a Boolean function describes a substitution if any set of variables that might become ground by further ....

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of Prop. In B. Le Charlier, editor, Static Analysis: Proc. First Int. Symp., volume 864 of Lecture Notes in Computer Science, pages 281-296. Springer, 1994.

....information does not allow to distinguish between e.g. bounded lists such as [1; X; 3] and open ended lists such as [1; 2jZ] because the open end can be situated at an arbitrary depth. Making such distinction requires to consider type information. The domain P os has been adapted to do so in [6] where each type was associated with an incarnation of P os. However, that analysis was for untyped programs and each incarnation was developed in somewhat ad hoc fashion. The model based analysis of [14] is also related. Here, the models of the program, based on different pre interpretations ....

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of prop. In Proc. SAS'94, volume 864 of LNCS, pages 281--296. Springer, 1994.

....transformations. Consequently, it is straightforward to provide similar functionalities using other top down frameworks based on [2] such as for example those described in [19] GAIA) and in [17] AMAI) The experiments described in this paper are based on three well known abstract domains: Prop [9, 11, 18], Sharing [16, 23] and ASub [24] For sharing analysis, data descriptions are represented as lists of lists of variables which appear as comments in the text of the program. The information describes properties of possible substitutions when execution reaches different points in the clause. The ....

....uses a ground representation of the domain relations. This results in large domain relations for predicates with a large arity and shows up in the table as a relative slow down for benchmarks with such predicates. We expect that a non ground representation of the domain relations (as in [9] or [25] would reduce the influence of large arities and on the whole benefit the implementation. However, with a rudimentary implementation of the gif domain, we arrive at a speed up of 2 on average over AMAI. The precision of gif in deriving groundness is exactly as for Share Free, as ....

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of Prop. In B. Le Charlier, ed., Proceedings of the first International Static Analysis Symposium, SAS'94, pp. 281--296. LNCS 864, Springer-Verlag, Sept. 1994.

....that, as shown recently in [30] the Pos domain can be interpreted as an implicational domain and can thus be used to capture any directional information for a downward closed property. In particular, it can capture dependencies between variables or argument positions, such as type dependencies [16], niteness analysis [5] and quite recently binding time analysis [11] We note that the last of these analyses was implemented using the approach described in this article. 2.3 Example 1: Goal dependent analysis with Pos Figure 5 illustrates a Prolog program which rotates the elements of a ....

M. Codish and B. Demoen. Deriving Polymorphic Type Dependencies for Logic Programs Using Multiple Incarnations of Prop. In B. Le Charlier, editor, Proceedings of the First International Symposium on Static Analysis, number 864 in LNCS, pages 281-297, Namur, Belgium, Sept. 1994. Springer-Verlag.

....Computer Science, K.U. Leuven, Belgium. bimbart cs.kuleuven.ac. be 1 Other analyses involving the domain Prop include various applications for suspension analysis described in [11] and simple type analyses described in [13] A non trivial application for polymorphic type analysis is described in [9]. In this paper we focus on groundness analysis as captured by abstract interpretation using the domain Prop. We apply an approach in which a program is analyzed by applying a concrete semantics to an abstraction of the program itself. This technique is referred to as abstract compilation in [24] ....

....that these implementations are more general than that of our groundness dependency analyzer. However, our approach is extremely simple to implement and is not restricted to the analysis presented in this paper. Recent results illustrate its application to a non trivial polymorphic type analysis [9]. The rest of this paper is organized as follows. The next section opens with an intuitive example which illustrates the basic idea of our approach. Section 3 sets the notational conventions used throughout and describes several background issues. Section 4 formalizes our groundness analysis and ....

[Article contains additional citation context not shown here]

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of prop. In Proceedings of the First Symposium on Static Analysis, volume 864 of Lecture Notes in Computer Science, pages 281--297. Springer-Verlag, September 1994.

....terms. The combination of program abstraction and ACI enhanced concrete evaluation results in a promising technique which is both theoretically clean and easy to implement using general purpose tools. The idea of obtaining static analyses by program abstraction is derived from ideas presented in [28, 19, 26, 15, 14] and has been applied in a variety of contexts, e.g. 44, 23, 7, 25, 24] Unique to our approach is the combination of a simple syntax directed program abstraction and a standard equality theory for type expressions. We pay special attention to the role of variables in abstract programs and in ....

....ACI unification. A bottom up polymorphic analysis based on the abstraction of the s semantics and several examples of its use are described in Section 5. Section 6 summarizes and provides our conclusion. This paper is an extended version of [17] which was inspired by the earlier work presented in [14]. 2 Preliminaries Syntax: In the following we assume a familiarity with the standard definitions and notation for logic programs as described in [35, 2] For any set of function symbols Sigma and variables V, we let T ( Sigma; V) denote the set of terms constructed using symbols from Sigma and ....

[Article contains additional citation context not shown here]

M. Codish and B. Demoen. Deriving Polymorphic Type Dependencies for Logic Programs using Multiple Incarnation of Prop. In Le Charlier [33], pages 281--296.

....de Waal, 1994; Janssens and Bruynooghe, 1992] However, in these frameworks the type definitions themselves are derived from the program clauses. We assume given type definitions and we check and or infer predicate typings using only these type declarations: in this respect, our work is closer to [Codish and Demoen, 1994] which also assumes the set of possible types to be given to the analysis. The closest in spirit to our type checking inference approach is the Mercury system [Somogyi et al. 1996] the Mercury compiler contains a type checker that also infers types for local predicates missing a pred declaration ....

Codish, M. and Demoen, B. (1994). Deriving polymorphic type dependencies for logic programs using multiple incarnations of prop. In Proc. of Static Analysis Symposium, number 864 in LNCS, pages 281--297. Springer-Verlag.

No context found.

M. Codish and B. Demoen. Deriving Polymorphic Type Dependencies for Logic Programs Using Multiple Incarnations of Prop. In B. Le Charlier, editor, Proc. SAS, volume 864 of LNCS, pages 281-296. Springer-Verlag, 1994.

No context found.

M. Codish and B. Demoen. Deriving Polymorphic Type Dependencies for Logic Programs Using Multiple Incarnations of Prop. In B. Le Charlier, editor, Proc. SAS, volume 864 of Lecture Notes in Computer Science, pages 281--296. Springer-Verlag, 1994.

No context found.

M. Codish and B. Demoen. Deriving Polymorphic Type Dependencies for Logic Programs Using Multiple Incarnations of Prop. In Proc. of the rst International Symposium on Static Analysis, volume 864 of Lecture Notes in Computer Science, pages 281-296. Springer-Verlag, 1994.

No context found.

M. Codish and B. Demoen. Deriving Polymorphic Type Dependencies for Logic Programs Using Multiple Incarnations of Prop. In Proc. of the rst International Symposium on Static Analysis, volume 864 of Lecture Notes in Computer Science, pages 281-296. Springer-Verlag, 1994.

No context found.

M. Codish and B. Demoen. Deriving polymorphic type dependencies for logic programs using multiple incarnations of prop. In SAS'94, pages 281--297. Springer-Verlag, 1994.

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