U. Kastens. Ordered attribute grammars. Acta Informatica, 13(3):229--256, 1980.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....From above features of MAGE2 system, we can develop easily complicated object oriented database systems which manage structured data eOEciently. 7. Future Works From the viewpoint of an execution speed of attribute evaluation, it is better to implement more eOEcientevaluator such ordered AGs [14] one. We plan to implement another evaluator using improved ordered AGs class OAG introduced in [15] ....

Uwe Kastens. Ordered attribute grammars. Acta Informatica, 13(3):229256, 1980.

....for context sensitive program analysis already existed in 1988 [13, 45] although we were not aware of them at the time. Our work was more influenced by Kastens s algorithm for building a collection of cooperating finite state machines for evaluating derivation trees of attribute grammars [24] (see below) The two other papers in the session at PLDI 88 in which our paper was presented [8, 12] actually foreshadowed a major direction that our work was to take subsequently: Callahan presented a context sensitive algorithm for a certain kind of dataflow analysis problem [8] Cooper and ....

.... the transitive dependences captured by the characteristic graphs of an attribute grammar s nonterminals; the algorithm for identifying summary edges was inspired by the first two passes of the algorithm for creating the TDS graphs that are used to define the class of ordered attribute grammars [24]. This yields a slicing algorithm that is quartic (i.e. O(N 4 ) in a certain size parameter of the problem. Given what we have learned in the last fifteen years, if we were to write the paper today, we would describe the work somewhat differently: in addition to describing the problem as a ....

U. Kastens. Ordered attribute grammars. Acta Inf., 13(3):229--256, 1980.

....3. definition of interfaces of sub traversals 4. dealing with interdependencies between sub traversals 5. generation of Gofer code The first step determines how arguments and results of functions may depend upon each other, and is based on well known attribute grammar analysis techniques [Kas80]. This step builds two kinds of dependencies: those among the arguments and results of one function and those among all arguments and results of all calls occurring in a function body. Once the dependencies are known we can see whether an argument in a call depends on a result in the same call. ....

Uwe Kastens. Ordered attribute grammars. Acta Informatica, 13:229--256, 1980.

....grammar [18] semantic analysis is achieved using attribute grammars [21] and code generation is often specified using a transformational grammar [14] A third compiler construction technique is tree transformation. Highly efficient tree manipulation is provided by ordered attribute grammars [20], by the TXL system [9] and the Gentle system [30] Tree manipulations provide a convenient and efficient way to implement a wide range of language based tasks such as translation, design recovery from source code, and automatic programming [10] 4. Case Study: Math Recognition As a case study, ....

U. Kastens, "Ordered Attribute Grammars," Acta Informatica, 13, 1980, 229-256.

....a symbol table by means of grammar rules with attributes for lookup values, an incremental implementation of the lookup function is obtained. An incremental implementation for ordered higher order attribute grammars is based on caching the results of visit functions for evaluating attributes [Kas80] rather than on storing attribute values in a tree. A visit function takes as first parameter a tree and part of the inherited attributes of the root of that tree. It returns a subset of synthesized attributes. An immediate consequence of this is a memoization of attribute values at physically ....

U. Kastens. Ordered attribute grammars. Acta Informatica, 13:229--256, 1980.

....to generate the visit procedures. Unfortunately, however, there will in general be much freedom in converting the partial orders to the required total orders. Our implementation is thus based on the smaller, but polynomially decidable class of ordered attribute grammars (OAGs) defined by Kastens [47]. 1 The OAG class is defined by a greedy strategy for computing the total orders, placing each attribute as early as possible in the total order for its phylum, subject to the requirement that it follow all of the attributes that precede it in the corresponding partial order. If more than one ....

....except possibly for I 1 and SN , and, in any tree, the instances of S j depend only on those of I j [ I i [ S i ; i j and the instances of I j depend only on those of I i [ S i ; i j. In the case of OAGs, there exists a standard and efficient method to compute these sequences, due to Kastens [47]. The sequence of pairs, or partition, defines a sequence of visits to be made to each instance of phylum X during the traversal and attribution of the AST. Upon entry to visit j, the instances of the I j will be available, as well as all instances computed in previous visits, thus the attributes ....

[Article contains additional citation context not shown here]

Uwe Kastens. Ordered attribute grammars. Acta Informatica, 13:229--256, 1980.

....nal states, if it has no dependency for the occurrence S:result . This process is iterated until there is no new dependency graph is generated. A graph based analysis of attribute dependency similar to the present 18 Nakano Nishimura scheme can be found in Kastens visit sequence construction [12], in which the attribute evaluation order is derived by an exhaustive enumeration of all possible attribute dependencies. The crucial dioeerence between his and ours is that his scheme preserves the original production rules while ours pastes together production rules to create a new one. The ....

U. Kastens. Ordered attribute grammars. Acta Informatica, 13(3):229256, 1980.

....grammar is not well defined. Testing if a grammar is well defined, however, requires exponential time [2] A sufficient condition for being well defined can be checked in polynomial time. This test defines the set of ordered attribute grammars as being a subset of the well defined grammars [5]. However, there is no constructive method to design such grammars. Hence, designing an or 10 Sabine Glesner and Wolf Zimmermann dered attribute grammar remains a difficult problem. For another class of attribute grammars it is required that all attributes can be evaluated during a constant ....

U. Kastens. Ordered Attribute Grammars. Acta Informatica, 13(3):229--256, 1980.

....with production. In the learning process the grammar, background semantic functions and examples can be used. In the current approach S attributed an L attributed grammars with simple rules can be learned. However, the background knowledge may contain more complex attribute grammars, as in OAG [8]. In this paper we present an application of the AGLEARN method to the part of speech (POS) tagging of Hungarian sentences. The disambiguation of possible tags is a non trivial task because many words may have different meanings e.g the word m ult 2 can be annotated as a verb, noun or adjective ....

Kastens, U. 1980. Ordered Attribute Grammars. Acta Informatica 13 (1980), 229--256.

....From above features of MAGE2 system, we can develop easily complicated object oriented database systems which manage structured data eOEciently. 7. Future Works From the viewpoint of an execution speed of attribute evaluation, it is better to implement more eOEcient evaluator such ordered AGs [14] one. We plan to implement another evaluator using improved ordered AGs class OAG introduced in [15] ....

Uwe Kastens. Ordered attribute grammars. Acta Informatica, 13(3):229256, 1980.

....variant, are a suitable formalism to specify such language based environments. Furthermore, attribute grammar are executable, i.e. eOEcient implementations are automatically obtained. In this paper we introduce a new strict, purely functional implementation for ordered attribute grammars [Kas80]. A strict model of attribute evaluation is attractive for twomainreasons: rstly, because we obtain very eOEcient implementations in terms of memory and time consumption. Secondly, because a rather eOEcient and simple incremental attribute evaluator can be derived from an attribute grammar: ....

....moved from the syntactic domain to the semantic domain. These characteristics make higher order attribute grammars particularly suitable to model interactive language based environments [TC90,Pen94,KS98,Sar99] It is known that the incremental attribute evaluator for ordered attribute grammars [Kas80,RT89] can be trivially adapted for the incremental evaluation of higher order attribute grammars. The adapted evaluator, however, decorates every instance of a higher order attribute separately [TC90] Note that in suchtraditional evaluators the usual representation for an attributed tree is an ....

[Article contains additional citation context not shown here]

Uwe Kastens. Ordered attribute grammars. Acta Informatica, 13:229256, 1980.

....Application Application Application Fig. 2. Components of the ARC system. Along with the component types are definitions for the various attributes and evaluation rules. Attributes are defined as a particular data type such as integer, real, etc. and attached to the different component types [5, 6]. Evaluations rules define how to calculate various performance metrics and aspects. These rules are composed of expressions, functions, and programming statements that may involve the various attributes of the different design components. It is important to note here that a performance model ....

U. Kastens. Ordered Attribute Grammars. In Acta Informatica, volume 13, pages 229--256, 1980.

....evaluator. Hence, more specialized evaluations methods are necessary to obtain fast semantic evaluation. In the literature, several methods have been proposed for AGs. The best known are the one sweep and multi sweep grammars, described in [8] and the ordered attribute grammars proposed in [13]. We think that these methods can be adapted to the evaluation of SRDs; in particular, the adaptations should concern the way dependencies determined by different branches of conditional rules affect the evaluation strategy and the computation of the evaluation visit sequences performed by the ....

U. Kastens. Ordered attribute grammars. Acta Informatica, 13:229--256, 1983.

....The selection algorithm is presented in Section 4. In Section 5, we discuss the correctness and the finest partitioning properties of the algorithm. Section 6 concludes this paper. 2. Notations In this section, we define the notations used in this paper. Basically, we adopt Kastens s notations [4]. An attribute grammar is built from a context free grammar (N; T ; P; S) where N is a finite set of nonterminals, T is a finite set of terminals, S is a distinguished nonterminal, called the start symbol, and P is a set of productions of the form: X ff, where X is a nonterminal and ff is a ....

Kastens, U., "Ordered attribute grammars," Acta Informatica, Vol. 13, pp. 229-256, 1980.

....graph GDS that provides a new approximation algorithm, which is different from the existing GAG and Eli Liga systems. We also show that we obtained good results with our experimental implementation. 1. Introduction Ordered attribute grammars (OAGs for short) introduced by Kastens in 1980 [Kas80] are a useful class of attribute grammars [Knu68] Knu71] AGs for short) since: ffl OAGs are large enough to include many practical AGs. ffl The problem whether a given AG is ordered is decided in polynomial time for the size of the AG. ffl OAGs can automatically generate efficient attribute ....

....NP complete. In such a case, the grammar is defined as l ordered AGs [EF82] In the manual of the Synthesizer Generator TM (SG for short) Gra96] circularities due to the approximation used are called type 3 circularities. In this paper, we follow the SG s terminology. Kastens stated in [Kas80] that type 3 circularities usually do not occur in practical applications such as compilers. This may be true for defining programming languages, but as we show later, it is not always true when defining programming environments using the SG. Actually, from our experience in implementing the MAGE2 ....

[Article contains additional citation context not shown here]

U. Kastens. Ordered attribute grammars. Acta Informatica, 13(3):229--256, 1980.

....the generic attribute grammars have to guarantee that all the attribute instances of a generic abstract syntax tree are eoeectively computable. That is, an order to compute such attribute instances must exist. Several methods exist that nd such an order for inormalj attribute grammars [Kas80, Alb91, Pen94]. Most of the algorithms that analyse attribute grammars for attribute dependencies can be used to handle generic attribute grammars as well. The key idea is to provide those algorithms with the dependencies between the inherited and the synthesized attributes of the generic symbols. Such ....

....the GenAg since the attribute equations dening the synthesized attributes of the generic symbols are not included in the GenAg. In this paper we will present a strict and purely functional implementation for generic attribute grammars. This implementation is based on the visit sequence paradigm [Kas80] and is restricted to the class of partitionable attribute grammars [Alb91] A straightforward strategy to approximate the dependencies of the generic symbols is to look at 190 Generic Attribute Grammars the evaluators of these symbols as functions from inherited to synthesized attributes. That ....

Uwe Kastens. Ordered attribute grammars. Acta Informatica, 13:229256, 1980.

....Attribute evaluators for attributed graph grammars are generated by analyzing the total dependency graph and determining an evaluation sequence that satisfies all dependencies. One approach is to completely determine an evaluation sequence from the set of attribute evaluation rules in the grammar ([31, 32, 33, 34, 35]) Such an evaluation sequence will, for a majority of graphs in the language, consistently bind values to attribute instances in the attribute instance graph. We call this static or compile time evaluator generation. Static evaluator generation has its advantages and disadvantages. Its advantage ....

U. Kastens, "Ordered attribute grammars," Acta Informatica, vol. 13, no. 3, pp. 229--256, 1980.

....production. In the learning process the grammar, the background semantic functions and the examples can be used. In the current approach S attributed an L attributed grammars with simple rules can be learnt. However, the background knowledge may contain more complex attribute grammars, as in OAG [Kastens 1980] for example. The main result of this paper is that by using the ILP concept a formal definition of the learning problem of semantic functions of attribute grammars can be provided. In addition a method for learning simple S attributed and L attributed semantic functions can also be completed. We ....

Kastens, U. 1980. Ordered Attribute Grammars. Acta Informatica 13 (1980), 229--256.

....grammar consisting of one nonterminal and one production must be constructed for each procedure. The attributes in the linkage grammar correspond to the parameters in the procedures. This attribute grammar is the input to an algorithm that is a slight modification of the algorithm described in [12]. The algorithm requires the construction of an auxiliary graph which expresses the dependences among the attributes of a production s nonterminal occurrences. Presently, we have implemented the algorithms that are proposed here as part of our Ghinsu tool. This tool, which is briefly discussed in ....

....the return statement at line [16] is not executed. Therefore, statement [18] is control dependent on statement [14] the return statement s control predicate) In addition, the definition of variable c at line [15] cannot reach 8 [ 1] int main( 11] int func(int a, int b, int c) 2] [12]. 3] int sum; 13] c = c 1; 4] int x; 14] if (c = 4) 5] int y; 15] c = 4 c; 6] x=4; 16] return c; 7] y=6; 17] 8] sum=func(x,y,4) 18] a = a c; 9] sum=sum 1; 19] return a; 10] 20] Table 2. An example program. statement [18] because, the return ....

[Article contains additional citation context not shown here]

Kastens, U. "Ordered Attribute Grammars". Acta Inf. 13,3, 1980.

....grammar is not well defined. Testing if a grammar is well defined, however, requires exponential time [5] A sufficient condition for being well defined can be checked in polynomial time. This test defines the set of ordered attribute grammars as being a subset of the welldefined grammars [6]. However, there is no constructive method to design such grammars. Hence, designing an ordered attribute grammar remains a difficult problem. For another class of attribute grammars it is required that all attributes can be evaluated during a single depth first, left to right traversal of the ....

U. Kastens. Ordered Attribute Grammars. Acta Informatica, 13(3):229--256, 1980.

....grammar is not well defined. Testing if a grammar is well defined, however, requires exponential time [Jaz81] A sufficient condition for being well defined can be checked in polynomial time. This test defines the set of ordered attribute grammars as being a subset of the well defined grammars [Kas80]. However, there is no constructive method to design such grammars. Hence, designing an ordered attribute grammar remains a difficult problem. For another class of attribute grammars it is required that all attributes can be evaluated during a constant number of DFS of the AST. These are the ....

U. Kastens. Ordered Attribute Grammars. Acta Informatica, 13(3):229--256, 1980.

No context found.

U. Kastens. Ordered attribute grammars. Acta Informatica, 13(3):229--256, 1980.

No context found.

U. Kastens. Ordered Attribute Grammars. In Acta Informatica,volume 13, pages 229#256, 1980.

No context found.

Kastens80. Kastens, U., Ordered attribute grammars, Acta Informatica 13(3) pp. 229-256 (1980).

No context found.

Kastens U.: Ordered Attribute Grammars. Acta Inf. 13,3 (1980) 229-256.

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