Glauert J.R., Kennaway J.R., Sleep M.R.: Dactl: An Experimental Graph Rewriting Language, in: [8], 378-395  Home/Search   Document Details and Download   Summary   Related Articles   Check

....visual languages and generating associated editors [4, 31] Besides specific applications of graph transformation, several programming languages have been developed that are based on graph transformation rules. Examples of such languages are Progres [43] Agg [22] Gamma [3] Grrr [39] and Dactl [24]. Programming tasks of a realistic size require large numbers of rules, making it imperative to develop systems from small components that are easy to comprehend. Recently, structuring concepts for systems of graph transformation rules have received some attention. A module concept for graph ....

R. W. Glauert, J. Kennaway, and M. Ronan Sleep, Dactl: An experimental graph rewriting language, in "Graph Grammars and Their Application to Computer Science, Selected Papers," Lecture Notes in Computer Science, Vol. 532, pp. 378--395, Springer-Verlag, New York/Berlin, 1991.

....the behaviour of an Actor grammar using graph transformations. Barthelmann and Schied also use graph grammars as the semantics of a parallel language, DHOP [7] and Glauert, Kennaway and Sleep have developed Dactl as a graph grammar based common target language for a number of other languages [42]. All of these approaches, while interesting and designed to deal with parallelism, take a relatively fine grained approach to parallelism, essentially rendering the partitioning problem moot. 48 10 Conclusions Our grammars cannot generate all graphs. A formalism which generates all graphs ....

J.R.W. Glauert, J.R. Kennaway, and M.R. Sleep. Dactl: An experimental graph rewriting language. In H. Ehrig, H.-J. Kreowski, and G. Rozenberg, editors, Proceedings of the 4th International Workshop on Graph Grammars and Their Application to Computer Science, number 532 in Lecture Notes in Computer Science, pages 378--395, Bremen, Germany, March 5--9, 1990. Springer-Verlag.

....combination, along with features such as serial rewriting and serial trigger initiation make Grrr a potentially useful system for inherently visual, but complex tasks such as graph drawing and algorithm animation. Previous graph rewriting languages include GOOD [15] Progres [18] Dactl MONSTR [7,1] and D grammar programming [11] each of which has a unique interpretation of programming with graph rewrites. These graph rewriting languages vary in several important aspects: the type of host graph that is to be rewritten may be any graph, or it may be restricted by disallowing duplicate nodes ....

Glauert J.R., Kennaway J.R. and Sleep M.R.: Dactl: An Experimental Graph Rewriting Language. Proceedings 4 th International Workshop on Graph Grammars and their application to Computer Science. LNCS 532. 378-395.

....or offer at least pattern matching constructs, and which are built upon a more or less graph like data model. Therefore, all so called graph rewriting systems are omitted that are merely term rewriting systems which allow sharing of subterms and which have a text oriented representation [20, 24]. Furthermore, we decided to disregard all those approaches which aim at the recognition (parsing) of diagram languages only and have a much more limited application domain [21, 27, 31] Using these criteria, a surveyable number of visual languages remain for inspection. Compared with PROGRES ....

Glauert J.R., Kennaway J.R., Sleep M.R.: Dactl: An Experimental Graph Rewriting Language, in: [15], 378-395

....programming facilities with a graph based data representation in a seamless manner. We also discuss some types of query that show this paradigm has benefits that are not available in most textual query languages. Graph rewriting derives from the mathematical area of graph grammars [6] Dactyl [7], is a textual programming language that uses graph rewriting. Derived from graph grammars are grammars, which have been used as a basis for a powerful visual graph rewriting language by Loyall and Kaplan [8] They give examples based both on state oriented and data oriented applications. The ....

....overlapping ways. A conflict could be avoided if only one subgraph was rewritten. PROGRES [9] takes this approach, arbitrarily selecting the rewritten subgraph out of the candidate matches. This unfortunately causes nondeterministic behaviour. A variation on this idea is that of Loyall and Kaplan [7], who rewrite the largest subset of non conflict causing subgraphs, this would also result in non determinism if implemented in Spider. In the example of figures 3.1 and 3.2 both the last two methods would leave a single node in the graph, but it is not defined which one it would be. Spider ....

Glauert J.R., Kennaway J.R. & Sleep M.R. (1991) Dactl: An Experimental Graph Rewriting Language. Proceedings 4th International Workshop on Graph Grammars and Their Application to Computer Science. LNCS 532. Springer-Verlag. pp. 378-395.

....rule, the name x 2 , and not the expression (2; 3) will be substituted for each occurrence of x, leading to the term: fx 1 = G(x 2 ; x 2 ) x 2 = x 3 ; x 4 ) x 3 = 2; x 4 = 3; in x 1 g : Thus, term graphs are simply a set of recursion equations. Similar notations appear in the literature [16, 18]. e.g. fx : F(x; y) y : G(x)g in the language DACTL [18] However, we insist on an equational notation, not just for the sake of style, but because we want to express term graph operations in terms of equational transformations [8] Definition 2.1. Term graph) Let Sigma be a first order ....

....substituted for each occurrence of x, leading to the term: fx 1 = G(x 2 ; x 2 ) x 2 = x 3 ; x 4 ) x 3 = 2; x 4 = 3; in x 1 g : Thus, term graphs are simply a set of recursion equations. Similar notations appear in the literature [16, 18] e.g. fx : F(x; y) y : G(x)g in the language DACTL [18]. However, we insist on an equational notation, not just for the sake of style, but because we want to express term graph operations in terms of equational transformations [8] Definition 2.1. Term graph) Let Sigma be a first order signature. A term graph defined over Sigma is defined ....

J. R. W. Glauert, J. R. Kennaway, and M. R. Sleep. Dactl: An experimental graph rewriting language. In Proc. 4th International Workshop on Graph Grammars and their Application to Computer Science, Bremen, Germany, Springer-Verlag LNCS 532, pages 378--395, 1990.

....restructuring operations, but its data model is just a binary tree with colored nodes. Therefore, we have to emphasize that PROGRES allows for the definition of arbitrary directed (attributed) graphs, and is not restricted to the case of attributed trees or directed acyclic graphs (as for instance [13, 17]) Directed, attributed graphs consist of typed nodes with attributes and directed typed edges between pairs of nodes. Each edge may be traversed in both directions (from source to target and vice versa) i.e. is equivalent to a pair of pointers of Pascal like programming languages. Furthermore, ....

....which are either rule oriented or offer at least pattern matching constructs, and which are built upon a more or less graph like data model. Therefore, all so called graph rewriting systems are omitted which are more or less just term rewriting systems with sharing of subterms (as for instance [13, 17]) Using these criteria, a surveyable number of visual languages remains for inspection. Compared with PROGRES their deficiencies come from the following sources: in many cases these languages focus on manipulation of data structures only and data definition sublanguages are not provided [3, 15, ....

Glauert J.R., Kennaway J.R., Sleep M.R.: Dactl: An Experimental Graph Rewriting Language, in: [9], 378-395

....as certain kinds of pushout. See [Ken87, HP88, HKP88, Ken91] Nevertheless, none of these constructions successfully describe the whole of the operational models of [BvEG 87] where term graph rewriting was introduced, or of its generalization in the language DACTL itself [GKSS88, GHK 88, GKS91, Ken90] The main stumbling blocks for all of these attempts have been examples such as the I combinator root:I[a] a when applied to a circular instance of itself x:I[x] None of the hitherto proposed categorical formulations of TGR adequately capture the DACTL version of the rewrite (which is, ....

J.R.W. Glauert, J.R. Kennaway, M.R. Sleep. DACTL: An experimental graph rewriting language. Graph Grammars and their Application to Computer Science, LNCS 532, pp. 378-395, Springer-Verlag, 1991.

....derivations . Parisi Presicce [1993] showed that this approach is strictly more powerful then the single pushout approach. One of the motivations [Kennaway, 1987] for developing the single pushout had already come from the context of algebraic modelling of the graph rewriting language DACTL [Glauert et al. 1990]. fiberr.tex; 9 05 1997; 19:27; no v. p.3 4 Wolfram Kahl L R A H B H As a further development of this approach, Banach [1993a; 1994] undertook a modelling through opfibrations the bold arrows are mappings from the rule sides to elements of their index categories. The advantage ....

.... as we indicated above (the weak opfibrations presented by the same author in another paper [Banach, 1993b] which might offer a solution in case of lacking composability, are therefore not needed) This proof of composability depends on garbage retention which also is a feature of DACTL [Glauert et al. 1990], the motivating system for Banach s approach. In the context of rewriting term graphs with bound variables [Kahl, 1996] garbage retention is not possible and composability of single heterogeneous pushouts therefore much harder to achieve. 6. Application of the Grothendiek Construction The first ....

J. R. W. Glauert, J. R. Kennaway, and M. R. Sleep. Dactl: An experimental graph rewriting language. In Ehrig et al. [1990c], pages 378--395.

....only few languages and tools available which are based on graph transformation. One of the very early visual programming languages based on graph transformation has been Plan2d [DFS75] Graph transformation based tools are Pagg [Got88] Graph Ed [Him91] Plexus [Wan91] Progres [Sch91b] Dactl [GKS91] and Agg [LB93] Most of them contain a graph editor for drawing graphs, and an interpreter for graph transformation. The editors usually support graphical representations, mouse menue driven interfaces, and different views on graphs, like on different sections of a graph, different layouts or ....

....an interpreter, a database and a graphical interface. In particular, the latter will play an important role in that it allows to use Grace as a visual specification and programming language. Since there are already some implementations of graph transformation systems (see [Sch90, Sch91a, Sch91b, GKS91, Him91, LB93, TB94] we will reuse their components as far as possible to build up the Grace environment. 3. Grace offers the choice among various graph transformation approaches. Furthermore, we will investigate the conditions under which approaches can be combined: As a final goal, we want to ....

John R. W. Glauert, J. Richard Kennaway, and M. Ronan Sleep. Dactl: An experimental graph rewriting language. In Ehrig et al. [EKR91], 378-- 395.

.... of programming languages has led to the theory of graph grammars (see [CER79, ENR83, ENRR87, EKR91] for a survey) A wide spectrum of graph grammar approaches exists within this theory and some of them are implemented like for example PROGRES ( Sch90, Sch91a, Sch91b] Graph Ed [Him91] Dactl [GKS91], and AGG [LB93, TB93] The new graph and rule based language GRACE is intended to be a specification and programming language that supports integration of different approaches of graph grammars as well as modularity (see also [EE93] Therefore it is supposed to be a candidate for making the ....

J.R.W. Glaubert, J.R. Kennaway, M.R. Sleep. Dactl: An experimental graph rewriting language. In H. Ehrig, H.-J. Kreowski, G. Rozenberg, eds., Graph Grammars and Their Application to Computer Science, Lecture Notes in Computer Science 532, 378--395, 1991.

.... object oriented languages ( 2] Furthermore, it has been shown to be capable of expressing the functionality of computational models such as Linear Logic and p calculus (see bibliography in [2] In addition, the implementation of TGRS themselves, by means of associated CTLs such as Dactl ([4]) and MONSTR ( 1] on a variety of (data flow and graph rewriting) machines such as Flagship ( 10] has been extensively studied. For more information on TGRS, Dactl and MONSTR the reader is advised to consult references [1,3,4,9] and [2] in these proceedings. In this paper we discuss some of the ....

....of TGRS themselves, by means of associated CTLs such as Dactl ( 4] and MONSTR ( 1] on a variety of (data flow and graph rewriting) machines such as Flagship ( 10] has been extensively studied. For more information on TGRS, Dactl and MONSTR the reader is advised to consult references [1,3,4,9] and [2] in these proceedings. In this paper we discuss some of the issues which we feel they have played an important role in the success of Dactl as an intermediate formalism bridging the gap between user level languages and (parallel) machine architectures. We believe that these issues are ....

J. R. W. Glauert, J. R. Kennaway, and M. R. Sleep, Dactl: An Experimental Graph Rewriting Language, Graph Grammars and Their Applications to Computer Science, LNCS 532, Springer Verlag, 1990, pp. 378395.

....by a connection which is principal for both of them is the analogue of the cut, and its elimination is a rewrite rule for such graphs. Thus arises the Interaction Net model of graph rewriting. MONSTR ( 2] originated from the desire to implement the generalised term graph rewriting language DACTL ([16,17]) on a distributed parallel machine, the Flagship machine ( 26] The demands of (even an imperfectly adhered to notion of) serialisability for DACTL executions, necessitated curtailing the expressive power of DACTL rules rather drastically. Because it was vital for the Flagship machine that the ....

....c8:Port f4:Port Fig. 8 Rew Rew Rew Rew 17 6 Examples and Comparison with Similar Work In this section we provide the MONSTR code for a number of concrete programming examples, covering both the cases of typed and untyped nets; this code has run successfully on the Dactl interpreter ([16,17]) We also compare the resulting MONSTR code with that produced when programs written in other computational models similar to Interaction Nets, are likewise translated to sets of MONSTR rewrite rules. This comparison further highlights some features of MONSTR and also puts each into a wider ....

[Article contains additional citation context not shown here]

Glauert J. R. W., Kennaway J. R., Sleep M. R., "DACTL:An Experimental Graph Rewriting Language", in [13], pp. 378-395.

.... by several application areas has led to the theory of graph grammars and graph transformation (see volume I of the Handbook and [1 7] for a survey) A wide spectrum of approaches exists within this theory and some of them are implemented (see, for example, Progres [8, 9] Graph Ed [10] Dactl [11], and Agg [12, 13] With the aim of enhancing the usefulness of graph transformation, we introduce approach independent structuring methods for building up large systems of graph transformation rules from small pieces. The methods are based on the notion of a transformation unit and its ....

John R. W. Glauert, J. Richard Kennaway, and M. Ronan Sleep. Dactl: An experimental graph rewriting language. In Ehrig et al. [4], pages 378-- 395.

.... to the theory of graph grammars and graph transformation (see [CER79, ENR83, ENRR87, EKR91, SE94, AEH 96, CEER96] for a survey) A wide spectrum of approaches exists within this theory and some of them are implemented (see, for example, PROGRES [Sch91a, Sch91b] Graph Ed [Him91] Dactl [GKS91] and AGG [LB93, TB94] With the aim of enhancing the usefulness of graph transformation, we propose a new approach independent structuring method for building up large systems of graph transformation rules from small pieces. The method is based on the notion of a transformation unit and its ....

John R. W. Glauert, J. Richard Kennaway, M. Ronan Sleep. Dactl: An experimental graph rewriting language. In Ehrig et al. [EKR91], 378--395.

....C.1.3, D.1.3, D.3.1, F.3.2, F.4. 2 1 INTRODUCTION This is the first of a series of papers about MONSTR, in which we will study the problem of implementing an extended term graph rewriting model of computation described by the language DACTL [Glauert et al. 1988a) Glauert et al. 1988b) [Glauert et al. 1990)] on a distributed store architecture, from a fairly theoretical vantage point. The issue arose in the late 80 s when the Flagship Project intended to use DACTL as a general purpose intermediate language for a distributed store multiprocessor, the Flagship Machine, Watson and Watson (1987) ....

Glauert J.R.W., Kennaway J.R., Sleep M.R., DACTL: An Experimental Graph Rewriting Language. in: Graph Grammars and their Application to Computer Science, Ehrig, Kreowski, Rozenberg, (eds.), LNCS 532, 378-395, Springer, (1990).

....which we feel are essential and must be supported by this formalism. The computational model in question which acts as an interface between languages and machine architectures is that of Term Graph Rewriting Systems ( 6] and, more to the point, its associated Compiler Target Language Dactl ([17,18,19]) The rest of the paper is organised as follows: the next section introduces briefly TGRS, the language Dactl and its role as an intermediate formalism and the following one discusses those features of the language which we judged have played an important role in the success of Dactl as an ....

....rewrite rules. It follows that a language based on rewriting theory has the potential of being an intermediate language for a number of declarative languages. A number of languages based on TGRS have been designed and implemented but here we will concentrate on one of them, namely Dactl ([17,18,19]) which we believe it is the most flexible of all. Dactl was essentially the target language of the Flagship project ( 27] part of the Alvey ( 37] and EDS ( 13] programmes, whose purpose was the design and implementation of a parallel computing system able to support different declarative ....

[Article contains additional citation context not shown here]

J. R. W. Glauert, J. R. Kennaway and M. R. Sleep, "Dactl: An Experimental Graph Rewriting Language", Proceedings Graph Grammars and Their Applications to Computer Science, LNCS 532, Springer Verlag, 1990, pp. 378395.

.... areas has led to the theory of graph grammars and graph transformation (see [CER79, ENR83, ENRR87, EKR91, SE93] for a survey) A wide spectrum of approaches exists within this theory and some of them are implemented (see, for example, PROGRES [Sch91a, Sch91b] Graph Ed [Him91] Dactl [GKS91], and AGG [LB93, TB93] With the aim of enhancing the usefulness of graph transformation, we propose This work has been supported by COMPUGRAPH II, ESPRIT Basic Research Working Group 7183. a new approach independent structuring method for building up large systems of graph transformation ....

J.R.W. Glauert, J.R. Kennaway, M.R. Sleep. Dactl: An experimental graph rewriting language. In Ehrig et al. [EKR91], 378--395.

....reduction is made available to its caller while the recursive call is executed in parallel. The second rule shows a lazy version; the recursive Append will remain dormant until the original caller activates it again. A more elaborate description of Dactl can be found in the accompanying paper (Glauert et al. 1997). 2 Concurrent logic languages Concurrent logic languages (CLLs) comprise a family of languages based on a subset of first order predicate calculus (Horn clauses) and utilizing stream parallelism. They provide a powerful computational model able to model reactive systems and easily amenable to ....

Glauert, J.R.W. Kennaway, J.R., Papadopoulos, G.A. and Sleep, M.R. (1997) Dactl: an experimental graph rewriting language. Journal of Programming Languages, 5, 75--98.

....adding parallel composition to a rewrite system can change its reduction relation. Whether the results of parallel composition of rewrites are useful is a subject for further study. Dactl is a graph rewrite language intended for parallel implementation (see [GHKPS88, GKS89] and in this volume [GKS90]) Its definition currently stipulates that the result of a computation must be identical to the result of some series of rewrites. This serialisability requirement was imposed to facilitate reasoning about Dactl programs. However, if the semantics of parallel rewrites is, as above, well defined ....

J.R.W.Glauert, J.R.Kennaway and M.R.Sleep "Dactl: An Experimental Graph Rewriting Language", these proceedings, 1990.

....reduction is made available to its caller while the recursive call is executed in parallel. The second rule shows a lazy version; the recursive Append will remain dormant until the original caller activates it again. A more elaborate description of Dactl can be found in the accompanying paper ([13]) 3. Concurrent Logic Languages Concurrent logic languages (CLLs) comprise a family of languages based on a subset of first order predicate calculus (Horn clauses) and utilizing stream parallelism. They provide a powerful computational model able to model reactive systems and easily amenable to ....

J. R. W. Glauert, J. R. Kennaway, G. A. Papadopoulos and M. R. Sleep, Dactl: An Experimental Graph Rewriting Language, Journal of Programming Languages, 1997 (to appear).

.... 9 PARLOG GHC GHC F CP family HOPE ML OO lang Dactl Computational Model (Term Graph Rewriting Systems) FLAGSHIP ALICE GRIP Sequential Implement TRS Linear Logic calculus p For more information on TGRS the reader is advised to consult [21,26] whereas for Dactl appropriate references are ([8,9,10]) The rest of the paper can be viewed as an attempt to add one more box to the top part of the figure shown above, that one of Object Oriented Programming. 3. TERM GRAPH REWRITING WITH RECORDS In order to be able to model some fundamental OOP techniques we propose the introduction of records ....

....logic languages in Dactl such as [2,8,11,12,19] We are given the opportunity to address further this point promptly. 5. 2 Implementation and Performance Aspects A prototype implementation of the framework developed in the previous section has been implemented on top of the Dactl system ([9,10]) The only aggregate data structure that Dactl supports is arrays and we have used it to represent a record structure. Names are mapped to unique integer values and operations on records are presented to the underlying system as operations on array elements. Thus, our sequential prototype ....

Glauert J. R. W., Kennaway J. R., Papadopoulos G. A. and Sleep M. R., Dactl: An Experimental Graph Rewriting Language, Journal of Programming Languages, 1997, (to appear).

No context found.

Glauert J.R., Kennaway J.R., Sleep M.R.: Dactl: An Experimental Graph Rewriting Language, in: [8], 378-395

No context found.

Glauert J.R., Kennaway J.R., Sleep M.R.: Dactl: An Experimental Graph Rewriting Language, in: [15], 378-395

No context found.

Glauert J.R.W., Kennaway J.R., Sleep M.R. (1991), DACTL: An Experimental Graph Rewriting Language.

No context found.

Glauert J. R. W., Kennaway J. R. and Sleep M. R., `Dactl: An Experimental Graph Rewriting Language', Proc. GRA GRA, LNCS 532, Springer Verlag, 1990, pp 378395.

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