Grammarware comprises grammars and all grammar-dependent software. The term grammar is meant here in the sense of all established grammar formalisms and grammar notations including context-free grammars, class dictionaries, and XML schemas as well as some forms of tree and graph grammars. The term grammar-dependent software refers to all software that involves grammar knowledge in an essential manner. Archetypal examples of grammar-dependent software are parsers, program converters, and XML document processors. Despite the pervasive role of grammars in software systems, the engineering aspects of grammarware are insufficiently understood. We lay out an agenda that is meant to promote research on increasing the productivity of grammarware development and on improving the quality of grammarware. To this end, we identify the problems with the current grammarware practices, the barriers that currently hamper research, and the promises of an engineering discipline for grammarware, its principles, and the research challenges that have to be addressed.

/pdf/toward-an-engineering-discipline-for-grammarware-4w9qo97p4f.pdf

Toward an engineering discipline for grammarware

Context-free hypergraph grammars and attribute grammars generate the same class of term languages. Extending the context-free hypergraph grammar with a context-free grammar and a semantic domain, a syntax-directed translation device is obtained that is equivalent to the attribute grammar.

Context-free hypergraph grammars have the same term-generating power as attribute grammars

The aim of this paper is to outline the support the theory of graph grammars offers to deal with parallelism and concurrency. The considerations in Part 1 are based on derivations in graph grammars (using a simple, but general framework). Sequentialization and parallelization of derivation steps are studied. From the point of view of concurrency, these constructions induce an equivalence on derivations. It turns out that each equivalence class is uniquely represented by a canonical derivation, which is minimal with respect to a delay index. Part 2 deals with a kind of non-sequential processes in graph grammars (overcoming the sequentiality of derivations).

Is parallelism already concurrency? Part 1: Derivations in graph grammars

Apex graph grammars are a particular type of directed node-label controlled (DNLC) graph grammars: the embedding edges are established between terminal nodes only. Apex graph grammars, slightly generalized, can generate the sets of dependency graphs of attribute grammars. The other way around, every apex graph language can be obtained from such a dependency graph language by a graph replacement (which is an operation analogous to a string homomorphism).

Apex graph grammars and attribute grammars

Bibliography on graph-rewriting systems (graph grammars)

Attribute grammars (atgs, for short) are string rewriting systems, allowing programming languages to be defined together with their context conditions and translations. In this paper graph grammars are used to specify, in a very general way, the evaluators which are generated from atgs. These graph grammars are correct with respect to the languages derived by atgs, and specify meaningful parsers and compilers if the atgs satisfy some weak conditions.

Modelling compiler generation by graph grammars

Derivations from Two-Level Grammars have been defined only in terms of the infinite context-free grammars that can be generated from them. This fact has proven a severe theoretical obstacle to solving the parsing problem for two-level grammars. So far, partial solutions exist for Van Wijngaarden Grammars, Extended Affix Grammars, and Affix Grammars which allow parsing in a single pass from left to right, because these special cases could be handled without developing a full theory of parsing for two-level grammars.

Berthold Hoffmann

Papers

Modelling compiler generation by graph grammars

A Mathematical Approach to Multi-Pass Parsing