Inference graphs: a computational structure supporting generation of customizable and correct analysis components

Abstract

Abstract-Amalia is a generator framework for constructing analyzers for operationally defined formal notations. These generated analyzers are components that are designed for customization and integration into a larger environment. The customizability and efficiency of Amalia analyzers owe to a computational structure called an inference graph. This paper describes this structure, how inference graphs enable Amalia to generate analyzers for operational specifications, and how we build in assurance. On another level, this paper illustrates how to balance the need for assurance, which typically implies a formal proof obligation, against other design concerns, whose solutions leverage design techniques that are not (yet) accompanied by mature proof methods. We require Amaliagenerated designs to be transparent with respect to the formal semantic models upon which they are based. Inference graphs are complex structures that incorporate many design optimizations. While not formally verifiable, their fidelity with respect to a formal operational semantics can be discharged by inspection.

Research Directions in Requirements Engineering

“…For example, model checking [25,52,137] checks behavioral models against temporal-logic properties about execution traces; and model satisfiability [76] checks that there exist valid instantiations of constrained object models, and that operations on object models preserve invariants. The notations listed in this column are notations that simplify and abstract the structure of the model to be verified [19,48], to facilitate automated verification.…”

Prescriptive Semantics for Big-Step Modelling Languages

“…An SIF, by its function, is a semantic definition language, and thus can be potentially compared with our semantic definition framework. Some TGFs adopt an existing formalism as their SIF; for example, higher-order logic [24,25], structural operational semantics [28], graph grammars [6], and forwarding attribute grammars [38]; others devise their own SIFs; for example, execution rules [81], which defines a semantics via its enabling, matching, and firing rules, and template semantics [65], which defines a semantics by instantiating values for semantic parameters and choosing or defining a set of composition operators. While TGFs strive for flexibility and extensibility, to accommodate new notations, I have strived to create a systematic semantic definition framework that clearly defines a BSML semantics.…”

“…For example, choosing a general SIF, such as structural operational semantics [28] or forwarding attribute grammars [38], might seem a good idea because it provides a certain level of systematicness and clarity, and hopefully flexibility and extensibility would follow. But I observe that researchers either report about supporting a limited set of notations (e.g., variations of "Lotos subset", without variables [28]), or report about difficulties with extensibility (e.g., difficulty in modelling the semantics of "events", because the semantic definition is "not trivial" and becomes "verbose" [38]). Conversely, devising a specific SIF, such as "execution rules" [81] and "template semantics" [65], might seem a good idea because it provides flexibility and extensibility, and hopefully systematicness and clarity would follow.…”

“…I call such a tool suite a tool-support generator framework (TGF), which takes the definition of a language, including its semantics, as input, and generates tool support, such as model checking and simulation capability for that language, as output [81,25,28,75,65,6,38,87]. TGFs differ in the semantic input formats (SIF) they use, and the procedure by which they obtain tool support for a language.…”

“…TGFs differ in the semantic input formats (SIF) they use, and the procedure by which they obtain tool support for a language. An SIF can be an existing formalism, such as higher-order logic [25], structural operational semantic format [28], or a new formalism, such as template semantics [75,74]. A TGF often strives for open-ended flexibility and extensibility, to accommodate new notations, and thus its SIF is a general, expressive format for semantic definition.…”