Despite its many unquestionable benefits, formal specifications are not widely used in industrial software development. In order to reduce the time and effort required to write formal specifications, in this paper we propose a technique for automatically discovering specifications from real code. The proposed methodology relies on the symbolic execution capabilities recently provided by the K framework that we exploit to automatically infer formal specifications from programs that are written in a non-trivial fragment of C, called KERNELC. Roughly speaking, our symbolic analysis of KERNELC programs explains the execution of a (modifier) function by using other (observer) routines in the program. We implemented our technique in the automated tool KINDSPEC 2.0, which generates axioms that describe the precise input/output behavior of C routines that handle pointerbased structures (i.e., result values and state change). We describe the implementation of our system and discuss the differences w.r.t. our previous work on inferring specifications from C code.

We propose an automated technique for inferring software contracts from programs that are written in a non-trivial fragment of C, called KernelC, that supports pointer-based structures and heap manipulation. Starting from the semantic definition of KernelC in the K framework, we enrich the symbolic execution facilities recently provided by K with novel capabilities for assertion synthesis that are based on abstract subsumption. Roughly speaking, we define an abstract symbolic technique that explains the execution of a (modifier) C function by using other (observer) routines in the same program. We implemented our technique in the automated tool KindSpec 2.0, which generates logical axioms that express pre-and post-condition assertions by defining the precise input/output behavior of the C routines.

A semi-on-line automated analysis for fermentation is presented. The system consists o f a modified commercial biosensor analyser YSI 27 and a Gilson dilutor injector. The results of direct glucose monitoring in fermentation broths are presented. Measurements, calibrations and washes are completely automated, with a m a x i m u m analysis frequency o f 16 per hour with dilutions. The detection limits are 50 mg/1 and 41 g/1 with g o o d linearity and a precision o f ___ 5%.

No abstract