Daniel Sundmark

Replay debugging of real-time systems using time machines

In this paper we present a new approach to deterministic replay using standard components. Our method facilitates cyclic debugging of real-time systems with industry standard real-time operating systems using industry standard debuggers. The method is based on a number of new techniques: A new marker for deterministic differentiation between e.g., loop iterations for deterministic reproduction of interrupts and task preemptions, an algorithm for finding well-defined starting points of replay sessions, as well as a technique for using conditional breakpoints in standard debuggers to replay the target system. We also propose and discuss different methods for deterministic monitoring, and provide benchmarking results from an industrial strength case study demonstrating the feasibility of our method. Previously published solutions to the problem of debugging real-time systems have been based on the concept of deterministic replay: where significant system events like task-switches of multitasking software and external inputs are recorded during run-time, and later replayed (re-executed) off-line. Previous works have been based on either non-standard hardware, specially designed compilers or modified real-time operating systems. The reliance on non-standard components has limited the success of the approach. Even though this idea has been around for 20 years, no industrial application for debugging of real-time systems of the method has been presented.

Factors Limiting Industrial Adoption of Test Driven Development: A Systematic Review

Test driven development (TDD) is one of the basic practices of agile software development and both academia and practitioners claim that TDD, to a certain extent, improves the quality of the code produced by developers. However, recent results suggest that this practice is not followed to the extent preferred by industry. In order to pinpoint specific obstacles limiting its industrial adoption we have conducted a systematic literature review on empirical studies explicitly focusing on TDD as well as indirectly addressing TDD. Our review has identified seven limiting factors viz., increased development time, insufficient TDD experience/knowledge, lack of upfront design, domain and tool specific issues, lack of developer skill in writing test cases, insufficient adherence to TDD protocol, and legacy code. The results of this study is of special importance to the testing community, since it outlines the direction for further detailed scientific investigations as well as highlights the requirement of guidelines to overcome these limiting factors for successful industrial adoption of TDD.

Model-Based Test Suite Generation for Function Block Diagrams Using the UPPAAL Model Checker

A method for model-based test generation of safety-critical embedded applications using Programmable Logic Controllers and implemented in a programming language such as Function Block Diagram (FBD) is described. The FBD component model is based on the IEC 1131 standard and it is used primarily for embedded systems, in which timeliness is an important property to be tested. Our method involves the transformation of FBD programs with timed annotations into timed automata models which are used to automatically generate test suites. Specifically we demonstrate how to use model transformation for formalization and model-checking of FBD programs using the UPPAAL tool. Many benefits emerge from this method, including the ability to automatically generate test suites from a formal model in order to ensure compliance to strict quality requirements including unit testing and specific coverage measurements. The approach is experimentally assessed on a train control system in terms of consumed resources.

A Framework for Comparing Efficiency, Effectiveness and Applicability of Software Testing Techniques

Software testing is expensive for the industry, and always constrained by time and effort. Although there is a multitude of test techniques, there are currently no scientifically based guidelines for the selection of appropriate techniques of different domains and contexts. For large complex systems, some techniques are more efficient in finding failures than others and some are easier to apply than others are. From an industrial perspective, it is important to find the most effective and efficient test design technique that is possible to automate and apply. In this paper, we propose an experimental framework for comparison of test techniques with respect to efficiency, effectiveness and applicability. We also plan to evaluate ease of automation, which has not been addressed by previous studies. We highlight some of the problems of evaluating or comparing test techniques in an objective manner. We describe our planned process for this multi-phase experimental study. This includes presentation of some of the important measurements to be collected with the dual goals of analyzing the properties of the test technique, as well as validating our experimental framework

Automated test generation using model checking: an industrial evaluation

In software development, testers often focus on functional testing to validate implemented programs against their specifications. In safety-critical software development, testers are also required to show that tests exercise, or cover, the structure and logic of the implementation. To achieve different types of logic coverage, various program artifacts such as decisions and conditions are required to be exercised during testing. Use of model checking for structural test generation has been proposed by several researchers. The limited application to models used in practice and the state space explosion can, however, impact model checking and hence the process of deriving tests for logic coverage. Thus, there is a need to validate these approaches against relevant industrial systems such that more knowledge is built on how to efficiently use them in practice. In this paper, we present a tool-supported approach to handle software written in the Function Block Diagram language such that logic coverage criteria can be formalized and used by a model checker to automatically generate tests. To this end, we conducted a study based on industrial use-case scenarios from Bombardier Transportation AB, showing how our toolbox CompleteTest can be applied to generate tests in software systems used in the safety-critical domain. To evaluate the approach, we applied the toolbox to 157 programs and found that it is efficient in terms of time required to generate tests that satisfy logic coverage and scales well for most of the programs.

Test case quality in test driven development: A study design and a pilot experiment

Background: Test driven development, as a side-effect of developing software, will produce a set of accompanied test cases which can protect implemented features during code refactoring. However, recent research results point out that successful adoption of test driven development might be limited by the testing skills of developers using it. Aim: Main goal of this paper is to investigate if there is a difference between the quality of test cases created while using test-first and test-last approaches. Additional goal of this paper is to measure the code quality produced using test-first and test-last approaches. Method: A pilot study was conducted during the master level course on Software Verification & Validation at Malardalen University. Students were working individually on the problem implementation by being randomly assigned to a test-first or a test-last (control) group. Source code and test cases created by each participant during the study, as well as their answers on a survey questionnaire after the study, were collected and analysed. The quality of the test cases is analysed from three perspectives: (i) code coverage, (ii) mutation score and (iii) the total number of failing assertions. Results: The total number of test cases with failing assertions (test cases revealing an error in the code) was nearly the same for both test-first and test-last groups. This can be interpreted as “test cases created by test-first developers were as good as (or as bad as) test cases created by test-last developers”. On the contrary, solutions created by test-first developers had, on average, 27% less failing assertions when compared to solutions created by the test-last group. Conclusions: Though the study provided some interesting observations, it needs to be conducted as a fully controlled experiment wit

Monitored software components - a novel software engineering approach

We propose monitoring of software components, and use of monitored software components, as a general approach for engineering of embedded computer systems. In our approach, a components execution is continuously monitored and experience regarding component behaviour is accumulated. As more and more experience is collected the confidence in the component grows; with the goal to eventually allow certification of the component. Continuous monitoring is also the base for contract checking, and provides means for post-mortem crash analysis; an important prerequisite for many companies to start use 3rd party component in their dependable systems. In this paper we show how four software engineering goals can be reached by monitoring four component properties.

Technical Debt in Test Automation

Automated test execution is one of the more popular and available strategies to minimize the cost for software testing, and is also becoming one of the central concepts in modern software development as methods such as test-driven development gain popularity. Published studies on test automation indicate that the maintenance and development of test automation tools commonly encounter problems due to unforeseen issues. To further investigate this, we performed a case study on a telecommunication subsystem to seek factors that contribute to inefficiencies in use, maintenance, and development of the automated testing performed within the scope of responsibility of a software design team. A qualitative evaluation of the findings indicates that the main areas of improvement in this case are in the fields of interaction design and general software design principles, as applied to test execution system development.

Impediments for Automated Testing -- An Empirical Analysis of a User Support Discussion Board

To better understand the challenges encountered by users and developers of automatic software testing, we have performed an empirical investigation of a discussion board used for support of a test automation framework having several hundred users. The messages on the discussion board were stratified into problem reports, help requests, development information, and feature requests. The messages in the problem report and help request strata were then sampled and analyzed using thematic analysis, searching for common patterns. Our analysis indicate that a large part of the impediments discussed on the board are related to issues related to the centralized IT environment, and to erroneous behaviour connected to the use of the framework and related components. We also observed a large amount of impediments related to the use of software development tools. Turning to the help requests, we found that the majority of the help requests were about designing test scripts and not about the areas that appear to be most problematic. From our results and previous publications, we see a clear need to simplify the use, installation, and configuration of test systems of this type. The problems attributable to software development tools suggest that testers implementing test automation need more skills in handling those tools, than historically has been assumed. Finally, we propose that further research into the benefits of centralization of tools and IT environments, as well as structured deployment and efficient use of test automation, is performed.

MOS: an integrated model-based and search-based testing tool for function block diagrams

In this paper we present a new testing tool for safety critical applications described in Function Block Diagram (FBD) language aimed to support both a model and a search-based approach. Many benefits emerge from this tool, including the ability to automatically generate test suites from an FBD program in order to comply to quality requirements such as component testing and specific coverage measurements. Search-based testing methods are used to generate test data based on executable code rather than the FBD program, alleviating any problems that may arise from the ambiguities that occur while creating FBD programs. Test cases generated by both approaches are executed and used as a way of cross validation. In the current work, we describe the architecture of the tool, its workflow process, and a case study in which the tool has been applied in a real industrial setting to test a train control management system.