Evaluating the algorithmic behavior of interactive systems is complex and time-consuming. Developers increasingly recognize the importance of accountability for their algorithmic creations’ unanticipated behavior and resulting implications. To mitigate this phenomenon, developers not only need to concentrate on the observable inaccuracies that can be measured quantitatively but also the more subjective outcomes that can perpetuate social bias, which are challenging to identify. We require a new approach that involves humans in scrutinizing algorithmic behavior. It leverages a combination of quantitative and qualitative methods to support an ethical, value-aligned design and a system’s lifecycle, informed by users’ perception and values. To date, the literature lacks an agreed-upon framework for such an approach. Consequently, we propose an oversight framework, Modular Oversight Methodology (MOM), which aids developers in assessing the behavior of their systems by involving a carefully crowdsourced society-in-the-loop. The framework facilitates the development and execution of an oversight process and can be tweaked according to the domain and application of use. Through such an oversight process, developers can assess the human perception of the algorithmic behavior under inspection, and extract valuable insights that will aid in assessing its implications. We present the MOM framework, as a first step toward tailoring more robust, domain-specific solutions to exercise human oversight over algorithms, as a means for software developers to keep the generated output of their solutions fair and trustworthy.

Product Lifecycle Management (PLM) is one of the most business-critical IT backbones of manufacturing companies. It often consists of numerous, rigidly interwoven monolithic applications and is seen as synonymous with costly maintenance, lack of extensibility, and poor scalability. This paper proposes an approach for transforming a monolithic PLM landscape into a federated Domain and Data Mesh. This enhances semantic interoperability and enables data-driven use cases by treating data as first-class citizens. User-centric PLM domains moreover help to increase productivity in the workplace.

The use of models for regulatory purposes requires accurate descriptions of important chemical fate processes, techniques for applying models for specific regulatory issues, techniques for parameterizing the model, and sufficient model distribution and support. Researchers (i.e. model developers), on the other hand, want to conceive more and more complex, sophisticated models in order to better represent the important environmental processes; unfortunately this invariably tends to lead to models with greater data requirements, parameters that are difficult to evaluate, greater expertise needed for application, and greater computational requirements, all of which makes for more difficult model application and use. The development of usable models for both current needs and future regulatory requirements demands that we achieve a balance between these ‘research needs' and ‘user needs'. Software engineering and code development issues can help to establish the needed bridge between research and user needs so that future models can more quickly escape the research arena and satisfy user needs as effective regulatory tools.

We combine a new data model, where the random classification is subjected to rather weakrestrictions which in turn are based on the Mammen−Tsybakov [E. Mammen and A.B. Tsybakov,Ann. Statis. 27 (1999) 1808–1829; A.B. Tsybakov,Ann. Statis. 32 (2004) 135–166.] small margin conditions,and the statistical query (SQ) model due to Kearns [M.J. Kearns, J. ACM45 (1998) 983–1006] to what we refer to as PAC + SQ model. We generalize the classconditional constant noise (CCCN) model introduced by Decatur [S.E. Decatur, inICML ’97: Proc. of the Fourteenth Int. Conf. on Machine Learn. MorganKaufmann Publishers Inc. San Francisco, CA, USA (1997) 83–91] to the noise modelorthogonal to a set of query functions. We show that every polynomial time PAC + SQ learning algorithm can beefficiently simulated provided that the random noise rate is orthogonal to the queryfunctions used by the algorithm given the target concept. Furthermore, we extend theconstant-partition classification noise (CPCN) model due to Decatur [S.E. Decatur, inICML ’97: Proc. of the Fourteenth Int. Conf. on Machine Learn. MorganKaufmann Publishers Inc. San Francisco, CA, USA (1997) 83–91] to what we call theconstant-partition piecewise orthogonal (CPPO) noise model. We show how statisticalqueries can be simulated in the CPPO scenario, given the partition is known to thelearner. We show how to practically use PAC +SQ simulators in the noise model orthogonal to the query space bypresenting two examples from bioinformatics and software engineering. This way, wedemonstrate that our new noise model is realistic.

This paper presents the development and application of a software tool for modeling knowledge to be used in knowledge-based systems or the Semantic Web. The inferential modeling technique, which is a technique for modeling the static and dynamic knowledge elements of a problem domain, provided the basis for the tool. A survey of existing knowledge modeling tools revealed they typically failed to provide support in four main areas: support for an ontological engineering methodology or technique, support for dynamic knowledge modeling, support for dynamic knowledge testing, and support for ontology management. Dyna, a Protégé plug-in, has been developed, which supports the Inferential Modeling Technique, dynamic knowledge modeling, and dynamic knowledge testing. Protégé and Dyna are applied for constructing an ontological model in the domain of selecting a remediation technology for petroleum contaminated sites. Dynamic knowledge testing in Dyna enabled creation of a more complete knowledge model.

Tracers provide users with useful information about program executions. In thisarticle, we propose a “tracer driver”. From a single tracer,it provides a powerful front-end enabling multiple dynamic analysis tools to beeasily implemented, while limiting the overhead of the trace generation. Therelevant execution events are specified by flexible event patterns and a largevariety of trace data can be given either systematically or “ondemand”. The proposed tracer driver has been designed in the contextof constraint logic programming (CLP); experiments have been made withinGNU-Prolog. Execution views provided by existing tools have been easily emulatedwith a negligible overhead. Experimental measures show that the flexibility andpower of the described architecture lead to good performance. The tracer driveroverhead is inversely proportional to the average time between two tracedevents. Whereas the principles of the tracer driver are independent of thetraced programming language, it is best suited for high-level languages, such asCLP, where each traced execution event encompasses numerous low-level executionsteps. Furthermore, CLP is especially hard to debug. The current environments donot provide all the useful dynamic analysis tools. They can significantlybenefit from our tracer driver which enables dynamic analyses to be integratedat a very low cost.

Rationale research in software development is a challenging area because although there is no shortage of advocates for its value, there is also no shortage of reasons for why rationale is unlikely to be captured in practice. Despite more than 30 years of research there still remains much uncertainty: how useful are the potential benefits and how insurmountable are the barriers? Will the value of the rationale (design and otherwise) justify the cost of collecting it? Although there have been numerous rationale research projects, many, if not most, received little or no empirical evaluation. There also have not been many studies examining what the needs are of the practitioners who would be supported by the rationale. This article discusses the “doom and gloom” predictions of rationale's failure, provides a survey of evaluations of rationale systems, and discusses what we hope is a brighter outlook for rationale research in the future. There are development standards and synergistic research areas that may help with rationale research and its acceptance in the software community with which we should be working. This article also presents the results of a pilot survey of software developers who were asked how they would envision using rationale and what they believe the most important barriers are. Although some results were as expected, there were also some surprises. Research on technology transfer indicates that, among other things, to transition successfully from research into practice we need to understand the need that is being met and demonstrate the value of our approach. Until we have determined how our work is needed by the people we are trying to help we will remain researching under uncertainty.

Refactoring is an established technique from the object-oriented (OO) programming community to restructure code: it aims at improving software readability, maintainability, and extensibility. Although refactoring is not tied to the OO-paradigm in particular, its ideas have not been applied to logic programming until now. This paper applies the ideas of refactoring to Prolog programs. A catalogue is presented listing refactorings classified according to scope. Some of the refactorings have been adapted from the OO-paradigm, while others have been specifically designed for Prolog. The discrepancy between intended and operational semantics in Prolog is also addressed by some of the refactorings. In addition, ViPReSS, a semi-automatic refactoring browser, is discussed and the experience with applying ViPReSS to a large Prolog legacy system is reported. The main conclusion is that refactoring is both a viable technique in Prolog and a rather desirable one.