Currently, adult higher education software engineering pedagogy isolates the student in a controlled environment during delivery, with application of their learning temporally distant from their professional practice. Delivering software engineering teaching that is immediately relevant to professional practice remains an open challenge. In this paper, we discuss a new pedagogical model which addresses this problem by embedding the validation of the student's learning within their rich professional context. We discuss our experience of applying the model to the design and delivery of a new post-graduate software development module, a core component in our new software engineering Masters qualification at the Open University, UK, a market leader in adult higher education at a distance.

The Open University is a distance-based higher education institution. Most of our students are in employment and study from home, contacting their tutor and fellow students via e-mail and discussion forums. In this paper, we describe our undergraduate and postgraduate modules in the software systems area, how we teach them at a distance, and our focus on shifting our students' minds into a reflective, critical, holistic socio-technical view of software systems that is relevant to their particular professional contexts.

The Journal Of Knowledge Engineering!As you will have noticed, Expert Systems has a new format, designed to provide a more capacious home for the increasing number of quality manuscripts that we receive.There are other changes as 2013 brings new opportunities for publishing: Expert Systems' authors will now have the option to publish using our Open Publishing model

Englishman Sir Paul Smith's catwalk shows are renown the world over for their style, brashness and audience as well as for their terrific clothes. He somehow manages to take, each and every year, a step away from the current fashion to define the next season. Fashion started way back in 1858 when another Englishman, Charles Frederick Worth, opened his high fashion shop in Paris. Today, the importance of the professional designer in a changing fashion world has increased to the point when almost all clothes in almost any high street shop, from bijoux boutique to colossal chain, is inspired by their work. Every season, the ‘Fashion Problem’ – that of coming up with tomorrow's fashions today – reinvents itself; the fashion dynamic is cyclic, recurring, revisited. As new fashion becomes the norm, last season's fashion problem is reborn to be solved anew; the Fashion Problem begets each new season's fashion problem. What changes from season to season? Simply that the context into which the new fashion must fit is changed by last season's fashion. The new fashion problem, while conceived of the old and tangled with it, has a new solution that sustains the next. It's wicked. 1973 saw Rittel and Webber coin the term ‘wicked’ to describe problems that are “difficult or impossible to solve because of incomplete, contradictory, and changing requirements that are often difficult to recognise.” Wicked problems have always fascinated me. There's a lot of mystique – fear even – around them: to me their characterisation as insolvable makes their status as problems dubious, and traditional explanations of their insolubility – such as “requirements are volatile” – while having some truth don't seem to explain everything: some people can handle problems with volatile requirements and some can't, but Rittel and Webber's definition never told me why. Jeff Conklin characterises wicked problems in another way: their stake-holders understand the problems differently and the things constraining the solution and the resources needed to solve the problem change over time. He adds, somewhat paradoxically, that they are not understood until after formulation of a solution yet are never solved. Our Fashion Problem's seasonal instances tangle together, last season's solution becomes part of this season's constraints; this season's problem can't be understood until last season's problem is solved, but the Fashion Problem per se is never solved. According to Conklin's criteria, our Fashion Problem is truly wicked. It's a bit like a Paul Smith scarf, tangling back on itself. Not even the Paul Smiths of this world can solve the Fashion Problem. No matter, because each season's instance of the fashion problem can be solved – if you're Paul Smith, at least. Perhaps the recognition of the tangling of instances of a wicked problem11 More generally, instances of a subproblem of a wicked problem. holds some hope for a new approach to their treatment?

Early software development approaches focused on program code as paramount, paying little attention to any analysis or understanding of its generative real-world context and needs. More modern problem-oriented approaches view software development as problem solving, and use many of the rich toolset and techniques of the knowledge engineering (KE) area. ‘When one encounters a new situation (or makes a substantial change in one's view of the present problem) one selects from memory a structure called a Frame. This is a remembered framework to be adapted to fit reality by changing details as necessary’. ‘We intuitively believe that experienced problem solvers construct new systems by modifying known systems’. Both Problem Frames and Domain Theory use modelling as a means for problem understanding and representation. In this respect, they are akin to the Common KADS framework (Breuker & van de Velde, 2004) that provides a suite of re-usable knowledge models for expert systems development. According to Studer et al. (1998), modelling marks an important transition from first-generation to second-generation expert systems, as the modelling approach to knowledge engineering had taken over the knowledge transfer approach. Problem-oriented approaches to software development have two essential attributes. Firstly, they recognise the need to work iteratively between real-world problem and software solution to increase understanding: solutions are sometimes straw men, allowing the developer to reflect on the problem to understand it better. Indeed, the importance of this knowledge building process in design has been widely acknowledged (Alexander, 1979; Dasgupta, 1991). Secondly, they view software development as problem solving and knowledge reuse. This special section collects the best papers presented at the International Workshop on Advances and Applications of Problem Orientation, held in Cape Town, South Africa, in May 2010. These papers represent a snapshot of current research and practice in problem-oriented approaches to software development. In the first paper, Luigi Lavazza addresses the problem that during analysis and development user requirements are elicited and perceived at various levels of abstraction, by proposing a systematic approach to relating such diverse requirements through the establishment of requirements hierarchies. The work is situated within Problem Frames and applied to a comprehensive case study. The work is particularly relevant to addressing business problems in which a need arises to relate business goals to software requirements. In the second paper, Xinyu Wang and his colleagues present an approach that guides a business analyst to derive a set of problem frames from a given problem description. The approach contains three steps. In the first step, business process models are developed to capture the behavioural view of the problem context. In the second step, object analysis models are used to capture the structural view of the problem context. Together, these two views collectively and comprehensively capture the early context knowledge. These two types of model are then used in a third step to construct context diagrams and problem frames. Finally, Jelena Marincic and her colleagues address the problem of how to construct adequate problem models, the starting point of solution seeking, through methods that properly account for both formal and informal modelling decisions and steps. Their approach is located within an industrial setting and the development of embedded systems. Action research was applied in such a setting to help improve problem analysis and problem model validation. Collectively, this special section aims to give a flavour of current thinking in this important area of research and practice. We hope this special section will continue to stimulate the debate and discussion.

Kenneth and Will Hopper’s book “The Puritan Gift: Triumph, Collapse and Revival of an American Dream” is an entertaining and thought provoking read that places the ills of global financial chaos squarely at the door of ‘(so-called) Experts’, modern followers of Frederick Winslow Taylor (1856–1915).

The current UK debate on Computing in schools is driven by a specific issue: what should form the basis of a modern Computing education. It’s a perfect storm that has been brewing for some time, but the recent comments of Eric Schmidt, the chairman of Google, have moved it into the political arena. Currently, the mandatory educational component of Computing (from years 11 to 16) is taught as Information and Communication Technologies (ICT) – effectively, Digital Literacy – which sees Computing, in a teleological light, as a tool that drives society.

Reading Dan Ariely's book Predictably Irrational1, it struck me that Computer Science professionals would benefit from having an oath, something to play the role that the Hippocratic oath once held for doctors. Ariely's book describes an experiment in which contemplation of a ‘moral benchmark,’ such as an oath, raised the general level of responsibility – social, ethical and moral – in a group of people. Such an oath would need to be at the entry point of the Computing profession, so that it delimited the world of the CS student/apprentice from that of the Computing professional: the word professional is derived from the Latin for ‘public declaration’. “I have entered the serious pursuit of new knowledge as a member of the community of graduate students at the University of Toronto. I declare the following: Pride: I solemnly declare my pride in belonging to the international community of research scholars. Integrity: I promise never to allow financial gain, competitiveness, or ambition cloud my judgment in the conduct of ethical research and scholarship. Pursuit: I will pursue knowledge and create knowledge for the greater good, but never to the detriment of colleagues, supervisors, research subjects or the international community of scholars of which I am now a member. By pronouncing this Graduate Student Oath, I affirm my commitment to professional conduct and to abide by the principles of ethical conduct and research policies as set out by the University of Toronto.” Davis et al.s’ oath is a response to their perception, shared by this author, that during their education we, as educators, miss the opportunity to prime our student and discuss with them the social, ethical and moral responsibilities in science research. Their oath emphasises three aspects of scientific research training at the graduate level: community, professionalism, and ethical conduct. In the oath we see three critical characteristics of the research (computer) scientist coming through: pride, in standing alongside others who dedicate their lives to science; integrity, in removing as motives for progression those that are known to damage the community; and, at the core of our community, the pursuit of knowledge and knowledge creation. That last point is telling: and we should note that, as knowledge is the basis in science with its creation the goal, knowledge engineering is the modern key to the creation of knowledge. As such we have a special responsibility to the community in which we practice. Perhaps we should take a step in the right direction by developing and swearing together a Knowledge Engineer's Oath. There are six articles this issue: In ‘A complete chronicle discovery approach: application to activity analysis’, Damien Cram et al. provide insight into the discovery of temporal patterns hidden in a sequence of events, presenting the first chronicle discovery algorithm that is complete. Scheme Emerger has been developed in order to implement the algorithm and to provide real-time graphical support for an interactive chronicle discovery process. Noise attenuation of biomedical signals with a quasi-cyclical character can be achieved through arithmetic averaging. In ‘On application of input data partitioning to Bayesian weighted averaging of biomedical signals’, Alina Momot presents an improvement on traditional arithmetic averaging techniques which assume constant noise power. The techniques employed include Bayesian weighted averaging with traditional (sharp) and fuzzy partition of the input data in the presence of non-stationary noise. In “A decision support system for fund raising management based on the Choquet integral methodology” Barzanti and Giove describe techniques that allow small and medium size not-for-profits to improve their strategies for fund raising based on the organisation's profile, characterised as a hierarchically organised decision tree. The techniques are validated and effectiveness confirmed. In ‘Combining classifiers under probabilistic models: experimental comparative analysis of methods’ Kurzynski and Wozniak present a review of the concept of classifier combination based on the combined discriminant function, in which several recognition algorithms are described. In doing so they introduce the original concept of information unification, to enable the formation of rules on the basis of learning set and vice versa, and go on to formulate new project guidelines for this type of decision-making system. The new techniques are validated both on computer generated data and on data from the medical diagnostics field. Tsai and Tung, in ‘Modified Smith predictor with a robust disturbance reduction scheme for linear systems with small time delays’ present a robust disturbance reduction scheme for linear systems with small time delays that does not require information about unknown load disturbance frequencies. An artificial neural network is used to approximate the unknown load disturbances. Simulation results demonstrate the effectiveness of the scheme as applied to various linear delay systems. Last, but by no means least, ‘Discovering patterns of online purchasing behaviour and a new-product-launch strategy’ by Lun-Ping Hung develops a system to analyse customers’ purchasing behaviour and track shifts in their interests as the basis of recommendation. From this, the author suggests a new-product-launch strategy. The new strategy is tested, and shows almost 40% of potential customers respond to the recommendation positively.

The number of areas to which knowledge engineering can be applied is growing rapidly. A curious knowledge engineer might question which patterns in techniques are discernible across those different areas. One such area of interest to the author is how useful scientific theory becomes engineering knowledge, as might underpin the techniques by which knowledge engineering, i.e., the ways that knowledge is integrated into computer systems, takes place. GFC Rogers1 observes that the knowledge in engineering design differs from that in science: “a theory can be of great practical use for design purposes, yet inadequate from a scientific point of view.” As illustration, Rogers mentions the preferential use of the older, known inadequate, corpuscular theory of light in the design of optical equipment, even though the more comprehensive wave theory that replaced it can deal with diffraction, interference and polarisation too. The modern process by which scientific knowledge loses its scientific validity but not necessarily its engineering utility – by which the corpuscular theory is replaced by the wave theory, for instance – is the subject of the youtube video2 in which Richard Feynman presents on the scientific method. Made in black and white, like many of the great comedy classics, the video is actually quite hilarious given the audience's reaction to what the great scientist is saying. Already, within 7 seconds, Feynman's got his first laugh: the audience guffawing at his assertion that science starts with a Guess. 13 seconds later, after mildly chastising the audience for their reaction, Feynman follows up with Compute the Consequences, and then Compare with Experience. Feynman is propounding Popper's theory of science central to which is falsifiability – that one must be able to test to failure a theory's predictions against observation. With Popper's characterisation of the intimate relationship between failure and science, science sets the bar very high for itself: as Feynman says, even the ‘most beautiful’ guess that cannot be compared to observation is not science. The corpuscular theory of light worked well when the path of light through a lens was required, but failed to predict how light interacted with small apertures – diffraction – with itself – interference – and with certain crystals – polarisation. The wave theory explains all four possibilities. Corpuscular theory, however, retains an arresting simplicity: light travels in straight-lines, like the balls on a billiard table, and that is sufficient for optical engineering. It is the simplicity of the theory – its fitness-for-purpose as applied to a particular3 problem – and not its completeness in explaining many things that is important. Returning to Rogers’ observation; that known ‘inadequate’ optical science can be a fit-for-purpose basis of practice in optical engineering suggests, by analogy, a knowledge science that is a fit-for-purpose basis of knowledge engineering. Such knowledge science may include, but would be more than computational (or computer) science; its topics would include decision making, reasoning, problem solving, and the nature of thought itself. Much has been written recently, not least in this journal, on the use of heuristics, theories and other optimisations that enable the knowledge engineer in their design of expert systems. It wouldn't be too much to say that knowledge engineering would not have reached its current state of development without the existence of these design techniques. One question remains: which is the knowledge science to which knowledge engineering should look for its heuristics, theories, tools and techniques? This is the last issue that will appear in its current small format. From Volume 30, Issue 1, Expert Systems will appear in a larger A4 format. The change will allow us to publish approximately 20% more papers which will allow us to cope with the increasing number of high quality papers. Of course, the vision, mission and scope of Expert Systems remains the same. I look forward to welcoming you to the the new format.

The convergence of 3D simulation and social networking into current multi-user virtual environments has opened the door to new forms of interaction for learning in order to complement the face-to-face and Web 2.0-based systems. Yet, despite a growing user community, design knowledge for virtual worlds remains patchy, particularly when it comes to an understanding of the particular nature of design in virtual environments, the relationship between virtual and real-world contexts of design, as well as the engineering issues it raises and the management of any related risks. In this article, we explore such issues based on our experience of socio-technical engineering of a novel learning programme for higher education with a substantial virtual component. The project's significance stems from the large number of stake-holders involved, the relatively large scale of the virtual world development and the strategic significance of such a development within the learning programme. Of particular novelty is our exploration of the relationship between virtual and real-world contexts of design, indicating when they align and differ, showing when tools and techniques translate and when new tools and techniques may be required.

In what must have been a pleasing outcome, 1950s brain researchers Olds and Milner found the ‘pleasure centre’, a region of the brain whose electrical stimulation gives great, if brief, pleasure. Rats, party animals that probably don't know any better – given the opportunity – choose to have their pleasure centre tickled. That's some feel-good factor. With more talk of quantitative easing (QE) in the United States, the global economy is taking its medicine. New cash will be injected in the hope that economic depression will pass. Too much QE, the currency is devalued, inflation grows uncontrollably to compensate, and disaster follows. Too little QE and the depression extends longer, perhaps growing deeper, and disaster continues. In the second round of QE, QE2, US politicians see the need for the feel-good factor money changing hands gives. The connectionist models of thought which underpin our current understanding of the brain, and which make sense of pleasure and pain, have struggled in their application in other fields. The application of connectionist models – which include Hopfield nets and multi-layer perceptrons, for instance – can give no more than an analytic capability, and so not a predictive capability, of complex social phenomena such as economics. One conclusion is that the ‘economic system can […] not be made rational or subject to “control”’.1 And yet – and yet – control is precisely what we need, and precisely what is intended though QE2. Is there a connectionist mechanism that might provide a predictive capability to underpin what we might call connectionist economics? Santiago Ramón y Cajal was first to identify the neuron as the ‘primary functional unit’ of the human nervous system, an insight that created the modern discipline of neuroscience, and for which he won the 1906 Nobel prize.2 A neuron in the brain gathers electrical stimuli from other neurons until there is enough stimulus to fire. This magnificently simple mechanism combines through sequences of firings, of loops and of tangles to produce thought. Gordon Rogers, who wrote the excellent but out-of-print book ‘The Nature of Engineering’, gave us a view of engineering that, I will argue, gives structures in the an economy that can be seen as mimicking neurons. It's an analogy rather than a direct correspondence, but perhaps – just perhaps – there are economic pleasure centres whose stimulation gives a feel-good factor, but for which the possibility of over-stimulation makes addiction to the stimulus possible. Rogers' definition of engineering talks of the physical world, of identified needs and of artifices. It leads, however, in a quite natural way, to the representation of engineering as solving the problem that is a need within and environment with a solution that, within the environment, satisfies that need. Here's what it means: In a world in which there are potatoes and a need for chips, someone will produce potato peelers. If there's a lack of skills to sharpen potato peeler blades – and the need for chips is great enough – then someone will learn how to sharpen a blade, in the process enabling the technology that makes it happen. With an adequate flow of chips, other problems are solved: the great British treat that is fish and chips is already some of the way into existence. Put simply, engineered solutions collect what they need from the environment and combines them to meet needs. Value is added by this process, and money flows in the opposite direction, back to the environment. That's a bit like how Santiago Ramón y Cajal's neurons work. Those combinations enter other environments to where they can be used to meet further needs. Problems are solved, with those solutions in turn allowing the solving of other problems. Again, value is added by this process, and money cascades from environment to environment. That, too, is a lot like how Santiago Ramón y Cajal's neurons work together. Problem solving is the basis of value creation, and value creation is the basis of our global economy. If engineered solution processes play the role of neurons, then we have identified a connectionist mechanism underpinning the global economy? As such we can ask questions of the economy that we ask of the brain: does the economy work, rest and play? Does it become exuberant and feel ‘pleasure’? Does it feel ‘pain’ and get sad? Is boom and bust a symptom of a depressive disorder? Is it currently depressed? Is QE2 an antidepressant? Is it the right antidepressant, or something more dangerous? We shall see! I'm very pleased to announce two very excellent and welcome additions to our associate editors: Chandra and Chrissanthi. Both Chandra and Chrissanthi have provided excellent service as reviewers to the journal over many years, and it is their dedication to the journal, skill in communicating with authors and professional demeanour that have recommended them for service. Chandra S. Amaravadi is the State Farm Professor of Information Systems at Western Illinois University and recipient of the best paper prize from the National Computer Centre, UK, for predictions of World and Business Computing in 2051. He holds a PhD from the University of Arizona and an MBA from the University of Minnesota. He has served as a consultant with Wellspring Computer Services, in Minneapolis and with Cognizant Technologies in Chennai, India. He regularly does consulting for small businesses. His research interests include neural network applications, knowledge representation for professional knowledge, information extraction for office documents, knowledge management, and ontologies for administrative domains. Chrissanthi Angeli is a professor at the Technological Education Institute of Piraeus, Greece, in the Department of Mathematics and Computer Science. She holds an MSc in Intelligent Systems from the University of Plymouth and a DPhil in Intelligent Fault Detection Techniques from the University of Sussex, both in the United Kingdom. Her research interests include artificial intelligence techniques for fault detection, the combination of artificial intelligence techniques with numerical methods for prediction of faults, real-time expert systems, web-based expert systems, fuzzy systems, neural networks and genetic computing. She has published two books in computing and chapters in scientific books regarding real-time expert systems and prediction of faults using the expert systems paradigm. Add to this many peer reviewed articles in international journals and conference proceedings, and her service as Programme Committee member in many international conferences regarding AI, modelling and simulation. She also serves as a reviewer for many international journals. She is a member of IEEE computer society, IASTED and EUROSIS and is currently the Chair of ASM2011 Conference to be held in Crete, Greece, 22–24 June 2011 (more details available from http://www.iasted.org/conferences/home-715.html). Outside of her scientific interests, Chrissanthi is an experienced photographer and traveller of many of the most beautiful places in the world, on which she has published articles and photographs in appropriate journals. The five papers in this issue voice a broad gamut of technological innovations and engineering application from the control of potentially very large mechanical systems through to facial expressions, medical decisions, tracking and image improvement. In Experimental evaluations of proportional– integral–derivative type fuzzy controllers with parameter adaptive methods for an active magnetic bearing system, Tsai and colleagues deal with the experimental adaptive control of highly non-linear rotating active magnetic bearing systems using fuzzy controllers. Wang and colleagues, in Semi-supervised classification of facial expression using a mixture of multi-variate t distributions, align labelled graph vectors generated from basic facial emotions to mixtures of multi-variate t distributions. The facial expression features are first extracted into labelled graph vectors using the Gabor wavelet transformation. The effectiveness of the new techniques is tested on the Japanese female facial expression database, Ekman's pictures of facial affect database and the AR database. Expert systems have, from early days, been aimed at medical diagnosis and new techniques – and new combinations and applications of traditional techniques – are always welcome. With Competitive/collaborative neural computing system for medical diagnosis in pancreatic cancer detection, Gorunescu and colleagues describe a new expert system that supports the medical decision process using medical imaging databases. Their techniques are applied and illustrated through the differential diagnosis of chronic pancreatitis and pancreatic cancer. In Heuristic particle filter: applying abstraction techniques to the design of visual tracking algorithms, Pantrigo and colleagues demonstrate how real-world visual tracking problems can be modelled as a dynamic optimisation problem within a general framework that hybridises heuristics/metaheuristics with particle filters. In Multiple scale neural architecture for enhancing regions in the colour image segmentation process, Díaz-Pernas and colleagues propose a dynamic multi-scale neural model for colour image enhancement. The authors demonstrate the quality of the visual results obtained and the robustness of the architecture in the presence of different levels of noise, benchmarked against the Berkeley segmentation dataset.

Hotel breakfasts are always alluring and vast; too much of them and you get vast and totally unalluring. That's the hotel breakfast problem for some. Another, more interesting, hotel breakfast problem is how they are ‘engineered’ as solution for feeding their myriad guests, and the role that architectures play in that engineering. Given that a large hotel may serve tens of thousands of breakfasts each year, a place setting for a hotel breakfast expresses expertise that arrives from vast reuse. You'll recognise a pretty standard place setting in the first figure: it places a napkin, a knife, a fork, a spoon and a cup and saucer to give what's needed to eat and drink what's chosen. It doesn't force the choice of what to eat or drink; rather it suggests only what is possible. Importantly, the place setting also copes with some of the variability in the breakfast offering, simultaneously capturing the shared ‘plate-ability’ of ‘English’ and ‘continental’ breakfasts, the ‘bowl-ability’ of cereal, fruit salad or yoghurt, and the ‘cup-ability’ of tea or coffee. A ‘breakfast architecture’. Of course, a place setting isn't breakfast in itself – for that, we need breakfast components, two of which are shown in the second figure; the architecture is simply suggestive of a solution's structure. As such it's a breakfast architecture. This is the general case: contrary to the received wisdom, architectures aren't solutions – but simply help arrive at a solution by suggesting organising structures, at a more or less detailed level, isolating stability in otherwise mutable problem and solution worlds. Two ‘plate-able breakfast components’: a croiscroissant and a piece of cake. So what about knowledge architectures? Effective knowledge architectures bind knowledge components together to solve knowledge problems. As an example, consider the Class architecture, shown below (third figure). Of great historical importance: I estimate that most of the world's students are educated within such structures, it is an especially effective knowledge architecture that places a teacher with content to study with students eager to learn. The Class Knowledge Architecture: Motivated Teacher, Eager Learner. Effective here is defined both with respect to the architecture and the properties of the components: remove either the motivation or the eagerness from teacher of student, and the effectiveness of an instantiated Class disappears. New technology is driven by innovation, science and real-world nous. It's incredibly difficult to arrive at. Even so, knowledge engineering now has a very complete knowledge technology tool-box to work with. No one doubts that early technological development should focus on the technology. However, as a technology matures, the needs of an architecture into which it can fit should become increasingly important, so that it can eventually be used to solve problems. Consider, for instance, an incredibly able expert system for medical diagnosis. As early as 1987, Engle and Flehinger built HEME2, a sophisticated diagnostic system for haematologists. HEME2 could diagnose 142 diseases and was sophisticated and flexible with its ‘automatic self-improvement feature.’ But, sadly for HEME2, its creators concluded that we should: ‘… stop trying to make computers act like diagnosticians …’ The reason they gave up was that they considered no diagnostician sufficiently patient1 to teach a system like HEME2, no matter how good it was at diagnosis. It is notable that theirs isn't a critique of HEME2's constituent technologies or its performance as a diagnostician. Rather they point to its failure as a component of, what we would call, a ‘knowledge solution,’ to coin a phrase. It was a student to a human teacher, arranged something like the Class Knowledge Architecture. What we may assume is that the initially expectant, patient and eager human teacher became demotivated when they realised the enormity of their instructional task. The lessons stopped. Either the learner wasn't sufficiently eager, or the teacher couldn't sustain their motivation for their relationship to continue. Or both. To approach our knowledge potential within a knowledge society we must therefore think not only of the technologies of which it will comprise, but also of the constraints placed by the effective knowledge architectures within which those knowledge components will work together. Building the knowledge society is something to which our community will – as knowledge engineers – make a defining contribution. Moreover – as knowledge technologists – we will contribute the novel knowledge technologies that define the knowledge engineering solution space. However, by analogy with breakfast, we are missing something if part of our contribution is not to identify the knowledge architectures that will structure the solutions for the knowledge society. Moreover, having identified them, it would be remiss if our knowledge technologies did not fit within them. HEME2, as a computational diagnostician, was a great achievement but that simply wasn't enough for HEME2 to be successful. Perhaps, with hind-sight, it would have been better to ‘… stop trying to make computers act like diagnosticians’ and make them act more like eager students. Together with the Special Issue Section, we have two ordinary papers. In the first, authors Ohbyung Kwon and Namyeon Lee, propose a methodology, based on Case Based Reasoning, for deducing the relationship between a user and an object. The methodology may help enable future smart spaces of ubiquitous computing technologies. In the second, Hakan Altinçay performs a deep comparison between three alternative neighbourhood definitions and four manifold-based distance measures and a novel computationally less demanding feature line-based method that exploits geometrical neighbourhoods of test samples for feature line construction which shows promise.

Software problems originate from real world problems. A software solution must address its real world problem in a satisfactory way. A software engineer must therefore understand the real world problem that their software intends to address. To be able to do this, the software engineer must understand the problem context and how it is to be affected by the proposed software, expressed as the requirements. Without this knowledge the engineer can only hope to chance upon the right solution for the problem. Application of problem-oriented approaches may well be a way of meeting this challenge.

Design rationale in software engineering fills in the gaps between the original requirements of a system and the finished product encompassing decisions, constraints and other information that influenced the outcome. Existing research corroborates the importance of design rationale to capture knowledge assets, particularly in the context of the global enterprise, with its increased risk of knowledge loss through staff movement and attrition. Despite this, the practice of design rationale capture and reuse is not as extensive as could be expected mainly due to time and budget constraints and lack of standards and tools. In this paper we report on and industrial study which tested the hypothesis that an emerging design approach - Problem Oriented Engineering (POE) - enables the capture of design rationale as the methodological by-product of software engineering activities, hence potentially addresses concerns over cost effectiveness of in situ design rationale capture. The results are encouraging, providing an early indication that using POE for solving software problems in an industrial setting may allow the capture of design rationale with no additional effort needed.

A very warm welcome to the first issue of volume 27 of Expert Systems: The Journal of Knowledge Engineering! This first issue of 2010 is a special issue on computer recognition systems, and I will very quickly hand over to the guest editors for their editorial. 2010 will see Expert Systems firmly located at the centre of knowledge engineering research with increasing recognition for it in that role. Our drive is to continue to enhance the value that the journal delivers to its readership with its description of leading edge research. Of course, none of this will be possible without the exceptional efforts of the reviewers, the associate editors, the editorial board and Wiley-Blackwell staff, as well as the authors. To all those who will contribute to another exciting year, here is a message in the form of a puzzle. Each letter you see in the grid should appear once in each line, in each column and in each emboldened 3 × 3 square. When you have completed filling in the squares, the message will be on the leading diagonal. Editorial board changes continue apace: It is my great pleasure to welcome Hojjat Adeli to the editorial board of Expert Systems. Professor Adeli is Professor of Civil and Environmental Engineering and Geodetic Science, Director of the Knowledge Engineering Laboratory, and the holder of the Abba G. Lichtenstein Professorship at the Ohio State University. He is also Professor of Aerospace Engineering, Biomedical Engineering, Biomedical Informatics, Electrical and Computer Engineering, Neuroscience and Neurological Surgery by courtesy. Renowned as one the founders of the field of knowledge engineering, his two-volume book Knowledge Engineering (McGraw-Hill, 1990) helped to establish the nascent field. He is the Founder and Editor-in-Chief of the research journals Computer-Aided Civil and Infrastructure Engineering and Integrated Computer-Aided Engineering, and Editor-in-Chief of the International Journal of Neural Systems. He has over 450 publications including 14 books such as Expert Systems for Structural Design – A New Generation (Prentice Hall, 1988) and Machine Learning – Neural Networks, Genetic Algorithms, and Fuzzy Systems (Wiley, 1995) that established the field of computational intelligence. He has been keynote speaker at 74 conferences held in 40 different countries over the past two decades. Professor Adeli is an elected Fellow of the American Association for the Advancement of Science (AAAS), ‘for distinguished contributions to computational infrastructure engineering and for worldwide leadership in computational science and engineering as a prolific author, keynote speaker, and editor-in-chief of journals’. Without further ado, I hand you over to the special issue's guest editors.

It is increasingly easy to imagine a future in which services are delivered through virtual worlds, such as Linden Lab's Second Life. Whereas one might always have to visit the dentist for a filling, the preliminary examination can be conducted through a mash-up of virtual and real images. In fact, with the right technology, nothing purely analytic is impossible virtualized: observation is simply the internalization of an item or an event. Even certain physical products, released from the need for distance-spanning delivery, will shuffle off their physical existence and join the virtual party: virtual visits to the cinema, theatre, ballet and opera. For physical synthesis, however, the situation is more complex: building in a virtual world with the intention of changing the real world is difficult to countenance. In one area, however, synthesis is as easy virtually as it is in the real world: knowledge engineering provides the perfect product for virtual-world synthesis. Knowledge, of course, is already a virtualized product running in the brain. Unfortunately, traditional notions of engineering do not easily translate to the virtual. Consider Rogers's (1983) practice-based definition of engineering:1 [Engineering is] the practice of organising the design and construction of any artifice [solution] which transforms the physical world around us to meet some recognised need. To work with this definition means one must consider virtual as embedded in the real, as illustrated in Figure 1. Rogers embeds the virtual world in the real. The Brain User is the human operator with, perhaps, knowledge needs. Rogers's model is correct in the analysis it enables – we experience the virtual only through physical interaction with the PC. However, it leaves something to be desired when it comes to synthesis in the virtual world, as building the virtual artifice rests on lengthy navigation back and forth through avatars, the PC embedded in the physical world which interacts with our Brain User in their physical world. Is there a simpler model that removes some of this complexity for easier synthesis? The metaphor I am suggesting for virtual engineering is illustrated in Figure 2. Here, we have held a mirror up to the real world to illustrate the symmetry that exists between the virtual world and the real world. It induces a paraphrasing of Rogers, from which we may define virtual engineering as A mirror on the problem: virtual engineering enabled through recognition of the virtualized identified need: the avatar as virtualized brain user. the practice of organizing the design and construction of any virtual artifice which transforms the virtual world around it and our virtual selves to meet some virtual recognized need. Notably, this definition suggests that our real-world engineering skills can and should be transferable to the virtual: there is no reason that organization in the virtual should differ substantially from that in the physical (unless we can get the avatars to do it!). It may also be, of course, that there are new organizational skills to learn from the virtual, to do with enhancing communications for instance, whence our real-world skills are augmented. What isn't clear from our simple restatement of Rogers is the nature of the process by which the real recognized need becomes its virtualized counterpart: more simply, how can we deduce the virtual need from the real one? Examples of the complexity of the relationship will be clear to anyone whose avatar has sat on the virtual beach of any second life world: if you haven't, I recommend the OU's new virtual MPhil on Second Life – hosted on the deep|think islands. The deep|think islands provide a virtual campus for research students where avatars sit on virtual deckchairs under virtual parasols sipping virtual cocktails and watching spectacular virtual sunsets. Virtual birds fly overhead singing virtual birdsong. Later, the avatars will dance the night away in a virtual disco, or go boating: most relaxing, your avatar drifting away into the sunset. Whilst relaxing there, one gets to thinking: for, although the virtual deckchair supports the avatar, it does not support the student whose avatar it is. On the other hand, the virtual birdsong does not calm the avatar, it calms the avatar's owner; just as the virtual sunsets burn themselves not into the avatar's retinas but into real memories. Each virtual artifice has both virtual and real effects, with each virtual effect having a physical emotional counterpart: indeed, it is the emotional response of the real that is reflected by the mirror! Outside of military simulations, 1982's best virtual reality was provided by Microsoft's Flight Simulator, and so it is not too difficult to understand why, for Rogers, the physicality of the real world was of critical importance to engineering. We now know that real-world engineering and virtual-world engineering are related; that the relation is based on a useful symmetry between them that allows skills from one to be used to good effect on the other; but that the translation of real recognized need to the virtual need is complex. As knowledge engineers, we should take a view on this complexity and try to describe it so that real and virtual engineers can come together, to think both outside and inside the box. The deep|think sunsets can be experienced by visiting http://slurl.com/secondlife/Deep%20Think%20East/228/29/34/. (Free registration is required to visit Second Life.) HTTP://www.welcome.ac.uk/News/2010/News/WTX058631.htm In their new ten-year plan, the Wellcome Trust will focus on supporting the best researchers to address five challenges. While each of the challenges is truly worthwhile and valuable, one might be of particular interest to the readers of Expert Systems, and especially to those researchers whose contribution defines it. It reads: Exploring how the billions of nerves in the brain allow us to think, learn and remember, so that we can find new approaches to treating mental illness and neurological disorders such as stroke, Parkinson's and Alzheimer's diseases. What is particularly welcome is the Wellcome Trust's recognition that researchers need freedom and security to pursue long term the questions that will provide answers to these challenges. Writing in their Strategic Plan, 2010–20, the Wellcome Trust writes: We will support research to improve understanding of how the brain functions and to find improved approaches for treating brain and mental health disorders. This will require the characterisation of how nerve cells function and interact in complex networks to enable specific cognitive and behavioural functions. It will also demand a fully integrated approach that links basic and clinical biomedical research, with key inputs from social sciences, humanities and the arts. More details at http://www.wellcome.ac.uk/news/2010/news/wtx058631.htm. Dr Juan I. Arribas, an Associate Editor for Expert Systems, is on his way to visit the Laboratory of Behavioral Neurophysiology, part of Barrow Neurological Institute (BNI, http://www.thebarrow.org/index.htm), in Phoenix, Arizona. The BNI is the largest neurological centre in the southwestern USA. Juan will travel at the behest of the Spanish government which has awarded him a fellowship to study computational models of blood flow measurements during seizure events and their use for neuroprotective therapies. The expected results of Dr Arribas have the potential to affect the quality of life of epilepsy patients everywhere. So, please join with me in wishing Juan a truly productive stay! Juan will, of course, continue his Associate Editor duties while he is away. Ronald Reagan, the 40th President of the United States, repeated this old Russian proverb many times in his long career as a politician. When asked by Mikhail Gorbachev ‘Why?’, he simply said, ‘I like it!’ Perhaps what he liked was the phrase's easy juxtaposition of the faith-based ‘’ (‘Trust’) with the scientist's creed ‘’ (verify). In legalese ‘Trust, but verify’ becomes the assonant, but otherwise unpoetic, ‘due diligence’. Expert Systems is currently trialling iThenticate. iThenticate is a web-based service which indicates overlaps between academic papers. It provides a way of verifying what we trust to be the case: that an author's writing is his/her own. As an example of the iThenticate service in use, I ran this short editorial through the iThenticate system. The system picked up that Rogers's definition also appears in other papers and, in particular, one from 1995 by Professor Tom Maibaum who cites Rogers and gives the same definition in a paper on the teaching of software engineering. Satisfyingly, that's precisely where I first came across Rogers's quote. We have five excellent papers for you in this issue, in a number of areas of our field. To whet your appetite for more – The first, ‘Identification and resolution of conflicts during ontological integration using rules’ by Biletskiy, Ranganathan and Vorochek, shows how larger ontologies can be built from smaller heterogeneous ones, even in the presence of certain semantic conflicts. The work is grounded in OWL, the Ontology Web Language, and also demonstrates how the newer SWRL, the Semantic Web Rule Language, can be used for executable ontology representations. In ‘Discovering hidden knowledge in data classification via multivariate analysis’, Chen, Ip, Li and Wang propose a new three-step classification algorithm based on multivariate analysis for the grading of school transcripts. Claiming high classification accuracy and intuitive data interpretation, this work is supportive of new ways of working in today's educational environment. Iskurt and Becerikli present their system for automatic calculation of the quantities of vascular plaque from intravascular ultrasound (IVUS) images. Validation of the new technique is provided by an algorithm for labelling IVUS images and comparison against expert-produced contours and the contours extracted by their system. The paper is entitled ‘Extraction of media and plaque boundaries in intravascular ultrasound images by level sets and min/max flow’. In ‘Method of classifier selection using the genetic approach’, Jackowski and Wozniak present a compound classifier system machine learning algorithm for sets of area classifiers, whose parameters were chosen through the application of an evolutionary algorithm. The results of experiments confirm that the proposed method outperforms each elementary classifier as well as a simple voting committee. Vague information is the bane of modern life. In our last paper, ‘Knowledge acquisition in vague objective information systems based on rough sets’, Feng, Wang and Li describe new theories, techniques and tools which permit the analysis of vague data sets based on the combination of rough and vague sets theories.

Crosswords are embedded in the British psyche; they were used to indicate which men (yes, men; only men) should attack the Enigma cipher at Bletchley Park (http://www.bletchleypark.org.uk/). From their genesis as a purely definitional form in the 1920s, cryptic crosswords, a construct1 of the sharp mind of one known as Torquemada (Edward Powys Mathers, 1892–1939), and honed to a razor edge by others including Ximenes (Derrick Somerset Macnutt, 1902–1971) and Azed (Jonathan Crowther, born 1942). A cryptic clue in the Ximinean tradition might be: Welcome odd dignitary on top form to middle-of-road club, it's said (12) (which forms the title of this section). It consists of: a precise definition– here ‘Welcome’, so that the answer to the clue means welcome – together with the number of letters in the solutions – here 12; a fair subsidiary indication– here ‘odd dignitary on top form to middle-of-road club, it's said’– which is a more or less complex wordplay. It is the fair subsidiary definitions from which all a cryptic crossword's fun stems! In the case of our section heading, the wordplay is thus: take the ‘odd’ letters of ‘dignitary’, to give ‘INTR’; add to them (‘on’) the middle letter (‘middle-of’) of the phrase ‘top form to’– the letter ‘O’; then add a word meaning ‘road’– could be ‘DUCT’; finishing with a ‘club’– here a golf club an ‘iron’– modified by the ‘it's said’– a phrase that means ‘sounds like’– to ‘ION’. The whole answer is, then, the combination ‘INTR-O-DUCT-ION’ which is, indeed, a 12-letter word meaning welcome! Cryptic crosswords share with other crossword forms the need for the solution to fit into a grid, where across solutions written horizontally tangle together with down solutions written vertically. As simpler clues in a cryptic crossword are discovered, they provide letters that help solve the harder clues. When solutions tangle, even in this simple way, earlier problem solving makes later problem solving easier. More explicit tangling is also used: the Guardian Cryptic Crossword #25046 contains, as clue 21 across: 21a The victims of 9? (12) in which the ‘9’ indicates a dependence on the solution of 9 across: 9a Hold over aspiring beneficiaries? (4,5) This form of interdependence almost always forces a temporal ordering in the solving process: 9 across should have a solution for 21 across to have a complete clue. The solution to 9 across is ‘WILL POWER’; that for 21 across ‘DISINHERITED’. In terms of the definition of a cryptic clue as a pair, the problem part of clue 21a overlaps the solution part of 9a, that is the problems overlap. In the previous editorial, we have suggested that problems tangle, a term intentionally suggestive of complexity, messiness and difficulty facing the solver. In this case, problem part and solution part overlap; a moment's thought longer will reveal only three possibilities for overlap: problem/problem overlap; problem/solution overlap; solution/solution overlap. Of these, the second is asymmetric, with that asymmetry introducing a linearization of the problem-solving process as the solution part should be available for the problem part to be complete. In contrast, the other situations are symmetric. The third form of tangling – solution/solution – gives crosswords their name; it is the sharing of letters between solutions. Note that the sharing in this and the general case is not of the whole solution, just a solution part – here a letter. Even so, given that shared letters assists the solver in resolving a solution from a clue's structure, solution/solution overlaps often simplify the solving process. (Of course, a shared letter with an incorrect solution may also complicate the solving process.) Problem/problem overlaps are far less typical of crosswords: indeed, the only class we are aware of is that of the aggregate clue, such as the following from Guardian Cryptic Crossword #24997: 2d,3d Boss taking the cake, as man on Muscle Beach? (10) 3d See 2d Here, the clue for 2d and that for 3d are the same, with 2d taking the first part of the answer, 3d the second part. Cryptic crosswords are fun and challenging word games. I'd argue that the solving of clues is a form of engineering, closest to knowledge engineering. And, although cryptic crosswords do not share the investment, risk, nor resource needs that more traditional engineering entails, their simple tangling of clues provides an effective illustration of the difficulties facing – and assistance provided to – the solver of tangled problems, whether in engineering or cryptic crosswords. In this issue With four out of six articles addressing the organization, there's something of a business efficiency and effectiveness focus on this month's articles. Read on if you think that organizational decision making and risk management is more difficult than it should be! First, however, in ‘Real-time smart surveillance using motion analysis’ Leo, Spagnolo, D'Orazio, Mazzeo and Distante describe a sophisticated combination of automatic video analysis technologies for monitoring restricted areas and suspicious objects. The context of the study is an Italian museum and its co-located archaeological site. A quadratic to linear reduction in the complexity of solving the modified primal problem for a proximal support vector machine is the goal of the next paper, entitled ‘Iterative support vector machine with guaranteed accuracy and run time’ by Ye, Zhao, Ye and Chen. And so to enterprise decision making! Ablanedo-Rosas, Gao, Zheng, Alidaee and Wang examine the relative efficiency of 11 major Chinese ports using a version of Data Envelopment Analysis (DEA) in ‘A study of the relative efficiency of Chinese ports: a financial ratio-based data envelopment analysis approach’. Their goal is to give managers better information upon which is base improvements in port operating efficiency. Choosing the best decision making unit is the topic of ‘A new DEA-based method for fully ranking all decision-making units’ by Du, Liang, Yang, Bi and Yu. The technique combines standard efficiency with inter-unit influence measures. In ‘Measurement of enterprise knowledge by state characterization’ Xu and Bernard focus on state based modelling of knowledge and knowledge management in an enterprise's approach to profit and the maintenance of competence. In our final paper of this issue, and this year, Xie develops a ‘A model-driven decision support system for product risk analysis’ which simulates the life cycle of products through their design and manufacturing phases. Thank you! The 27th volume of Expert Systems: The Journal of Knowledge Engineering has come to a close. For those who were in at the start, back in 1984, you will now be able to build a cube of your (paper) copies that is three volumes wide, three volumes deep and three volumes high. And there are far too many people to whom the journal owes its debt of gratitude to be able name them all. To our reviewers and readers go our most heartfelt thanks: our community simply would not exist without you. The names of reviewers of 2010 papers appear on the journal homepage (http://wileyonlinelibrary.com/journal/EXSY), down by the twitter link. Onward into Volume 28!

In this position paper, we present an initial Problem Oriented Engineering foundation for a small part of Pattern-Oriented Analysis and Design (POAD). We argue for the interpretation of software design patterns as instructions for transforming quality requirements into an architecture and a 'stress test'; and show, incidentally, how to reuse an existing software component and its justification in a new development.

The paper reports on the application of Problem Oriented Engineering (POE) to the design of a highly innovative post-graduate research programme for the Open University, UK, a world leader in supported distance higher education. The new programme, to be launched in October 2009, is a part-time Master of Philosophy (MPhil) to be delivered entirely at a distance, supported by a blend of synchronous, asynchronous and immersive internet and web technologies. POE is a framework for engineering design under development at the Open University. After a brief description of the project and the task at hand, the paper discusses the overall engineering approach taken, key aspects of product design, the mapping between requirements and technology, and the development of one key technological component.

1983 and all that 1983 was the year that World War III almost began, at least if you believed the film 'War Games.' Matthew Broderick, in one of his best roles, was a troublesome computer geek that brought the world to the brink of total meltdown.By connecting to the the US military's top-secret War Operations Plan Response (WOPR) system for evaluating what to do in the event of a Soviet attack, Broderick mistakenly starts playing 'Global Thermonuclear War' with the US military's computer system.1983 saw the publication of Edward Feigenbaum and Pamela McCorduck's technological thriller 'The Fifth Generation: Artificial Intelligence and Japan's Computer Challenge to the World,' (Feigenbaum and McCorduck, 1983).Feigenbaum and McCorduck's book wasn't-or wasn't meant to be-fiction, but a serious discourse on the way that Japan's technological wizardry would change the world.But, like the cherry blossom, Japan's Computer Challenge wilted and fell away.Another dip into 1983 provides us with a breakthrough in the theory of engineering, a definition of what it means to engineer.GFC Rogers (Rogers, 1983) told us:

Not all beaches attract the sun, but when they do, they attract sunbathers. The photograph was taken near Otranto, Italy, on a wonderfully sunny day. It was difficult to find a place to put even a tiny beach towel. From the air the patchwork of parasols and towels soaked up almost every square centimetre of sun. It was a feat that a tree canopy would have been proud of. Let us say that Otranto beach is about 1 kilometre long and 50 metres wide. A lazy sunbather, lying on a towel under a flowery parasol, covers no more than a couple of square metres of sand. Those that gambol in the surf or impress with beach volleyball may cover more like 25 square metres. With around 4 metres depth of sand, and with your average grain of sand being a tiny 0.7 mm3– allowing for gaps – there's about 1011 grains per active bather. By coincidence, that is approximately the number of neurons packed – much more tightly – into any one sunbather's brain. Of course, there is a big, big difference between brain and beach. Whereas one grain of sand touches only a few others, each neuron in a human brain can count as many as 7000 neighbours: around 7000 tangled synaptic tendrils reach out from each neuron. Our beach is simply 1000 times less connected than a human brain. Moreover, a grain of sand and a neuron are very different: the physical states that a grain of sand can be in are very few (orientation together with wet and dry spring to mind), whereas a neuron in the brain has a whole sigmoid function of possibilities. Humans think in the sun, sand just bakes. Although each grain of sand is a tiny chip of silicon, as a silicon chip it computes a big round nothing, zilch, nada; touching only its nearest neighbours does not make even the slightest hum, never mind turning on the lights. Making the right connections is how difficult, precisely? Dummer first conceived of connecting silicon with the ‘integrated circuit’ in 1953, and since then larger and larger scale connection has been the goal of IT; and, although Joseph ‘Lick’ Licklider's 1962 fantastically named ‘Intergalactic Computer Network’ contained most of the ideas that later became the Internet, we had to wait for Tim Berners-Lee's invention of HTML to teach us how even a small sandpit-worth of computers – the 100 million or more processors that exist today – could communicate with each other. Expert-systems-wise, networks of neurons have been the focus of simulation for over 25 years, starting with John Joseph Hopfield's artificial neural networks (ANNs) in 1982. From that basis, and with many technological revisions, papers in Expert Systems currently present excellent results with ANNs based on a few hundred neurons. But it is still a long way from getting a ‘beachful of silicon’ thinking. Not because it is the most polished or complete but because it is focused on leveraging artificial neural nets through the Internet, the Artificial Intelligence System project of Intelligence Realm Inc. intends to create – at least a small – thinking beach. Based on the Berkeley Open Infrastructure for Network Computing (BOINC) – the same technology as the basis of the much better known SETI1 project – the aim of the Artificial Intelligence System is to simulate the human brain, in real time. If their current experiments pay off, and it is early days yet, they will next cross the five-sunbather threshold: 500 billion neurons-worth of brain power. You can find out more about the project at http://www.intelligencerealm.com/aisystem. There are six excellent papers in this issue covering areas as diverse as rock engineering and dairy science. After Guest Editor Elif Übeyli's special issue last month on Advances in Medical Decision Support Systems, the first paper in this issue continues her theme of medical decision using expert systems with ‘Probabilistic neural networks combined with wavelet coefficients for analysis of electroencephalogram signals’ in which probabilistic neural networks classify electroencephalogram (EEG) signals. The result is that the wavelet coefficients obtained accurately characterize EEG signals, with a total classification accuracy of 97.63%. In ‘Automatic tuning of L2-SVM parameters employing the extended Kalman filter’ Mu and Nandi validate the method of using extended Kalman filters to tune a 2-norm support vector machine (the eponymous L2-SVMs), leading to higher classification accuracies, faster training speeds and reduced sensitivity to initial conditions. Gokceoglu, Sonmez and Zorlu report on the improvement over traditional simple and multivariate regression techniques traditionally employed in rock engineering given by fuzzy inference systems and artificial neural networks. They build predictive models of clay-bearing rocks. Moving on to dairy science, Magdalena, Fernández, Martín, Soria, Martínez, Navarro and Mata in their paper ‘Qualitative analysis of goat and sheep production data using self-organizing maps’ analyse the relationship between different small ruminant livestock production systems with different levels of specialization using self-organizing maps. New links between more specialized farms that have greater milk income per goat are reported that have guided improvements in farm performance. In ‘An automated hierarchical gait pattern identification tool employing cross-correlation-based feature extraction and recurrent neural network based classification’, Dutta, Chatterjee and Munshi use Elman's recurrent neural network for automatic identification of healthy and pathological gaits, and for the diagnosis of neurological disorders. The results allow the segregation of healthy, Parkinson's disease, Huntington's disease and amyotrophic lateral sclerosis subjects. Finally, Furst and Trichet, in their paper ‘Axiom-based ontology matching’, give a new approach to ontology matching covering overlapping domains.

… hours of processing time to sustain seconds of self-awareness. (True Names, Vernor Vinge) There's been some chatter in the popular IT press recently about the ‘Singularity’. Sounds like science fiction? Something out of a Star Trek movie, perhaps? ‘Entering the Singularity now, captain; phasers on stun.’ Well, it is science fiction, but it's not that sort of singularity: there are no wormholes in sight. No, it's much more like The Matrix plot or the future context of The Terminator: superintelligent machines, created by humans, taking over the world. Artificial-intelligence-wise, it's a relatively old concept, first written about by Vernor Vinge in his 1981 novella True Names:1 in 1981 it was unthinkable, apparently, that a computational intelligence could sustain self-awareness in real time. The postulate behind the Singularity is that there is a point in the future when in silica intelligence has higher cognitive power than in corpo intelligence. From that point on, it's all downhill for humans: the new society based on computational intelligence becomes too complex to understand. We quickly cease to be the dominant form of ‘life’ and, well, a while later a digital David Attenborough prompts adoring reviews from a digital press for its description of the human, i.e. lower, form of life. Unless there's a mad genius out there – and in the readership there are perhaps many who fit the bill – able to thrust a computational intelligence able to dominate the world fully formed and ticking into it, such a takeover is either going to be (a) an emergent property or (b) criminally irresponsible knowledge engineering. Asimov had it right in his examination of three laws of robotics – scratch a superintelligence with the goal of protecting humans and you find a domineering parent who takes all dangerous toys away. Before the Singularity, however, there is a major role for responsible knowledge engineering. The papers published in this journal describe the forward steps you have made to, for instance, assist professionals in their duties, improve engineering, make prosthetics better, make society a better place to live in and, well, simply make a better life for many people. Long before any hard, controlling, singular computational intelligence, there is the knowledge engineering community – soft, beneficent, singular intelligences – self-aware in real time. This month's papers are typical of these themes. Übeyli continues her investigation of electrocardiogram signal diagnostics, this time with support vector machines using two types of signal from the Physiobank database. The study shows that support vector machines trained on electrocardiogram power spectral densities obtained by eigenvector methods lead to accurate classifications. With uncanny timing, Kim, Lee, Oh and Kim present ‘An early warning system for financial crisis using a stock market instability index’, which measures the differences between current and past market conditions. The authors study comparative time series modelled by an asymptotic stationary autoregressive model using artificial neural networks. To see whether they say ‘I told you so’ you'll have to read the paper. Our third paper this month is a ‘Hybrid expert–fuzzy approach for evaluation of complex systems’ by Öztürk. The complex systems of the title include complex engineered systems, such as an air defence system, systems that are typically complicated and time consuming to evaluate. The author describes a hybrid expert–fuzzy system, called INES, for evaluation of complex systems which was developed for speeding up the evaluation, thus saving time, reducing cost and providing reusability. The fourth paper is ‘Fuzzy sliding mode control of a finger of a humanoid robot hand’ by Arslan, Hacioglu and Yagiz in which a fuzzy sliding mode controller is applied to a model of a human finger in order to tune the control gain of the sliding mode controller leading to an efficient adaptive controller.

Submit an article entitled thus to your publishing editor and they will raise the query: Give the eponymous sum to a schoolchild and, provided that he/she has learned to add (or has learned to use a calculator), he or she will also tell you the answer is 105721. Alan Turing was unabashed, in a 1950 article1 in Mind, to put the incorrect answer into the mouth of a fictional player of the ‘imitation game’. The article was entitled ‘Computing, machinery and intelligence’. In it Turing asked ‘Can a machine think?’ The game came to be known as the Turing test, through which a congruence of machine and human intelligence could be established. Turing's suggested route to intelligent machines was to mimic the machinery of a child's brain, and educate the resulting machine. That suggestion continues to inspire researchers – and robots.2 And now there's a new kid on the block! WolframAlpha is described – carefully, without mentioning intelligence – as a ‘computational knowledge engine’. Its goal is to3 build on the achievements of science and other systematizations of knowledge to provide a single source that can be relied on by everyone for definitive answers to factual queries. WolframAlpha is said to contain more than 10 trillion pieces of data, more than 50,000 algorithms and models, and linguistic capabilities from more than 1000 domains. It's great fun to play with. One commentator has suggested that WolframAlpha is possibly an ‘emerging artificial intelligence and a step towards a self-organising internet’. That's most unlikely, at least if we use the Turing test as the criterion for judging the emergent artificial intelligence. But it's only a failure if the aim is that AI should compete with HI – human intelligence. It appears very much the mission of this journal that the aim is to complement and complete HI. It is with great pleasure that we welcome to the board two new members: Professors Adrian Hopgood (De Montfort University, UK) and John Fox (University of Oxford and University College London, UK). Professor Adrian Hopgood is Dean of the Faculty of Technology at De Monfort University, Leicester, UK. His research interest concerns artificial intelligence, including knowledge-based and distributed multi-agent systems, computational intelligence (artificial neural networks, genetic algorithms, fuzzy logic) and their practical applications. He has a particular passion for hybrids, which bring together the best techniques for particular applications. As well as a PhD, Adrian has a Diploma in French and an MBA from the Open University. He is also author of the best selling Intelligent Systems for Engineers and Scientists. Professor John Fox attended Durham (UK) and Cambridge Universities (UK) and held postdoctoral fellowships at Carnegie-Mellon and Cornell Universities in the USA. After returning to the UK in 1975 he worked on decision making and artificial intelligence in medicine, joining Cancer Research UK (then ICRF) in 1981 to set up an interdisciplinary group in artificial intelligence, computer science and medicine. John's interests include cognitive science, computing and biomedical engineering. A recent book Safe and Sound: Artificial Intelligence in Hazardous Applications deals with the use of artificial intelligence in medicine and other safety-critical fields. He was founding editor of the Knowledge Engineering Review. John is now at Oxford where he has set up a new collaboration in cognitive science and systems engineering (http://www.cossac.org). Please welcome both Adrian and John to the board. This month we have four excellent papers. In ‘Fuzzy based fast dynamic programming solution of unit commitment with ramp constraints’, Patra et al. give insight into the use of fuzzy logic in the solution to difficult-to-solve multi-stage decision-making problems, showing that fuzzy models can perform as well on these problems as more traditional techniques. The difficulty of obtaining realistic simulated real-world data suitable for the validation of models used in knowledge engineering of medical solutions is well known. In ‘Preliminary evaluation of electroencephalographic entrainment using thalamocortical modelling’, Cvetkovic et al. consider a number of theoretical models for their ability to replace actual data. Importantly, the authors identify a number of shortcomings that should be overcome for advances in this area of knowledge engineering to be made. In ‘Modified mixture of experts employing eigenvector methods and Lyapunov exponents for analysis of electroencephalogram signals’, Übeyli presents another in her excellent series of papers dedicated to detecting variability of electroencephalogram signals, this time with a classification accuracy of 98.33%. Continuing the theme of the second paper, if not the approach, Yang and Kecman consider the use of an adaptive local hyperplane algorithm as a suitable classifier when only a small data set is available to learn from. The proposed classifier outperforms, on average, all the other four benchmarking classifiers in this area. Enjoy!

Eugene M. Izhikevich's stimulating paper on polychronization (Izhikevich, 2006) describes a number of neuro-computational properties observed in simulations of neural networks based on his computationally tractable model of the neuron (Izhikevich, 2003). ‘Polychronization’ is coined by Izhikevich to refer to the reproducible time-locked (but not synchronous) firing patterns that occur in networks of ‘spiking cortical neurons’. Imagine a juggler juggling three balls. The release of a ball as it rises from the juggler's hand is neither synchronized with the release of the other two nor random; throws must simply occur at precisely the right time relative to each other. Polychronization is like neurons juggling: it is the effect of unsynchronized neuron firings that cascade because they are timed correctly with respect to each other. Santiago Ramón y Cajal was the first to identify the neuron as the ‘primary functional unit’ of the human nervous system, an insight that created the modern discipline of neuroscience and for which he won the 1906 Nobel prize.1 With Turing unborn, and the computer many years' distant, the possibility of modelling the neuron's function lay far into the future. Today it is very different: with a multitude of excellent simulation tools, complex simulations are possible, some might even say easy. Now it's perfecting the detail of the model that's the challenge. Enter Izhikevich with his neuron model – its computationally tractable firing behaviour allows simulations of tens of thousands of neurons with millions of synaptic connections in real time on an ordinary PC. Moreover, by altering his model's parameters, there arise neural firing patterns that closely resemble those observed in the rat's motor cortex. And that is where we come in: they are also familiar in closely resembling the different ways that the products of engineered manufacturing processes sustain a modern economy. Rogers' definition of engineering (see Editorial 26 (1)) leads, quite naturally as it happens, to the representation of an engineering problem as a triple (Hall & Rapanotti, 2008) consisting of an environment, a solution and a need. In combination, the environment's elements are used to engineer a solution to meet a need. As illustration, consider a world that needs, say, potato peelers: in an environment in which suitable raw materials, skills and finance exist a solution can be engineered that produces potato peelers. If the environment, say, lacks the ability to sharpen the potato peeler blade – and that ability isn't part of the engineered solution – then any need for potato peelers may go unfulfilled. Otherwise, the environment is adequate, and solutions, such as potato peelers or, more generally, products, services or whatever, appear in the environment of other problems – unsynchronized but arriving at the right time – to drive their solutions. In seeing it this way, perhaps there is an inkling of a suspicion of an analogy between an engineering problem and a neuron that can link Izhikevich's spiking neuron model to a polychronic economy. Can an economy be said to be a polychronic brain of sorts with engineered solution processes playing the role of neurons? Does the economy have brain-like activity? Does it rest, work and play? Does it become exuberant? And is the global crash simply a symptom that the economy is currently suffering from a depressive disorder? To fix the analogy in science is a matter of playing with Izhikevich's model to loosen – or, better, scale – the timescales over which it reacts and to accommodate the changing overlaps between problem solvers in the global economy. The benefit? A better understanding of the economy could bring better models for economic management. Perhaps a ‘happier, more satisfied’ economy can be built using tools for the analysis and synthesis of polychrony. And, even if they already contribute greatly, knowledge engineering and expert systems would have much more to contribute. It's a no brainer, really. It is a great pleasure to welcome a new board member: Dr Desheng Dash Wu has joined the board as an active associate editor. Dash is a tenured assistant professor at Reykjavik University, an affiliated professor at RiskLab and director of the RiskChina Research Center, both at the University of Toronto. His research interests focus on enterprise risk management, performance evaluation and business decision support. His work has appeared in journals including the International Journal of Production Research, the IEEE Transactions on Knowledge and Data Engineering, the Journal of the Operational Research Society, the European Journal of Operational Research, the Annals of Operations Research, the International Journal of Production Economics and the International Journal of System Science, to name but a few. He has two books to his name, has served as editor and guest editor for several journals, and has chaired many conferences. He is also a member of the Professional Risk Managers' International Association's Academic Advisory Committee and is a steering committee member. Dash is currently preparing a special issue for Expert Systems, and his work as a guest editor on this issue has identified him as someone who will contribute much as an associate editor. We have a bumper crop of six papers this month. In ‘Extracting new patterns for cardiovascular disease prognosis’, Mena, Gonzalez and Maestre identify abnormal blood pressure variability as a cardiovascular risk factor and assess its ability to classify sick people correctly with other, more traditional algorithms. In ‘Colon segmentation and colonic polyp detection using cellular neural networks and three-dimensional template matching’ by Kilic, Ucan and Osman the authors describe an improved detection system using three heuristic optimization techniques. The third paper by Kong, Xu, Liu and Yang, ‘Applying a belief rule-base inference methodology to a guideline-based clinical decision support system’, contributes to the debate of how one arrives at an accurate clinical decision in the light of uncertain medical domain knowledge and clinical symptoms. Through a case study, the paper applies the generic rule-base inference methodology using evidential reasoning (RIMER) approach to model clinical guidelines and clinical inference processes in a clinical decision support system. Moving on to the use of ontologies within service-oriented architectures (or SOAs), Goodwin and Russomanno, in their paper ‘Ontology integration within a service-oriented architecture for expert system applications using sensor networks’, describe how ontology-enhanced search can better match discovered services to requests. Back to clinical uses of expert systems with work on the classification of Pap smear cells by Marinakis, Marinaki, Dounias, Jantzen and Bjerregaard in their paper ‘Intelligent and nature inspired optimization methods in medicine: the Pap smear cell classification problem’. Tested on two large data sets, the methods used include tabu search, genetic algorithms and particle swarm and ant colony optimizations. The last paper of Volume 26 is entitled ‘Unlabelled extra data do not always mean extra performance for semi-supervised fault prediction’. In it, Catal and Diri investigate and benchmark a number of high performance classifiers for software fault prediction with limited fault data. With each employed within YATSI (Yet Another Two Stage Idea), a meta algorithm, the authors show that the YATSI versions may improve on naive Bayes for large data sets.

The design of assurance cases has been hampered by the posit-and-prove approach to software and systems engineering; it has been observed that, traditionally, a product is produced and then evidence from the development is looked for to build an assurance case. Although assured development is possible, there has been a recent drive towards the proactive design of the assurance case. Assurance-driven design is a development method that considers the assurance argument - similar to an assurance case - to be as important as the product. In this paper, we introduce assurance-driven design, as part of our problem oriented engineering framework, and show how software and assurance argument can be built together.

We are very pleased to be associated with the four papers on problem frames (Jackson, 2001; Cox et al., 2005) that appear in this special issue of Expert Systems, The Journal of Knowledge Engineering. In its 25 years of publication, Expert Systems has seen many changes in the landscape of expert systems. The technologies have matured from their early forms to become robust and the basis of many fine systems. The areas in which they have been applied have expanded and many areas of modern life now contain them. The design of solutions based on knowledge engineering technologies continues apace in many areas of modern relevance: recent volumes of Expert Systems have seen applications to financial market volatility (Oh et al., 2006), heart beat instability (Ubeyli, 2007), water quality improvement (Hatzikos et al., 2007) and even adolescent well-being (Wang et al., 2007). The application of knowledge engineering techniques depends to a large extent on the skills of software engineers; and few would argue against a view of software engineering as overlapping knowledge engineering. Software requirements engineering, for instance, has learned much from reflection on core knowledge engineering techniques of elicitation, representation and classification. The flow is both ways, of course, and the reflective knowledge engineering community is reappraising the work on expert systems in the light of what software engineering has achieved; Sagheb-Tehrani (2006), for instance, provides an investigation of the processes of design of expert systems to determine why many of them have been unsuccessful at their primary task. The critique points to the work of the designer, or knowledge engineer, coming increasingly under scrutiny as technological achievement needs converting into products and services, with notions of fitness-for-purpose becoming of heightened importance in what is now a large industry. The role of designer as knowledge engineer is broad both in scope and application. In terms of scope, we distinguish the designer as radical inventor at one end of the spectrum from normal designer at the other: both are creative, both work iteratively, both work as knowledge engineers, both have a view of the product or service to be provided, but it is techniques and support for those towards the normal design end–those whose role requires an ability to build a product or service that can be assured, in all probability, to satisfy its design goals first time–that can be the basis of a sustainable engineering discipline. Normal design was identified by Vincenti (1990), and is likely to play an increasingly important role in the theory of design, including that of expert systems. One might ask how normal design is practised. One way is through the use of ‘patterns’ (see, for instance, Moynihan et al., 2006), with a pattern describing successful recurring solutions to common problems. Problem frames are a more recent and successful addition to the pattern literature, and are designed to assist in the software requirements phase of software intensive system development by capturing recurring ‘problems’ in software; this theme is continued in Jackson's Foreword to this issue (Jackson, 2008). One important aspect of the normal practice of capturing recurrent problems, rather than solutions, in traditional engineering disciplines is that they lead to practical, rather than technological, specialists, and that is another powerful thesis of Jackson. We hope that you enjoy reading the papers in this issue and any subsequent reflection on your own discipline. The four papers in this volume illustrate the broad applicability of problem frames. Strunk and Knight, in their paper ‘The essential synthesis of problem frames and assurance cases’, tackle the difficult issue of building an assurance case, i.e. a collection of evidence and argument in defence of a system's fitness for purpose, during early problem analysis. A convincing assurance case is the basis of a decision of whether to commission a system and, ultimately, its value to the customer. Lavazza and del Bianco, in ‘Enhancing problem frames with scenarios and histories in UML-based software development’, show how effective and beneficial it is to consider not only the structure and behaviour of a problem's context, but also its typical scenarios and past history. They find that scenarios offer good support for the merging of subproblems, for addressing frame concern issues and for design models, a result of obvious potential importance in problem frames and other knowledge engineering areas. Jørgenson, in ‘Coloured Petri nets and graphical animation: a proposal for a means to address problem frame concerns’, demonstrates how some very polished Petri net tools can be employed to make the understanding of a problem, expressed in the problem frames' notation, easier to a stakeholder. One exciting element of this work is that it includes not only the behaviour of the system but also the frame concerns, i.e. those arguments that show that early problem analysis has been successful. Lencastre, Araujo, Moreira and Castro, in ‘Towards aspectual problem frames: an example’, show how another modern development technology, aspect orientation, works to provide a more modular, evolvable approach to problem framing. Begun in 2004, the IWAAPF workshop series has encouraged those with an interest in requirements engineering to come together, biannually, to discuss their areas of expertise. The 2006 IWAAPF workshop, at which the original versions of these papers were presented, was successful due to the hard work of everyone involved. Many of the original reviewers helped in the preparation of this special issue, and we thank them greatly for their help. Thanks also to those in Wiley-Blackwell's Expert Systems' office who have made this special issue available in good time. Last, but by no means least, to Michael Jackson go our thanks for presenting such a stimulating avenue of research.

Problem Orientation is gaining interest as a way of approaching the development of software intensive systems and yet a significant example that explores its use is missing from the literature. In this paper, we present the basic elements of Problem Oriented Software Engineering (POSE) which aims to bring both non-formal and formal aspects of software development together in a single framework. We provide an example of a detailed and systematic POSE development of a software problem, that of designing the controller for a package router. The problem is drawn from the literature, but the analysis presented here is new. The aim of the example is twofold: to illustrate the main aspects of POSE and how it supports software engineering design, and to demonstrate how a non-trivial problem can be dealt with by the approach.

A key challenge for software engineering is to learn how to reconcile the formal world of the machine and its software with the non-formal real world. In this paper, we discuss elements of problem oriented software engineering (POSE), an approach that brings both non- formal and formal aspects of software development together in a single theoretical framework for software engineering design. POSE presents development as the representation and step-wise transformation of software problems. It allows for the identification and clarification of system requirements, the understanding and structuring of the problem world, the structuring and specification of a hardware/software machine that can ensure satisfaction of the requirements in the problem world, and the construction of adequacy arguments, convincing both to developers and to customers, users and other interested parties, that the system will provide what is needed. Examples are used throughout the paper to illustrate how formal and non-formal descriptions are reconciled under POSE.

Software problems originate from real world problems. A software solution must address its real world problem in a satisfactory way. A software engineer must therefore understand the real world problem that their software intends to address. To be able to do this, the software engineer must understand the problem context and how it is to be affected by the proposed software, expressed as the requirements. Without this knowledge the engineer can only hope to chance upon the right solution for the problem. Application of the Problem Frames approach may well be a way of meeting this need.

Understanding the roles that rigour and formality can have in the design of critical systems is critical to anyone wishing to contribute to their development. Whereas knowledge of these issues is good in software development, in the use of hardware -- specifically programmable logic devices (PLDs) and the combination of PLDs and software -- the issues are less well known. Indeed, even in industry there are many differences between current and recommended practice and engineering opinion differs on how to apply existing standards. This situation has led to gaps in the formal and rigorous treatment of PLDs in critical systems.In this paper we examine the range of and potential for formal specification and analysis techniques that address the requirements for verifiable PLD programs. We identify existing formalisms that may be used, and lay out the areas of contributions that academia and industry in collaboration can make that would allow high-integrity PLD programming to be as practicable as high-integrity software development.This paper also touches briefly on some important practical, technical, organisational, social, and psychological aspects of the introduction of formal methods into industrial practice for hardware and system design. It also provides an update and summary of the recent UK Defence Standard 00-56, as it relates to hardware.

FPGAs are being used in increasingly more complex roles in critical systems, interacting with conventional critical software. Established safety standards require rigorous justification of safety and correctness of the conventional software in such systems. Newer standards now make similar requirements for safety-related electronic hardware, such as FPGAs, in these systems. In this paper we examine the current state-of-the-art in programming FP-GAs, and their use in conventional (low-criticality) hardware/software systems. We discuss the impact that the safety standards requirements have on the co-development of hardware/software combinations. and suggest adaptations of existing best practice in software development that could discharge them. We pay particular attention to the development and analysis of high-level language programs for FPGAs designed to interact with conventional software.

The reference model of Gunter et al, 2000, provides a framework for describing and analyzing key software engineering artifacts and their properties. We propose a reification of this framework in which behaviour is explicitly trace-based. We find that this benefits the formalism in adding structure in ways which are meaningful and practical from an engineering viewpoint. In particular, we develop notions of points of introduction and reachability in the new framework, and show how they strengthen the properties of the reference model.

Problem frames provide a means of analyzing and decomposing problems. They emphasise the world outside of the computer, helping the developer to focus on the problem domain, instead of drifting into inventing solutions. However, even modestly complex problems can force us into detailed consideration of the architecture of the solution. This is counter to the intention of the problem frames approach, which is to delay consideration of the solution space until a good understanding of the problem is gained. We therefore extend problem frames, allowing architectural structures, services and artifacts to be considered as part of the problem domain. Through a case study, we show how this extension enhances the applicability of problem frames in permitting an architecture-based approach to software development. We conclude that, through our extension, the applicability of problem frames is extended to include domains with existing architectural support.