Sandeep Purao

Product metrics for object-oriented systems

We survey metrics proposed for object-oriented systems, focusing on product metrics. The survey is intended for the purposes of understanding, classifying, and analyzing ongoing research in object-oriented metrics. The survey applies fundamental measurement theory to artifacts created by development activities. We develop a mathematical formalism that captures this perspective clearly, giving appropriate attention to the peculiarities of the object-oriented system developmenr process. Consistent representation of the available metrics, following this mathematical formalism, shows that current research in this area contains varying coverage of different products and their properties at different development stages. The consistent representation also facilitates several analyses including aggregation across metrics, usage across metrics, equivalent formulation of metrics by multiple researchers, and exploitation of traditional metrics for object-oriented metrics. We also trace the chronological development of research in this area, and uncover gaps that suggest opportunities for future research.

Improving Analysis Pattern Reuse in Conceptual Design: Augmenting Automated Processes with Supervised Learning

Conceptual design is an important, but difficult, phase of systems development. Analysis patterns can greatly benefit this phase because they capture abstractions of situations that occur frequently in conceptual modeling. NaA¯ve approaches to automate conceptual design with reuse of analysis patterns have had limited success because they do not emulate the learning that occurs over time. This research develops learning mechanisms for improving analysis pattern reuse in conceptual design. The learning mechanisms employ supervised learning techniques to support the generic reuse tasks of retrieval, adaptation, and integration, and emulate expert behaviors of analogy making and designing by assembly. They are added to a naA¯ve approach and the augmented methodology implemented as an intelligent assistant to a designer for generating an initial conceptual design that a developer may refine. To assess the potential of the methodology to benefit practice, empirical testing is carried out on multiple domains and tasks of different sizes. The results suggest that the methodology has the potential to benefit practice.

Educating reflective systems developers

Abstract. Systems development research shows that practitioners seldom follow methods and that the competencies required for successful development of computer‐based systems go well beyond those represented in contemporary methods. These insights make us question the role that methods should play in educating would‐be developers. Pedagogical theories, such as situated learning and double‐loop learning, complement these insights. Integrating the two, we argue that students need to complement the simplified accounts that methods express, with reflections on methods‐in‐use and on development practice in general. We present operationalizations of this idea in two quite different academic settings. Based on a retrospective analysis of our experiences in these settings, and a comparison and evaluation of the two approaches, we propose a number of lessons that can be used to improve the education of would‐be developers.

Situating evaluation in scenarios of use

We report on the use of scenario-based methods for evaluating collaborative systems. We describe the method, the case study where it was applied, and provide results of its efficacy in the field. The results suggest that scenario-based evaluation is effective in helping to focus evaluation efforts and in identifying the range of technical, human, organizational and other contextual factors that impact system success. The method also helps identify specific actions, for example, prescriptions for design to enhance system effectiveness. However, we found the method somewhat less useful for identifying the measurable benefits gained from a CSCW implementation, which was one of our primary goals. We discuss challenges faced applying the technique, suggest recommendations for future research, and point to implications for practice.

A theoretical investigation of the emerging standards for web services

Currently, standards for web services are being developed via three different initiatives (W3C, Semantic web services and ebXML). To the best of our knowledge, no theoretical perspectives underlie these standardization efforts. Without the benefit of a strong theoretical basis, the results, within and across these initiatives, have remained piecemeal. We suggest ‘Language–Action Theories’ as a plausible perspective that can effectively define, assess and refine web services standards. In this paper, we first investigate the existing initiatives to identify commonalities that point to theories of ‘Language–Action’ as an appropriate theoretical basis for web services standards. Next, we adapt work from these theories to develop a comprehensive reference framework for understanding web services standards. Finally, we use this reference framework to assess the three initiatives, and analyze the findings to provide insights for future development and refinement of web services standards.

CONQUER: A Methodology for Context-Aware Query Processing on the World Wide Web

A major impediment to accurate information retrieval from the World Wide Web is the inability of search engines to incorporate semantics in the search process. This research presents a methodology, CONQUER (CONtext-aware QUERy processing), that enhances the semantic content of Web queries using two complementary knowledge sources: lexicons and ontologies. The methodology constructs a semantic net using the original query as a seed, and refines the net with terms from the two knowledge sources. The enhanced query, represented by the refined semantic net, can be executed by search engines. This paper describes the methodology and its implementation in a prototype. An empirical evaluation shows that queries suggested by the prototype produce more relevant results than those obtained by the original queries. The research, thus, provides a successful demonstration of the use of existing knowledge sources to enhance the semantic content of Web queries. The paper concludes by identifying potential uses of such enhancements of search technology in organizational contexts.

A heuristic-based methodology for semantic augmentation of user queries on the web

As the World Wide Web continues to grow, so does the need for effective approaches to processing users’ queries that retrieve the most relevant information. Most search engines provide the user with many web pages, but at varying levels of relevancy. The Semantic Web has been proposed to retrieve and use more semantic information from the web. However, the capture and processing of semantic information is a difficult task because of the well-known problems that machines have with processing semantics. This research proposes a heuristic-based methodology for building context aware web queries. The methodology expands a user’s query to identify possible word senses and then makes the query more relevant by restricting it using relevant information from the WordNet lexicon and the DARPA DAML library of domain ontologies. The methodology is implemented in a prototype. Initial testing of the prototype and comparison to results obtained from Google show that this heuristic based approach to processing queries can provide more relevant results to users, especially when query terms are ambiguous and/or when the methodology’s heuristics are invoked.

The Sciences of Design: Observations on an Emerging Field

The boundaries and contours of design sciences continue to undergo definition and refinement. In many ways, the sciences of design defy disciplinary characterization. They demand multiple epistemologies, theoretical orientations (e.g. construction, analysis or intervention) and value considerations. As our understanding of this emerging field of study grows, we become aware that the sciences of design require a systemic perspective that spans disciplinary boundaries. The Doctoral Consortium at the Design Science Research Conference in Information Sciences and Technology (DESRIST) was an important milepost in their evolution. It provided a forum where students and leading researchers in the design sciences challenged one another to tackle topics and concerns that are similar across different disciplines. This paper reports on the consortium outcomes and insights from mentors who took part in it. We develop a set of observations to guide the evolution of the sciences of design. It is our intent that the observations will be beneficial, not only for IS researchers, but also for colleagues in allied disciplines who are already contributing to shaping the sciences of design.

Software Architectures: Blueprint, Literature, Language Or Decision?

This study questions the traditional view of software architecture as a specification that needs only be understood by software architects and engineers. Based on an intensive study of three software-producing organizations, we identify multiple metaphors (‘Blueprint,’ ‘Literature,’ ‘Language’ and ‘Decision’) that stakeholders use to understand the term software architecture, which in turn, allows them to effectively participate in its creation and use. Our results point to new research directions that may better encompass a broader view of software architecture.

Towards an understanding of the use of problem and design spaces during object- oriented system development

The importance of, and distinction between, problem and design spaces has been recognized in studies of information systems development (ISD). With increasing acceptance of objectoriented techniques, which promise close mirroring of real-world concepts in the IS artifacts, this distinction becomes even more important. In spite of a rich literature stream dealing with the general notion of ‘design,’ several inadequacies remain in our understanding of the ISD processes—one of these is the recognition and impact of problem and design spaces. In this paper, we analyze processes followed by two developers engaged in a non-trivial development task using the object-oriented modeling techniques—with a view to understanding their use of, and explorations in, problem and design spaces. Our analysis provides evidence for a distinction between the two spaces, and interprets the developers’ engagements and behaviors to structure the two spaces with the help of object-oriented modeling techniques. Several interesting findings emerge from our analysis, including the overlapping of spaces due to the use of object-orientation, disturbing patterns such as design fixation, interplay between simulation, expansion and validation in the design space, and the varying use of modeling techniques to structure the two spaces. Our analysis, supported by prior literature, provides a grounded description of some phenomena that have, hitherto, had only intuitive or prescriptive support. Based on these, we argue for more methodological and tool support for explorations of problem