Abstract: Current desktop environments provide weak support for carrying out complex user-oriented tasks. Although individual applications are becoming increasingly sophisticated and feature-rich, users must map their high-level goals to the low-level operational vocabulary of applications, and deal with a myriad of routine tasks (such as keeping up with email, keeping calendars and web sites up-to-date, etc.). An alternative vision is that of a personal cognitive assistant. Like a good secretary, such an assistant would help users accomplish their high-level goals, coordinating the use of multiple applications, automatically handling routine tasks, and, most importantly, adapting to the individual needs of a user over time. In this paper we describe the architecture and its implementation for a personal cognitive assistant called RADAR. Key features include (a) extensibility through the use of a plug-in agent architecture (b) transparent integration with legacy applications and data of today's desktop environments, and (c) extensive use of learning so that the environment adapts to the individual user over time.

Abstract: Notations and techniques for architectural modeling and analysis have matured considerably over the past two decades. However, to date these approaches have primarily focused on architectural properties and behavior that can be precisely defined. In this paper we argue that it is possible to augment existing architecture description languages (ADLs) to support reasoning and analysis in the presence of uncertainty. Specifically, we outline two basic extensions to formal architecture descriptions that take advantage of probabilistic specifications to support architecture-based analyses such as simulation, detection of behavioral drift, and reasoning about the expected outcomes of uncertain behavior. An important property of these specifications is that they allow incremental refinement � as more is known about the behavior of the system, specifications can be extended without invalidating previous analyses.

Abstract: Autonomic, or self-adaptive, systems are increasingly important. One of the most prevalent techniques is to adopt a control systems view of the solution: adding a runtime, separate control unit that monitors and adapts the system under consideration. A problem with this paradigm for system engineering is that the control and the system are loosely coupled, introducing a variety of sources of uncertainty. In this paper we describe three specific sources of uncertainty, and briefly explain how we address those in the Rainbow Project.

Abstract: Dynamic adaptive systems based on multiple concurrent applications typically employ optimization models to decide how to allocate scarce resources among the applications and how to tune their runtime settings for optimal quality-of-service according to the preferences of an end user.

Traditionally, such systems have avoided dealing with uncertainty by assuming that current snapshots of the relevant inputs are precise and by solving for an optimal system point. To achieve dynamic behavior, a system performs an optimization loop upon discovering changes in the input variables (e.g. changes in the available level of resources) and adapts the applications according to the new optimal solution. Unfortunately, when certain adaptation actions incur costs, such reactive adaptation strategies suffer from a significant shortcoming: several locally optimal decisions over time may often be less than optimal globally.

By using predictive information about the future values of the problem inputs, we can model and implement an anticipatory adaptation strategy that helps improve the global behavior of the system in many situations. However, modeling predictions requires representing and dealing with uncertainty from different sources. In this paper, we describe our proposed approach to represent multiple sources of uncertainty and outline algorithms for solving the anticipatory configuration problem with predictive inputs.

Abstract: Little is known about how developers think about design during code modification tasks or how experienced developers’ design knowledge helps them work more effectively. We performed a lab study in which thirteen developers worked for 3 hours understanding the design of a 54 KLOC open source application. Participants had from 0 to 10.5 years of industry experience and were grouped into three “experts” and ten “novices.” We observed that participants spent their time seeking, learning, critiquing, explaining, proposing, and implementing facts about the code such as “getFoldLevel has effects”. These facts served numerous roles, such as suggesting changes, constraining changes, and predicting the amount of additional investigation necessary to make a change. Differences between experts and novices included that the experts explained the root cause of the design problem and made changes to address it, while novice changes addressed only the symptoms. Experts did not read more methods but also did not visit some methods novices wasted time understanding. Experts talked about code in terms of abstractions such as “caching” while novices more often described code statement by statement. Experts were able to implement a change faster than novices. Experts perceived problems novices did not and were able to explain facts novices could not. These findings have interesting implications for future tools.

Abstract: Self-adapting systems based on multiple concurrent applications must decide how to allocate scarce resources to applications and how to set the quality parameters of each application to best satisfy the user. Past work has made those decisions with analytic models that used current resource availability information: they react to recent changes in resource availability as they occur, rather than anticipating future availability. These reactive techniques may model each local decision optimally, but the accumulation of decisions over time nearly always becomes less than optimal.

 In this paper, we propose an approach to self adaptation, called anticipatory configuration that leverages

predictions of future resource availability to improve utility for the user over the duration of the task. The approach solves the following technical challenges: (1) how to express resource availability prediction, (2) how to combine prediction from multiple sources, and (3) how to leverage predictions continuously while improving utility to the user. Our experiments show that when certain adaptation operations are costly, anticipatory configuration provides better utility to the user than reactive configuration, while being comparable in resource demand.

Abstract: Over the past decade there has been considerable progress in the development of formal methods to improve our confidence in complex systems. Today the use of such methods in certain fields, such as hardware design, or nuclear power control systems, is de rigueur, with commensurate improvements in quality and reliability.

Regrettably, however, the use of formalism in the medical device domain is relatively sparse. This is due in large part to the perceived difficulty of using formal methods by ordinary engineers and domain specialists, and by the lack of training in how best to apply existing tools to solve the problems faced in that domain.

Over the past few years we have been developing educational materials to help bridge this gap. Specifically we have developed a course in formal modeling for practicing engineers. A core component of this effort is a set of exercises drawn from the medical device domain, which are used to

a) show how formal modeling can be used as an effective technique to improve quality and reliability of software-intensive systems

b) provide guidelines on selecting appropriate modeling approaches for the problem at hand

c) give students hands-on experience in modeling and tool-assisted analysis

In this paper we outline our use of medical device challenge problems in achieving these goals. We argue that such exercises (and the underlying concepts) can go a long way towards bridging the gap between theory and practice, and could be used more generally to improve the state of the practice in developing high-confidence systems, in general, and medical devices, in particular.

Abstract: In a ubiquitous computing environment augmented reality would be an ideal choice for a display for the user. An augmented reality display assists users by adding computer generated information to their perception of reality, thus making it ideal for ubiquitous computing. Unfortunately augmented reality is technically difficult and costly to implement even when the application is designed for its use from the ground up. However, many of the necessary devices for a low fidelity implementation of augmented reality are readily available in a ubiquitous computing environment. Our research focuses on using Aura, a ubiquitous computing framework, to marshal the available devices in the environment. These devices can then be connected within the framework to provide an augmented reality display to the user at the best fidelity possible, given the available resources in a user's environment.

Abstract: Frameworks are a valuable way to share designs and implementations on a large scale. Client programmers, however, have difficulty using frameworks. They find it difficult to understand non-local client-framework interactions, design solutions when they do not own the architectural skeleton, gain confidence that they have engaged with the framework correctly, represent their successful engagement with the framework in a way that can be shared with others, ensure their design intent is expressed in their source code, and connect with external files. A design fragment is a specification of how a client program can use framework resources to accomplish a goal. From the framework, it identifies the minimal set of classes, interfaces, and methods that should be employed. For the client program, it specifies the client-framework interactions that must be implemented. The structure of the client program is specified as roles, where the roles can be filled by an actual client program¡Çs classes, fields, and methods. A design fragment exists separately from client programs, and can be bound to the client program via annotations in their source code. These annotations express design intent; specifically, that it is the intention of the client programs to interact with the framework as specified by the design fragment. The thesis of this dissertation is: We can provide pragmatic help for programmers to use frameworks by providing a form of specification, called a design fragment, to describe how a client program can correctly employ a framework and by providing tools to assure conformance between the client program and the design fragments. We built tools into an IDE to demonstrate how design fragments could alleviate the difficulties experienced by client programmers. We performed two case studies on commercial Java frameworks, using demo client programs from the framework authors, and client programs we found on the internet. The first case study, on the Applet framework, yielded a complete catalog of twelve design fragments based on our analysis of fifty-six Applets. The second case study, on the larger Eclipse framework, yielded a partial catalog of fourteen design fragments based on our analysis of more than fifty client programs. This work provides three primary contributions to software engineering. First, it provides a new technique to help programmers use frameworks. Second, it provides a systematic way to increase code quality. Design fragments provide a means to communicate known-good designs to programmers, and, unlike simple copying of examples, a means of influencing the uses of that design so that revisions can be propagated. Third, it provides an empirically-based understanding of how clients use frameworks, which aids researchers in choosing research directions and aids framework authors in delivery of new frameworks.