We present a new technique, failure-oblivious computing, that enables servers to execute through memory errors without memory corruption. Our safe compiler for C inserts checks that dynamically detect invalid memory accesses. Instead of terminating or throwing an exception, the generated code simply discards invalid writes and manufactures values to return for invalid reads, enabling the server to continue its normal execution path. We have applied failure-oblivious computing to a set of widely-used servers from the Linux-based opensource computing environment. Our results show that our techniques 1) make these servers invulnerable to known security attacks that exploit memory errors, and 2) enable the servers to continue to operate successfully to service legitimate requests and satisfy the needs of their users even after attacks trigger their memory errors. We observed several reasons for this successful continued execution. When the memory errors occur in irrelevant computations, failure-oblivious computing enables the server to execute through the memory errors to continue on to execute the relevant computation. Even when the memory errors occur in relevant computations, failure-oblivious computing converts requests that trigger unanticipated and dangerous execution paths into anticipated invalid inputs, which the error-handling logic in the server rejects. Because servers tend to have small error propagation distances (localized errors in the computation for one request tend to have little or no effect on the computations for subsequent requests), redirecting reads that would otherwise cause addressing errors and discarding writes that would otherwise corrupt critical data structures (such as the call stack) localizes the effect of the memory errors, prevents addressing exceptions from terminating the computation, and enables the server to continue on to successfully process subsequent requests. The overall result is a substantial extension of the range of requests that the server can successfully process. 1

Interactive tabletops have been previously proposed and studied in the domain of co-located group applications. However, little fundamental research has been done to explore the issue of size. In this paper we identify a number of size considerations for tabletop design, and present an experiment to explore some of these issues, in particular the effects of group size and table size on the speed at which the task was performed, the distribution of work among group members, issues of shared resources, and user preference for table size. Our findings shed light on (1) how work strategies are affected by group size, (2) how social interaction varies with respect to table size, and (3) how the speed of task performance is influenced by group size but not by table size. In addition, our experiments revealed that for larger groups, designers might need to add additional vertical displays for shared information. This finding opens the door for extending singledisplay groupware to shared-display groupware settings that involve multiple, shared displays.

accuracy issues more directly. Moreover, this research attracts a great deal of attention because of what many take to be its dismal implications for the accuracy of human social reasoning. These implications are illusory, however, because an error is not the same thing as a "mistake. " An error is a judgment of an experimental stimulus that departs from a model of the judgment process. If this model is normative, then the error can be said to represent an incorrect judgment. A mistake, by contrast, is an incorrect judgment of a real-world stimulus and therefore more difficult to determine. Although errors can be highly informative about the process of judgment in general, they are not necessarily relevant to the content or accuracy of particular judgments, because errors in a laboratory may not be mistakes with respect to a broader, more realistic frame of reference and the processes that produce such errors might lead to correct decisions and adaptive outcomes in real life. Several examples are described in this article. Accuracy issues cannot be addressed by research that concentrates on demonstrating error in relation to artificial stimuli, but only by research that uses external, realistic criteria for accuracy. These criteria might include the degree to which judgments agree with each other and yield valid predictions of behavior. The accuracy of human social judgment is a topic of obvious

Visual search is the subject of countless psychology studies in which people search for target items within a scene. The bulk of this literature focuses on the individual with the goal of understanding the human perceptual system. In life, visual search is performed not only by individuals, but also by groups – a team of doctors may study an x-ray and a team of analysts may study a satellite photograph. In this paper, we examine the issues one should consider when searching as a group. We present the details of an experiment designed to investigate the impact of group size on visual search performance, and how different display configurations affected that performance. We asked individuals, pairs, and groups of four people to participate in a baggage screening task in which these teams searched simulated x-rays for prohibited items. Teams conducted these searches on single monitors, a row of four monitors, and on a single horizontal display. Our findings suggest that groups commit far fewer errors in visual search tasks, although they may perform slower than individuals under certain conditions. The interaction between group size and display configuration turned out to be an important factor as well.

also inspired by discussions about current work on related topics, with Marianne Bertrand, Eldar Shafir,

A number of studies have shown that students are often more willing to participate in educational conversations online than in the classroom. However, other studies have shown that online environments have poor student participation Why is this the case? What causes participation to vary from one environment to another? To explore these phenomena, we borrow a concept from social psychology, the bystander effect, which explains why individuals are less likely to helpinanemergencyifothersarepresent. Although the bystander effect specifically applies to helping behavior in emergency situations, we use this construct as a lens through which to view nonemergency situations such as educational environments. The bystander effect has 4 key components: self-awareness, social cues, blocking mechanisms, and diffuse responsibility. Focusing on these mechanisms can help us more fully characterize participation patterns observed in different educational environments and leverage this knowledge in the design of such systems. We present a case study of two students in both classroom and online French learning environments and show how the psychological mechanisms of the bystander effect help us understand observed behavioral changes. Any dialog—large or small, written or oral, scientific or casual—involves two well-defined roles: speaker and listener. In productive discourse, participants must play each of these roles at some point. For a speaker to speak, there must be an audience; for a discussion to continue, the audience must respond. Authors from many aspects of academia have attempted to define conversation and discussion in many ways, but these two features remain consistent. Ong (1982, p. 176) wrote: Correspondence and requests for reprints should be sent to James M. Hudson, College of Computing,

on the World Trade Center and the Pentagon, many scholars, journalists, and public officials expressed fear about the security of nuclear facilities in the

Recently, it has been argued that the evidence in social science research suggests that deceiving subjects in an experiment does not lead to a significant loss of experimental control. Based on this assessment, experimental economists were counseled to lift their de facto prohibition against deception to capture its potential benefits. To the extent that this recommendation is derived from empirical studies, we argue that it draws on a selective sample of the available evidence. Building on a systematic review of relevant research in psychology, we present two major results: First, the evidence suggests that the experience of having been deceived generates suspicion which in turn is likely to affect judgment and decision making of a non-negligible number of participants. Second, we find little evidence for reputational spillover effects that have been hypothesized by a number of authors in psychology and economics (e.g., Kelman, 1967; Davis and Holt, 1993). Based on a discussion of the methodological costs and benefits of deception, we conclude that experimental economists ’ prohibition of deception is a sensible convention that economists should not abandon.

Participants played four rounds of a social card game in which they formed groups to make card hands and earn money. When isolates—players left out when groups formed—earned nothing, self-organized groups frequently included these extra people, even though this decreased earnings for group members. Although group composition remained fluid, with time most populations settled into a pattern of including everyone. When isolates received a small welfare payment, group formation was less ordered and predictable and exclusion remained common. Although exact repeats of the same group membership were relatively rare in both conditions, group stability was associated with more cooperative behavior by members. When membership was unstable, members were more likely to make excessive claims on group earnings, increasing their individual payoff at the expense of fellow group members.

We would like to thank all the many researchers who contributed to the ProMES database over