Long-running Programs Research Articles

Two Ways of Being Italian on Global Television

Two Ways of Being Italian on Global Television

Exceptional situations and program reliability

It is difficult to write programs that behave correctly in the presence of run-time errors. Proper behavior in the face of exceptional situations is important to the reliability of long-running programs. Existing programming language features often provide poor support for executing clean-up code and for restoring invariants. We present a data-flow analysis for finding a certain class of exception-handling defects: those related to a failure to release resources or to clean up properly along all paths. Many real-world programs violate such resource usage rules because of incorrect exception handling. Our flow-sensitive analysis keeps track of outstanding obligations along program paths and does a precise modeling of control flow in the presence of exceptions. Using it, we have found over 1,300 exception handling defects in over 5 million lines of Java code. Based on those defects we propose a programming language feature, the compensation stack, that keeps track of obligations at run time and ensures that they are discharged. We present a type system for compensation stacks that tracks collections of obligations. Finally, we present case studies to demonstrate that this feature is natural, efficient, and can improve reliability.

Translation as an Ecological Tool for Instrument Development

In this paper the authors outline the translation process involved in Macro International's evaluation of the Department of State's International Visitor Leadership Program (IVLP) in Russia, Ukraine, Kazakhstan, and Georgia. IVLP is a long-running program in which professionals and prospective leaders from around the world participate in funded short-term visits to the United States to learn first-hand professional practices and values of American society and democracy. The authors highlight the importance of attending to the theoretical issues in, discuss contextual factors inherent in, and outline specific phases of the translation process, and present the modified decentering translation technique adapted for the project. They describe the types of translation equivalencies that were addressed and present findings that attest to the quality of the translation. They underscore the importance of the translation process as a qualitative tool for the instrument development that maps the contexts of people's lives, documents emic-etic aspects of cross-cultural research, and fosters collaborations with all stakeholders of the research project.

Artemis

A number of hardware and software techniques have been proposed to detect dynamic program behaviors that may indicate a bug in a program. Because these techniques suffer from high overheads they are useful in finding bugs in programs before they are released, but are significantly less useful in finding bugs in long-running programs on production systems -- the same bugs that are the most difficult to find using traditional techniques. In this paper we propose the Artemis 1 is the Greek goddess of the hunt and wild animals. Our framework guides the hunt for wild bugs. compiler-based instrumentation framework that complements many pre-existing runtime monitoring techniques. The Artemis framework guides baseline monitoring techniques toward regions of the program where bugs are likely to occur, yielding a low asymptotic monitoring overhead. Artemis also facilitates system-load aware runtime monitoring that allows the monitoring coverage to be dynamically scaled up to take advantage of extra cycles when the system load is low, and dynamically scaled down to monitor only the most suspicious regions when the system load is high. Our experiments show that Artemis' asymptotic overhead can outperform the performance floor overhead of random sampling for many tools, and that Artemis can effectively guide a monitoring tool to the buggy regions of a program. Our experimental results show that Artemis applied to a hardware-based PC-invariance monitoring scheme and a value-based invariance detection and checking scheme significantly improves their runtime monitoring overhead (by up to 4.6 times) with moderate impact on their bug-detecting capabilities.

Bristol-Myers Ends No-Strings Grants

BIOMEDICAL RESEARCH Scientists are mourning the cancellation of a long-running research grants program funded by a major drug company. The Bristol-Myers Squibb (BMS) Freedom to Discover program, begun in 1977, only supports about 50 biomedical scientists at a time. But the grants, about $6 million a year recently, come with no strings attached. That feature, which allows for high-risk research, is particularly welcome at a time when U.S. funding for biomedical research is tightening. Some scientists are troubled that the company is pulling the plug in part because of the growing global debate over the ethics of corporate payments to academic physicians. “I think they've gone overboard and are tanking a wonderful program,” says grantee Carl June, a cancer researcher at the University of Pennsylvania. BMS says that Freedom to Discover is the largest corporate-funded, unrestricted research grants program in the world. Scientists can't just bid for the grants, however. Instead, BMS scientists identify potential recipients doing work of interest to the company in six biomedical fields and invite them to compete. The winners, chosen largely based on their track records, receive $100,000 a year for 5 years. The grantees also meet annually to choose a distinguished scientist to receive a $50,000 lifetime achievement award considered to be among the most prestigious in their fields. The resulting flexibility to follow one's hunches is extremely rare, says Johns Hopkins University neuroscientist Michela Gallagher. She says her search for neurobiological markers that explain why some rats remain mentally sharp into old age might be seen as a “fishing expedition” by a U.S. National Institutes of Health study section. Others have used the company's money to support postdocs until they get their first grant or to collect chimp fecal samples in Africa for an HIV study. BMS makes no claim to any of the findings. “There's lots of payola within the pharmaceutical industry, but this is one of the few programs that is really squeaky clean,” says immunologist W. Allan Walker of Harvard Medical School in Boston, who is also a recipient. ![Figure][1] Fishing license. Michela Gallagher says a BMS grant lets her explore promising ideas. CREDIT: JOHNS HOPKINS UNIVERSITY Earlier this year, the company began telling grantees, many of whom are not physicians, that it was changing some rules to avoid the perception of any conflict of interest. Spouses could no longer attend the awards selection meeting for free, for example, and grantees were asked to sign an agreement saying they were consultants to BMS. BMS spokesperson Rebecca Taylor says the program was killed in order to expand efforts such as a $150 million, multi-year program that funds pediatric AIDS clinics in Africa. But “an increase in compliance regulations affecting the global pharmaceutical industry” is a contributing factor, she adds. Some recipients say they were told that BMS lawyers felt the company could run afoul of new, restrictive regulations in Europe on corporate gifts to physicians. [1]: pending:yes

Creative Communities after Television: The Collective Authorship of Channel 101

Creative Communities after Television: The Collective Authorship of Channel 101

Bio-Power: CBC Television's Life & Times and A&E Network's Biography on A&E

In a section of Imagined Communities Benedict Anderson presents biography as part of the documentation of identity that is based on a forgetting and then repetition of origins that cannot be remembered. The nation itself is a narrative made of the collection of individual stories. Anderson's image of biography as the site of a forgetting and as the means of national suturing is an excellent way to examine how the genre of biography works as a social practice within national discourse in two popular biography shows on television: the American cable network A'E's flagship series Biography on A'E and the Canadian television network CBC's long-running biography programme Life & Times. In this paper, I look at selected programmes by each network about artists, political figures and sports stars as one of the latest attempts to make “biographies of the nation” on television in Canada and the United States. In my analysis, genre is understood as a set of social practices that constructs and limits many types of cultural meaning for audiences and cultural producers. “Biography” as a genre has been translated into the medium of television without losing sight of its origins in popular collective biography of the Victorian era.

Dynamic hot data stream prefetching for general-purpose programs

Prefetching data ahead of use has the potential to tolerate the grow ing processor-memory performance gap by overlapping long latency memory accesses with useful computation. While sophisti cated prefetching techniques have been automated for limited domains, such as scientific codes that access dense arrays in loop nests, a similar level of success has eluded general-purpose pro grams, especially pointer-chasing codes written in languages such as C and C++. We address this problem by describing, implementing and evaluating a dynamic prefetching scheme. Our technique runs on stock hardware, is completely automatic, and works for general-purpose programs, including pointer-chasing codes written in weakly-typed languages, such as C and C++. It operates in three phases. First, the profiling phase gathers a temporal data reference profile from a running program with low-overhead. Next, the profiling is turned off and a fast analysis algorithm extracts hot data streams, which are data reference sequences that frequently repeat in the same order, from the temporal profile. Then, the system dynamically injects code at appropriate program points to detect and prefetch these hot data streams. Finally, the process enters the hibernation phase where no profiling or analysis is performed, and the program continues to execute with the added prefetch instructions. At the end of the hibernation phase, the program is de-optimized to remove the inserted checks and prefetch instructions, and control returns to the profiling phase. For long-running programs, this profile, analyze and optimize, hibernate, cycle will repeat multiple times. Our initial results from applying dynamic prefetching are promising, indicating overall execution time improvements of 5.19% for several memory-performance-limited SPECint2000 benchmarks running their largest (ref) inputs.

Cost-effectiveness of harm reduction in preventing hepatitis C among injection drug users.

Hepatitis C (HCV) has emerged as a major epidemic among injection drug users (IDUs), with observed prevalence exceeding 70% in many American and European cities. This article explores the potential of syringe exchange programs (SEPs) to reduce HCV incidence and prevalence. A random-mixing epidemiological model is used to examine the potential impact of harm reduction interventions. Steady-state analysis is used to scrutinize the impact of SEP on HCV incidence and prevalence and to examine the accuracy of short-term incidence analysis in predicting long-run program effects. SEP is predicted to have little impact on HCV incidence and prevalence within realistic populations of IDUs. Short-term incidence analysis substantially overstates SEP effectiveness and cost-effectiveness in preventing HCV. More comprehensive harm reduction models, coupled with referral of active IDUs to treatment, must complement syringe exchange to successfully contain highly infectious blood-borne diseases.

NSF Urban Systemic Initiatives

In their Editorial “Science learning, science opportunity ( Science 's Compass, 8 Oct., p. [237][1]), National Science Foundation (NSF) Director Rita R. Colwell and National Science Board Chairman Eamon M. Kelly give a glowing but what seems a largely unfounded report on the accomplishments of the NSF systemic reform initiatives. The authors cite as evidence an increase in student performance in mathematics in Chicago, Illinois (61 out of 62 high schools), and a tripling of students passing science and mathematics advanced placement (AP) tests in Dallas, Texas. Although it is true that Chicago public high schools have shown increases in standardized (Illinois Goals Assessment Program) test scores in mathematics over the past 2 years (in 51 out of 60 schools reporting), increases in the same students' reading scores have been even greater (57 out of 60). It seems plausible that Chicago's lower performing students are doing better on math tests because they can read the problems, not necessarily because of NSF systemic reform. Only the three top-scoring high schools in Chicago exceeded the Illinois average for math scores in 1997, and all three showed a drop in math scores the following year. Seven out of the nine Chicago schools with falling math scores started above the district average. Student scores on science and math AP tests have also risen in Dallas, but once again the role of NSF systemic reform is questionable. The Dallas-based O'Donnell Foundation has developed a long-running incentive program in which AP teachers receive special course training and a salary supplement, and AP students receive a subsidy for exam fees and a $100 minischolarship if they pass the AP College Board exam. The O'Donnell Foundation has tracked and reported on the efficacy of their work and can account for the student gains. In their Editorial, the NSF director and the National Science Board chairman appear to take credit for what is very likely to be another group's accomplishment. “Discovery learning” methods of science and math instruction may be widely embraced by the education community, but if the scores from Chicago and Dallas are the most robust data that NSF can assemble to justify their systemic initiatives, then the program should be carefully scrutinized. # Response {#article-title-2} Metzenberg's letter questions the specific examples we used to illustrate our comments on the NSF-supported Urban Systemic Initiatives (USIs). The nature of systemic reform activity—aligning and integrating the uses of different types of resources—makes explicit cause-and-effect attribution difficult. There are always several related indicators of performance and multiple contributors to changes in performance for any one indicator. The factors Metzenberg cites in Chicago and Dallas may well contribute to the phenomena we described. However, we are confident, on the basis of the reports of similar results across the USIs, that the systemic reform activities played an important role as well. We are pleased that NSF has helped so many urban public school systems to move forward in improving their science and mathematics education. Our examples—drawn from public testimony at a National Science Board hearing on kindergarten through grade 12 education held in Chicago—are just the tip of an iceberg of data coming from Chicago and Dallas; Detroit, Michigan; El Paso, Texas; Memphis, Tennessee; and many other locations. [1]: /lookup/doi/10.1126/science.286.5438.237

Checkpointing and rollback recovery for network of workstations

Network of workstations (NOW) now becomes one of the main trends of parallel computing. But for long-running scientific programs, it needs effective fault tolerance for its changing property. Checkpointing and rollback recovery is a solution to this problem. First the main problems upon rollback recovery are discussed, the different checkpointing techniques for NOW are analyzed, and then the design and implementation of ChaRM (checkpoint-based rollback recovery and process migration) system are described. The comparison of three coordinated checkpointing systems is given.

TRIO: The Unwritten Legacy

In this article, the authors argue for and operationalize the adoption of auto-ethnographic research methods to explore the impact of TRIO participation on the development of social, political, and professional capacities among educators who have been students and staff members of the programs. They invite the reader on a journey of reacquaintance with their individual educational histories, presented in the form of a textual collage. Through these reflections, they explore the intersections of cultural and racial/ethnic identity and discuss the implications of understanding such histories and identities for the administration of TRIO programs today. ... the future of educational research- . a greater role for imagination, a greater reliance on metaphorical thinking, and a greater openness to the visions of human possibility opened by our artists in the present and the past. (Greene, 1998, p. 35) ... the possibilities of storytelling are precisely those of understanding the human experience. (Momaday, cited in Fischer, 1986, p. 225) The federal TRIO programs have been studied from multiple angles with various research methodologies since the inception of the earliest TRIO initiatives in the 1960s. For example, many articles published in scholarly journals have focused on the practices and outcomes of TRIO's largest and longest-running program, Upward Bound (Billings, 1968; Butler & Gipson, 1975; Bybee, 1969; Dottin, Linton, & Roberts, 1981; Egeland, Hunt, & Hardt, 1970; Exum & Young, 1981; Frost, 1967; Garns, 1971; Gill, 1969; Helyar, 1977; Herson, 1968; Joseph, 1968; Lewenstein, 1974; McCormick & Williams, 1974; Mims, 1985). The literature on Upward Bound also includes numerous doctoral dissertations (Allen, 1975; Bemak, 1976; B. Brown, 1976; J. Brown, 1993; Burris, 1969; Coron, 1969; Dease, 1979; Dixon, 1982; Ehrbright, 1969; Farrow, 1976; Geisler, 1968; Hattman, 1974; Hollis, 1974; Jackson, 1976; James, 1979; Jawa, 1969; Jones, 1991; Lewis, 1982; McCormick, 1971; Moore, 1974; Mullins, 1974; Nash, 1974; Palmer, 1979; Powe, 1990; Richardson, 1974; Roland, 1981; Seehe-Fields, 1972; Slaughter, 1983; Smith, 1967; Stewart, 1978; Thompson, 1973; Waite, 1968; White, 1991; Young, 1973). Additionally, many evaluation studies of Upward Bound have been commissioned and conducted (Burkheimer, French, Levinsohn, & Riccobono, 1977; Greenleigh Associates, 1970; MacKenzie, 1983; Moore, Fasciano, Jacobson, Myers, & Waldman, 1997). This literature, though dominated by quantitative research addressing student outcomes, offers some qualitative research addressing the TRIO programs' structure, operations, and student outcomes. What absent, however, auto-ethnographic research such as that described by Denzin (1997a), Ellis (1996), Ellis and Bochner (1996), Thompson and Sageeta (1996), White and Shelley (1996), and White, Mogilka, and Slack (1998). Arguably, this methodology could be used to address the impact of TRIO programs on the future professional lives of educators who have been students and staff members in the programs.1 The absence of this type of research understandable, given that the dominance of Eurocentric and positivist research paradigms has only recently been sufficiently challenged and interrupted to encourage experimentation with novel methodologies (for examples, see Eisner, 1997; Scheurich & Young, 1997; Tyson, 1998). Auto-ethnographic research underscores the crucial need for individuals to write their own histories. As Lincoln (1983) explains, the sender or storyteller/historian fulfills this role in Native American cultures. He notes that by characterizing the people's collective reality-tonaliz[ing], lighten[ing], spiritualiz[ing], brighten[ing], and darken[ing] [their] experience, all the while working with the reality that is (p. 223)-- word senders strive to create a living story [with] the personal inflection embodied and embraced in communal history (p. …

Performance debugging shared memory parallel programs using run-time dependence analysis

We describe a new approach to performance debugging that focuses on automatically identifying computation transformations to reduce synchronization and communication. By grouping writes together into equivalence classes , we are able to tractably collect information from long-running programs. Our performance debugger analyzes this information and suggests computation transformations in terms of the source code. We present the transformations suggested by the debugger on a suite of four applications. For Barnes-Hut and Shallow, implementing the debugger suggestions improved the performance by a factor of 1.32 and 34 times respectively on an 8-processor IBM SP2. For Ocean, our debugger identified excess synchronization that did not have a significant impact on performance. ILINK, a genetic linkage analysis program widely used by geneticists, is already well optimized. We use it only to demonstrate the feasibility of our approach to long-running applications.We also give details on how our approach can be implemented. We use novel techniques to convert control dependences to data dependences, and to compute the source operands of stores. We report on the impact of our instrumentation on the same application suite we use for performance debugging. The instrumentation slows down the execution by a factor of between 4 and 169 times. The log files produced during execution were all less than 2.5 Mbytes in size.

Optimal tracing and replay for debugging message-passing parallel programs

A common debugging strategy involves reexecuting a program (on a given input) over and over, each time gaining more information about bugs. Such techniques can fail on message-passing parallel programs. Because of nondeterminacy, different runs on the given input may produce different results. This nonrepeatability is a serious debugging problem, since an execution cannot always be reproduced to track down bugs. This paper presents a technique for tracing and replaying message-passing programs. By tracing the order in which messages are delivered, a reexecution can be forced to deliver messages in their original order, reproducing the original execution. To reduce the overhead of such a scheme, we show that the delivery'order of only messages involved inraces need be traced (and not every message). Our technique makes run-time decisions to detect and trace racing messages and is usuallyoptimal in the sense that the minimal number of racing messages is traced. Experiments indicate that only 1% of the messages are often traced, gaining a reduction of two orders of magnitude over traditional techniques that trace every message. These traces allow an execution to be reproduced any number of times for debugging. Our work is novel in that we adaptively decide what to trace, and trace only those messages that introduce nondeterminacy. With our strategy, large reductions in trace size allow long-running programs to be replayed that were previously unmanageable. In addition, the reduced tracing requirements alleviate tracing bottle-necks, allowing executions to be debugged with substantially lower execution time overhead.

Optimal tracing and incremental reexecution for debugging long-running programs

Optimal tracing and incremental reexecution for debugging long-running programs

A personal multicomputer for parallel processing of long-running programs

Long-running programs executing a large number of numerical calculations require as high a performance as possible of the computer used. Monte Carlo programs, in particular, are among these, and they are applied, e.g. for simulating statistical physical processes. The performance of a personal computer is in most cases insufficient, but it has nevertheless been possible to increase it to more than ten times that of an IBM-AT-compatible computer. To achieve this, a personal computer of this type was supplemented by several plugged-in single board PC68020 microcomputers operating in parallel, each exhibiting two to three times the performance of the personal computer. The software developed for operating this multicomputer system allows programs to be run simultaneously, in particular those in Fortran or Pascal, each on one of the microcomputers and on the personal computer. The program running of the personal computer can also communicate with the other ones. For a Monte Carlo task, the programs simultaneously running on the microcomputers may be identical and merely use different starting values of the random number generator; the results of the calculations are afterwards combined. The identical programs can also run with different values of a physical parameter so that its influence can be studied. This is of prominent importance in developing a model for the physical process investigated.

Comments on “The History of Modern Shipbuilding Methods: The U.S.-Japan Interchange” and “The Productivity Problem in U.S. Shipbuilding”

Comparisons of technology and productivity between U.S. and Japanese shipyards are made in two papers in the first issue of the Journal.23 Such comparisons are very difficult and, to some extent, questionable because of the differences in ships and in cost-keeping practices. They are complicated further by the intervention of governments in shipyard operations. In any case, the comparisons show that equalling foreign productivity would not make U.S. shipyards cost-competitive. A coherent government policy implemented by coordinated action is necessary to ensure the operating stability which is essential for productivity and innovation. This was shown by the U.S. wartime program. Clear aims and policies and coordination by government and industry—the entire industry, not led by any one organization or individual, such as Kaiser—produced both extraordinary output and innovation. Many new technologies can only be applied successfully in long-run stable programs which make the productivity gains from both innovation and series production possible.

A technique for finding storage allocation errors in C-language programs

This paper describes a simple tool for diagnosing allocation errors in C-language [Kern 78] programs. Our tool consists of a traced replacement for the standard memory allocator, and a program to analyze traces. We have found this tool quite helpful in making long-running programs (e.g. text editors) more robust.

A technique for finding storage allocation errors in C-language programs

This paper describes a simple tool for diagnosing allocation errors in C-language [Kern 78] programs. Our tool consists of a traced replacement for the standard memory allocator, and a program to analyze traces. We have found this tool quite helpful in making long-running programs (e.g. text editors) more robust.

Evaluating Human Service and Correctional Programs By Modeling the Timing of Recidivism

This paper develops and illustrates a new statistical model of recidivism which enables program evaluators to (1) examine short-run program impact on the postponement of recidivism, through estimates of the average time at which recidivism occurs; (2) measure long-run program impact on the prevention of recidivism, through estimates of the ultimate probability ofrecidivism and (3) help determine when individuals have been successful long enough to be considered "safe," through estimates of their conditional probability of future recidivism.