Showing papers in "ACM Computing Surveys in 1981"

Concurrency Control in Distributed Database Systems

Abstract: In this paper we survey, consolidate, and present the state of the art in distributed database concurrency control. The heart of our analysts is a decomposition of the concurrency control problem into two major subproblems: read-write and write-write synchronization. We describe a series of synchromzation techniques for solving each subproblem and show how to combine these techniques into algorithms for solving the entire concurrency control problem. Such algorithms are called "concurrency control methods." We describe 48 principal methods, including all practical algorithms that have appeared m the literature plus several new ones. We concentrate on the structure and correctness of concurrency control algorithms. Issues of performance are given only secondary treatment.

The Recovery Manager of the System R Database Manager

Abstract: The recovery subsystem of an experimental data management system is described and evaluated. The transactmn concept allows application programs to commit, abort, or partially undo their effects. The DO-UNDO-REDO protocol allows new recoverable types and operations to be added to the recovery system Apphcation programs can record data m the transaction log to facilitate application-specific recovery. Transaction undo and redo are based on records kept in a transaction log. The checkpoint mechanism is based on differential fries (shadows). The recovery log is recorded on disk rather than tape.

The Psychology of How Novices Learn Computer Programming

Abstract: This paper examines the current state of knowledge concerning how to increase the novice's understanding of computers and computer programming. In particular, it reviews how advances in cognitive and educati(~nal psychology may be applied to problems in teaching nonprogrammers how to use computers. Two major instructional techniques are reviewed: providing a concrete model of the computer and encouraging the learners to actively put technical information into their own words.

Formal Models for Computer Security

Abstract: Efforts to build "secure" computer systems have now been underway for more than a decade. Many designs have been proposed, some prototypes have been constructed, and a few systems are approaching the production stage. A small number of systems are even operating in what the Department of Defense calls the "multilevel" mode some information contained m these computer systems may have a clasmfication higher than the clearance of some of the users of those systems. This paper revmws the need for formal security models, describes the structure and operation of military security controls, considers how automation has affected security problems, surveys models that have been proposed and applied to date, and suggests possible d~rectlons for future models

Garbage Collection of Linked Data Structures

Abstract: A concise and unified view of the numerous existing algorithms for performing garbage collection of linked data structures is presented. The emphasm is on garbage collection proper, rather than on storage allocatlon. First, the classical garbage collection algorithms are reviewed, and their marking and collecting phases, with and without compacting, are discussed. Algorithms descnbing these phases are classified according to the type of cells to be collected: those for collecting single-sized cells are simpler than those for varimzed cells. Recently proposed algorithms are presented and compared with the classical ones. Special topics in garbage collection are also covered: the use of secondary and virtual storage, the use of reference counters, parallel and real-time collections, analyses of garbage collection algorithms, and language features whlch influence the design of collectors. The bibhography, wlth topical annotations, contains over 100 references.

Human Factors Studies of Database Query Languages: A Survey and Assessment

Abstract: Empirical studies have been undertaken to measure the ease-of-use of a query language, compare two or more such languages for ease-of-use, study controversial issues in query language design, and provide feedback to designers for improving a language. Some primitive at tempts at constructing abstract models related to query languages also exist. This paper discusses some of the techniques that have been used and results obtained. A primary goal is to show the reader unfamiliar with behavioral research what the results do and do not mean.

Guest Editor's Introduction: An Applied Psychology of the User

Abstract: This special issue of COMPUTING SURVEYS on \"The Psychology of the Computer User\" represents the debut of user psychology, as a coherent subfield of computer science, to the computing community. In the sixties only a handful of researchers were exploring the psychology of the computer user (e.g., SACK70, WEIr71). During the decade of the seventies the researchers venturing into this difficult area began to number in the tens. By the end of the decade a large enough body of work had been amassed for us to begin to think of the area as a distinct field of study, as evidenced by substantial reviews [RAMS79], books [SHNES0, RousS0, CARD81], and the establishment of the \"Human Aspects of Computing\" department in the Communicat ions of the ACM. This special issue thus presents a review and assessment of the field of user psychology; 1980. It should serve as a good starting point for anyone interested in the topic, either as a potential worker in the field or as a consumer of knowledge about the user. In putting together this special issue I have had to make several restrictions on its scope. Only limited, but crucial, aspects of user psychology are treated here. This special issue is concerned with the psychology of the indiv idual user and does not treat the user as a social agent. Further, the focus is on the cognitive aspects of the user, which include learning, performing, and reasoning, and not on the motivational, emotional, or personality aspects. This does not imply that these other aspects are unimportant, but it does reflect the judgment that the cognitive aspects form a coherent body of phenomena on which we can make scientific and practical progress. Finally, all the work surveyed here is based on empirical research and not on casual observation and speculation--a hard discipline to work under, especially in a free-wheeling field like computer science, but necessary if we are ever to have a science of the user. In addition to the usual task of outlining the papers in the special issue, I would like to present a few orienting ideas~regarding both the substance of user psychology and the nature of user psychology as a discip l i n e t o put the various papers in perspective. Section 1 reflects on computer science's attitude toward the user and how psychology can contribute. Section 2 sketches some rudimentary concepts that are helpful to keep in mind when thinking about user behavior. Section 3 discusses the nature of the field by contrasting different approaches to applying user psychology to system design. Section 4 summarizes the papers in this issue, thus giving a picture of the current status of the field. And finally, Section 5 briefly assesses the current status of the field of user psychology: its results, its directions, and its prospects.

Representation of Three-Dimensional Digital Images

Abstract: Three-dimensional digital images are encountered m a variety of problems, including computed tomography, biological modeling, space planning, and computer vision. A wide spectrum of data structures are available for the computer representation of such images. This paper is a tutorial survey of three-dimensmnal spatial-data representation methods emphasizing techniques that apply to cellular (or voxel-based) images. We attempt to unify data structures for representing interior, surface, and structural information of objects in such images by companng their relative efficmncy. The derivation of high-level representatmns from serial sectmn images is also discussed The representations include topological representations (Euler characteristic and adjacency trees), geometrical representatmns {borders, medial axes, and features), and spatial organization representations {generalized cyhnders and skeletons).

Human-Computer Interaction in the Control of Dynamic Systems

Abstract: Modes of human-computer interaction in the control of dynamic systems are discussed, and the problem of allocating tasks between human and computer considered. Models of human performance in a variety of tasks associated with the control of dynamic systems are reviewed. These models are evaluated in the context of a design example involving human-computer interaction in aircraft operations. Other examples include power plants, chemical plants, and ships.

Image Models

Abstract: Models of spatial variation used m the computer processing of pictorial information are surveyed. Models of images depictmg homogeneous textures are reviewed under the categories of pixel-based and region-based models. Plxel-based models are further divided into one-dimensional tune series models, random field models, and syntactic models The random field models mcorporate either global or local properties of an image. To put the role of the two low-level models in perspectwe, several high-level models making use of semantic reformation are also described.

Behavioral Aspects of Text Editors

Abstract: Theoretical models and experimental results relevant to the study of behavioral issues in the use of text editors--including both those intended primarily for computer program development and those intended for manuscript preparation--are examined. Models can predict editing task time in terms of elementary activities, in an error-free environment, to an accuracy comparable to the variability between subjects. In a realistic setting, however, unpredictable user activities account for between 25 and 50 percent of the task time, an amount that is comparable to individual variations due to errors. Variations in computer response time appear to affect users more than mere delay does. Command options improve expert performance but degrade the performance of beginners. The surface syntax of an editor can have considerable impact on ease of use. Ergonomic aspects of keyboard and display terminal design and use are well understood, with little hope for significant improvement, but there is no experimental evidence to support guidelines for display format design. Among analog pointing devices the mouse appears to have a small edge over the light pen, joystick, and track ball; human pointing performance using these devices approaches known psychophysical limits. Optimum ambient conditions, including temperature, noise, work-station layout, illumination, and work-rest cycles derived for professional key entry operators and for other interactive tasks, are probably also valid for editing. Gaps in the application of cognitive psychology and human engineering to text editors in the literature are indicated, and promising research areas are delineated.

Verifying Security

Abstract: Software correctness and reliability are considered especially important when the protection of confidential data is at stake. Data protection is critical in any multipleuser system having files that may be shared on a limited basis. Since one function of an operating system is to control and provide access to files, the operating system often enforces the desired protection policy. The study of operating system security is the study of how to design systems that successfully protect confidential data, and how to verify that a proposed design and implementation will provide the required protection. Program specification and verification techniques have influenced, and been influenced by, work in the area of operating system security. Operating systems, considered as large bodies of software, are still not quite within the practical reach of program verification, but the designs of their dataprotection mechanisms can be verified today. A few of the recent activities in this area are documented in WALK80, McCA79, and SCHA80. In these and other projects only a portion, rather than the whole operating system, is scrutinized. The portion, which contains enough of the critical functions to protect both itself and user files from unauthorized access, is sometimes called the \"security kernel.\" Automated specification and verification environments have been developed to aid in producing correct and reliable software. This paper focuses on the way they are used to prove security properties of an operating system design. We survey and compare a few prominent systems that have been, or could reasonably be, used for this purpose. A recent collection of short summaries covering the verification technology area more broadly may be found in VERK80. This paper discusses the following four systems:

Testing and Debugging Custom Integrated Circuits

Abstract: Although designing custom integrated circuits does not seem to be any harder than writing computer programs, debugging integrated circuit designs is much more cumbersome than testing and debugging programs. This paper examines some of the approaches to testing and debugging integrated circuits and describes how simulation and access to the internal state of a chip can make testing and debugging custom integratedcircmt designs as easy as testing and debugging software.

Surveyor's forum: augmentation or dehumanization?

Abstract: We would like to take issue with the introductory paper \"An Applied Psychology of the User\" in COMPUTING SURVEYS 13, 1 (March 1981), 1-11. We feel that the article treats questions of user psychology from a wrong perspective and with a wrong approach. In the view presented in the article, users (= people) are reduced to (machinelike) error-prone components of the human-computer system. If research is pushed on in the direction suggested in the article, the results may contribute to further dehumanizing the interaction between people and computer-based systems. The argument of the article relies on a doubtful division of \"user aspects,\" which necessary in order to focus attention exclusively on cognitive aspects of the user acting as an individual (e.g., the defmition of cognitive aspects in the article does not take into account that human learning is partly determined by motivation and by cooperation with other people). While this division clearly facilitates the building of calculational models, there is no user available on earth who exhibits these traits alone--and does not at the same time exhibit \"personality aspects\" and act as a social agent; thus there can be no basis (no \"objects\") for empirical studies strictly along the lines suggested by the author. It is great to have models, but what or rather whom are these models modeling? \"Studying the user objectively,\" as required in the article, implies that only those dimensions of user behaviorwhich are measureable can be considered. Predicting and controlling user behavior on the basis of the models thus obtained will contribute to increasing the performance of the system at the level of individual transactions, but actual user needs will not be taken into account. Even if system performance is the overriding goal, this approach will produce misleading results, since the overall system performance will only be satisfactory if the system does not force the user either to decompose his work into awkward steps or to work against or around the system in order to maintain a reasonable work style. Clearly there is no way to build calculational models of human behavior, unless this behavior is artificially restricted to a small set of patterns which are amenable to being modeled in those terms that we, being computer scientists, are used to work with; this tendency is in keeping with the hope expressed in the article to get help from artificial intelligence. Though the article gives no definition, the terms \"behavior,\" \"task,\" and others used in the article to describe users only make sense to us if they are meant to refer to the view that the computer system has of the user, that is, the view of a man as \"seen\" by the machine. The approach described in the article amounts to controlling human behavior in order to subordinate it to the purpose of computer system performance irrespective of the goals that people wish to achieve with the help of the system. We raise objection to this kind of control as a goal of scientific research. As an alternative, we suggest that we study applied psychology in order to find out how to make computer-based systems more adequate as tools for their human users. Taking this approach, one would focus on such questions as, \"How should computer-based systems be embedded into the working context of their users? . . . . What kind of computer system models are appropriate for the users? .... How can we cent ~bute to making systems more transparent to

Surveyor's Forum: Augmentation or Dehumanization?

Abstract: transactions and streams of transactions best be arranged to correspond to meaningful work steps? .... What does 'meaningful' mean, psychologically?\" Taking this approach, we also will start by considering individual transactions, but we will pose different questions and interpret the answers in a different context. We will not be led to a user psychology which has its home base in computer science (and thus draws primarily on computer science concepts and methods). Instead, we see the need for psychologists who work from users' needs toward establishing requirements for computer-based systems and with whom system designers cooperate. Floyd, Keil, and Nullmeier seem to be making three points: (1) it is misleading to focus on the cognitive aspects of user behavior; (2) information-processing models of users reduce users to machine components; and (3) there is an alternative approach to a user psychology that is not based on computer science. I will take up each point in turn. Focus on Cognition. Floyd et al. claim that focusing on the cognitive aspects of the user neglects other more important aspects , such as his motivational and social aspects. Thus, a cognitively oriented user psychology will neglect \"actual user needs,\" forcing the user into an awkward \"work style.\" Further, since there is no such thing as a purely cognitive user, there cannot be an empirical science of such a user. But of course we must focus in studying anything so complex as the computer user, for no matter what we decide to study there will always be other aspects which someone else will consider more important. Thus, even if we include social interaction among groups of users in our studies, someone might argue that the real issues are organizational. Or economic. Or political. In my article I did point out that there are many aspects of the user other than cognition that must be studied. The article states that my focus on the cognitive aspects \"does not imply that these other aspects are unimportant, but it does reflect the judgment that the cognitive aspects form a coherent body of phenomena on which we can make scientific and practical progress.\" The fact is that users of current computer systems have massive cognitive problems-most systems are difficult to learn and remember, it is impossible for anyone but system hackers to understand them, they allow users to make fatal errors, and so on. These must be attacked on …

Surveyor's Forum: Generating Solutions

Abstract: In \"Some Techniques for Solving Recurrences\" [LUEKB0], one of the suggested approaches is to use generating functions. Generating functions are properly considered as f o rmal power series, so convergence considerations need usually not apply. Indeed, Lueker's principal reference for asymptotics [BEND74] discusses methods for the manipulation of generating functions with a zero radius of convergence. Such manipulation is possible since the set of formal power series forms an integral domain; Lueker's checks for convergence are unnecessary, and his \"caution\" [LuEKB0, p. 431] is invalid. There are very simple algori thms, employing generating functions, which can be used to solve inhomogeneous linear difference equations, thus avoiding the guessing technique or heuristic manipulations suggested in LUEKB0. We give one example. In

Surveyor's Forum: Interpreting Experimental Data

Abstract: more dehumanizing than modeling human blood flow as hydraulics. Although there is a sense in which the human is being \"re-duced,\" this is the normal way in which science attempts to come to a theoretical understanding of its object of study. Today, any cognitively oriented user psychology must begin with an information-processing view of the user if it wants to build on the work of modern cognitive psychology. Treating the user as an information-processing system interacting with the computer in no way makes the user a component of the computer. The user's information processing capabilities are radically different from the computer's. That is the point of a user psychology-to understand the user on his terms. My article was quite explicit on this: \"We can make no sense of [the user's] behavior without knowing his [i.e., the user's] goal.\" Quite the opposite from subordinating the user to the computer , a theoretically based user psychology provides the crucial element-an understanding of the user-to subordinate the computer to the user's ends. The Alternative Approach. Floyd et al. propose an alternative approach to a user psychology based on \"psychologists who work from users' needs\" to make computers \"more adequate as tools for their human users.\" They propose that these psychologists should not themselves be based in computer science, in order to best represent user interests in the design process. Their alternative approach is no alternative at all at the level of objectives, for, as stated above, the kind of user psychology I am advocating does work from the user's goals. Their alternative appears to be more political (or institutional) than scientific, at least in its emphasis. They seem to view the user psychologist as a negotiator-an advocate of the user on the design team. I view the user psychologist as a scientist developing models of the user and user-oriented design techniques to give to the designer. I do not see any necessary conflict in these views, since there could be user psychologists of both kinds. The more general issue here has to do with the dual nature of an applied science. Surveyors' Forum • 493 A good applied science has needs on both its science side and on its application side. The science side requires powerful tools (ideas and techniques) to be able to affect the world, and the application side requires the good taste (aesthetic, political, moral) to affect the world for …

Surveyor's Forum: Generating Solutions

Abstract: In \"Some Techniques for Solving Recurrences\" [LUEKB0], one of the suggested approaches is to use generating functions. Generating functions are properly considered as f o rmal power series, so convergence considerations need usually not apply. Indeed, Lueker's principal reference for asymptotics [BEND74] discusses methods for the manipulation of generating functions with a zero radius of convergence. Such manipulation is possible since the set of formal power series forms an integral domain; Lueker's checks for convergence are unnecessary, and his \"caution\" [LuEKB0, p. 431] is invalid. There are very simple algori thms, employing generating functions, which can be used to solve inhomogeneous linear difference equations, thus avoiding the guessing technique or heuristic manipulations suggested in LUEKB0. We give one example. In