Elizabeth A. Unger

A predictor for success in an introductory programming class based upon abstract reasoning development

The purpose of this study was to create and validate a tool which could be administered to students enrolled in or considering enrollment in an introductory programming course to predict success in the course or alternatively to segregate enrolled students into fast and slow paced sections. Previous work which met the criteria of a self contained predictive tool included the work of Barry Kurtz [5] of the University of California, Irvine using abstract reasoning development as the predictive measure. The test Kurtz developed had been tested only on a small sample (23 students) in a controlled environment (one instructor - the researcher) and the test required up to 80 minutes to complete. This study modified the Kurtz test to require 40 minutes and administered it to 353 students learning two different languages from a variety of instructors. This predictor successfully predicted the advanced students from average to below average students. When used in conjunction with other known factors, e. g., GPA, the authors feel it is a viable tool for advising and placement purposes.

Entropy as a measure of database information

An estimate of the information a database contains and the quantification of the vulnerability of that database to compromise by inferential methods is discussed. Such a measure could be used to evaluate the deterrent value of extant protection methods and provide a measure of the potential for inferential compromise through the use of one of the known attack tools. The authors explore the use of the concept of entropy as defined for information by C.E. Shannon (1948; 1951), for the purpose of quantifying information content in a database and develop a measure of vulnerability based on entropy. Use of the measure, when exact disclosure through the use of a tracker is anticipated, is characterized for both static and dynamic databases at design and operational time.<<ETX>>

Another look at motivating data processing professionals

Motivation is a crucial management issue. The productivity of the U.S. worker is still dropping in the 1980s. The industrial revolution (automation of labor intensive industries) maintained productivity and increased standards of living earlier in this century. The computer revolution has a similar potential for the U.S. and the world. Another look at motivation thus appears indicated. Analysis of the data processing professionals personality profile as developed by recent studies (COU78, COU80, FIT78, LED80) in light of established theories of behavior results in several suggestions for managing computer personnel.

An extended memoryless inference control model: accounting for dependence in table-level controls

Memoryless inference controls are an important class of inference control methods for on-line statistical databases. Other inference controls are usually too complex to use in on-line systems. A subclass of memoryless inference controls which have been shown to require a reasonable level of computation are table level controls. Table level controls described in the literature make some assumptions about the structure of the database OH which they are used. We formalize these assumptions as database restrictions, and formally prove that these table level controls provide an accurate measure of the identification risk of a query. In addition we extend Lhcse controls to account for interdependence among attributes.

Natural change in dynamic databases as a deterrent to compromise by trackers

There has been considerable research in database security and inference control; however, that research deals primarily with statistical databases. Although some of the mechanisms proposed to deter compromise artificially introduce dynamicism into these otherwise static databases, very little effort has been devoted to the area of natural dynamics. The purpose of the work presented is to determine a deterrent value for natural change in dynamic databases. More specifically, the authors investigate how the changing value of an attribute prevents the inference of the confidential value of that attribute for any particular record. Assuming a restricted query set size, they calculate bounds of deterrence to compromise by the individual, general, and double trackers.<<ETX>>

Manifolds: a very high-level conceptual framework of interprocess synchronization and communication

This paper presents a new linguistic mechanism for expressing and controlling inter-process interactions. Unlike existent designs such as monitor- or message-based mechanisms, which generally provides a set of relatively low-level primitives for users to implement various inter-process synchronization constraints, this mechanism provides a general conceptual framework to help users visualize and specify inter-process synchronization and communication from a higher-level vantage point. This conceptual framework highlights two distinct and frequently-used high-level concepts in parallel programming : condition synchronization and scheduling and, in most of the useful cases, users can simply specify the synchronization constraints imposed by these concepts in a given problem, without having to implement the constraints with some lower-level primitives.

Design for integration of a DBMS into a network environment

This research deals with the methodology of integrating a data base management system (DBMS) into a distributed network system. The focus of concern is upon the interfaces and communications among a DBMS, an operating system with associated network interface features, and other such components at distributed nodes. The environment of a generalized data base management system is extended to a distributed generalized data base management system. The definition of this system by the authors allows the DBMS, schema, subschema, and application program to be distributed individually throughout nodes in the network and for the data base itself to be distributed. A network interface is defined consisting of a command interpreter and a command handler to effect and coordinate the communication of the units of the system (e.g., applications program with schema, etc.). The message transmission is studied for the various topological distributions, and for a restricted set of these topologies a selection algorithm is developed with considers simple priorities and estimated costs. The distribution of a generalized data base management system with the flexibility of locating the schema, subschema, application programs and data appears technically possible, and for some topologies and environments, economically feasible. Further study is indicated.

Security issues with TCP/IP

An introduction to network security , basic definitions and aa brief discussion of the architecture of TCP/IP as well as the Open System Intercornnection(OSI) Reference Model open the paper. The relationship between TCP/IP and of some OSI layers is described. An indepth look is provided to the major protocols in TCP/IP suite and the security features and problems in this suite of protocols. The secutiy problems are discussed in the context ofthe protocol services.

A survey of security issues in office computation and the application of secure computing models to office systems

With the proliferation of office information systems (OIS) into almost every area of industry and government, it is important to design systems that offer a guarantee of privacy and security to their users. The same solutions and research pertaining to traditional data-processing environments cannot, in most cases, be applied directly to the OIS envitonment. Many OIS do not provide the hardware/software controls necessary to protect information from anyone who gains physical access to the system. Furthermore, users of an OIS cannot be expected to possess a clear understanding of the system, its operating characteristics, or even the implication of interconnecting component devices. This paper examines the typical OIS environment with a view toward the provision of a secure operating architecture. Important security problems faced by OIS are enumerated and explained. We argue that the OIS environment presents a different problem to solve in a security sense from when working with a traditional nondistributed system. Existing solutions found in large scale operating systems and networks cannot simply be scaled down and moved to an OIS, if for no other reason than the architectures inability to support locks and multiple system states. An architecture of both software and hardware controls must be built for this new environment using concepts from large scale operating systems but recognizing the limitations and constraints of OIS. While we emphasize the security issues, we look at alternative technologies that can be combined to provide a solution.

Integrity in object-oriented database systems

This paper develops a model of integrity for object-oriented database management systems. The integrity system is layered over an extended object-oriented data model. The resulting model is shown to handle common state integrity constraints, including domain, key, and structural constraints. This model can enforce value-based functional dependencies from the relational model, both within an object and across object classes. A new relationship structural constraint, reference relationship, is presented which generalizes 1:1, 1:N, and N:M constraints in addition to total versus partial participation.