Software Engineering Journal
Volume 6, Issue 4, July 1991
Volumes & issues:
Volume 6, Issue 4
July 1991
A C interpreter for Scheme — an exercise in object-oriented design
- Author(s): William Campbell
- Source: Software Engineering Journal, Volume 6, Issue 4, p. 130 –136
- DOI: 10.1049/sej.1991.0016
- Type: Article
- + Show details - Hide details
-
p.
130
–136
(7)
We describe a production quality implementation of Scheme. Its design, and therefore the program, is based around objects although it is written in C, a language not specifically intended for object-oriented programming. This was a fortuitous choice as it allowed us to recover from a design mistake that was not discovered until implementation time. Moreover, as we designed the program around objects and were not constrained by a strict object-oriented notation, this created a program that others have found easy to understand, extend and tune for improved performance.
HCLIE: a logic-based requirement language for new software engineering paradigms
- Author(s): Jeffrey J.-P. Tsai and Thomas Weigert
- Source: Software Engineering Journal, Volume 6, Issue 4, p. 137 –151
- DOI: 10.1049/sej.1991.0017
- Type: Article
- + Show details - Hide details
-
p.
137
–151
(15)
It is a paradigm of software engineering, which is especially popular among researchers involved in automatic programming, that a complete representation of the problem domain (called requirement theory) is repeatedly refined by correctness-preserving transformations into the program itself. The design of a requirement theory has to allow for such a design methodology. We argue that currently available languages intended to yield requirement theories fail to meet these demands. In this paper, we present a logic-based requirement language, which can be used to specify the requirements for software systems under the emerging methodology. The language presented here is based on first-order predicate logic but augments standard first-order logic by introducing hierarchies and exceptions to its generalisations, in order to allow for a more natural description of the problem domain. The proposed requirement language is valid and can be determined to be internally consistent. With the support of the requirement language, software development using new software engineering paradigms can be easily achieved.
A model for IS quality
- Author(s): Inger Eriksson and Aimo Törn
- Source: Software Engineering Journal, Volume 6, Issue 4, p. 152 –158
- DOI: 10.1049/sej.1991.0018
- Type: Article
- + Show details - Hide details
-
p.
152
–158
(7)
Different concepts of software quality are reviewed and discussed in this paper. Based on this, a hierarchical model of information system (IS) quality concepts developed in the research project SOLE* is presented. This model aims at a division of quality concepts consistent with the different decision makers and decisions made during the software life-cycle. The main division is into IS cost effectiveness, IS use quality and IS work quality. The last two are further divided into requirement quality, interface quality, and efficient IS management, evolution quality and operation quality, respectively.
A CCS case study: a safety-critical system
- Author(s): Jean Baillie
- Source: Software Engineering Journal, Volume 6, Issue 4, p. 159 –167
- DOI: 10.1049/sej.1991.0019
- Type: Article
- + Show details - Hide details
-
p.
159
–167
(9)
A level-crossing control system is specified in CCS, motivated by a temporal logic specification of the safety requirements. We show that, with certain reservations, these can be satisfactorily stated entirely within CCS. The crossing system is divided into two smaller subsystems, which are shown to be equivalent to the original single system, and whose behaviour is then analysed using the methods of the calculus. We also show that the crossing satisfies the safety requirements; there is no bisimulation although the system is both ‘may’ and ‘must’ equivalent to the specification. Some of the implications of this are discussed.
The HCI component of dependable real-time systems
- Author(s): A. Burns
- Source: Software Engineering Journal, Volume 6, Issue 4, p. 168 –174
- DOI: 10.1049/sej.1991.0020
- Type: Article
- + Show details - Hide details
-
p.
168
–174
(7)
The operator interface is an important aspect of any real-time system. If the system is a safety-critical one, the HCI components must themselves be dependable. Such applications constrain the time an operator has to assimilate and respond to system data. This paper focuses on the specific constraints these hard real-time applications present. Attention is focused on the interface between the application code and the HCI components. An architecture is outlined that allows these interfaces to be specified, and thus enables the application software and the HCI components to be designed and implemented separately.
Customers do not want frozen specifications
- Author(s): I.T.A. Spence and B.N. Carey
- Source: Software Engineering Journal, Volume 6, Issue 4, p. 175 –180
- DOI: 10.1049/sej.1991.0021
- Type: Article
- + Show details - Hide details
-
p.
175
–180
(6)
A technique is described that has been used to introduce flexibility into the specification of an interactive computer system. For the most part, the specification is written in English, but the portion of interest here uses a graphical notation which has a precise meaning, yet can readily be understood by a non-specialist. This portion of the specification can easily be changed in a controlled manner. The technique was carried right through to the design and implementation, which retained the same flexibility. It is described here by means of a simple illustrative example.
Restructuring operations for data-flow diagrams
- Author(s): Ming-Jie Chen and Chyan-Goei Chung
- Source: Software Engineering Journal, Volume 6, Issue 4, p. 181 –195
- DOI: 10.1049/sej.1991.0022
- Type: Article
- + Show details - Hide details
-
p.
181
–195
(15)
When defining and designing software with structured analysis and design methods, we need to restructure data-flow diagrams. Using basic editing operations to restructure large systems with voluminous data-flow diagrams is tedious, laborious and error-prone. It is necessary to have data-flow diagram editors that provide editing operations specific for restructuring. This paper proposes and formally specifies a set of operations sufficient for all restructuring needs. It also confirms that the properties of consistency and completeness are observed by all the restructuring operations, and that both models of data-flow diagrams are equivalent, before and after each of the restructuring operations.
A behavioural semantics for Linda-2
- Author(s): Paul Butcher
- Source: Software Engineering Journal, Volume 6, Issue 4, p. 196 –204
- DOI: 10.1049/sej.1991.0023
- Type: Article
- + Show details - Hide details
-
p.
196
–204
(9)
This paper presents a behavioural semantics for the language Linda, written in the specification language Z. The language defined is the dialect commonly referred to as Linda-2. Linda is a co-ordination language, and therefore has to be embedded in a computation language in order to produce a full parallel language. This specification attempts to be (as far as is possible) base language-independent, i.e. as few assumptions as possible are made about the computation language. We do, however, make some assumptions, in particular that the language is typed. This paper assumes no knowledge of Linda, but an understanding of Z (or, at least, set theory and predicate calculus) is assumed.
From objects to classes: algorithms for optimal object-oriented design
- Author(s): Karl J. Lieberherr ; Paul Bergstein ; Ignacio Silva-Lepe
- Source: Software Engineering Journal, Volume 6, Issue 4, p. 205 –228
- DOI: 10.1049/sej.1991.0024
- Type: Article
- + Show details - Hide details
-
p.
205
–228
(24)
The contributions of this paper are two-fold: first, we introduce a novel, axiomatically defined, object-oriented data model called the Demeter kernel model; and secondly, we present abstraction and optimisation algorithms and their relationships for designing classes from objects in the kernel model. We analyse several computational problems underlying the class design process, which is divided into two phases; a learning phase and an optimisation phase. This study focuses on approximation algorithms for the optimisation phase and aims to lead to a better understanding and a partial automation of the object-oriented design process. The algorithms and the theory presented here have been implemented in the C++ Demeter System, a CASE tool for object-oriented design and programming.
Most viewed content for this Journal
Article
content/journals/sej
Journal
5
Most cited content for this Journal
We currently have no most cited data available for this content.
