On the Utilisation of Persistent Programming Environments

There is a growing gap between the supply and demand of good quality software, which is primarily due to the difficulty of the programming task and the poor level of support for programmers. Programming is carried out using software tools which do not match very well either real world understanding of a problem or even the other tools which need to be used. In every phase of software production, the programmer must master new tools which function in a different way from each other. The Persistent Programming Paradigm attempts to reduce these problems by providing a programming environment which gives consistent methods of accessing program values of various kinds. Long-term and short-term data are treated in the same way. Numbers, text, graphical values and even program objects are all referred to in the same consistent way. Languages which support persistence provide considerable power within a simple environment, so that programmers can perform most if not all parts of the programming task in a coherent and uniform manner. This thesis tests the hypothesis that programmers do in fact derive some benefit from this - the simplification of the program and faster implementation of complex programs. The persistent language PS-algol is introduced and used to build: user-interface and compiler tools; a database application; some data modelling tools, both relational and semantic; a rapid prototyping system; an object-oriented language; and software support systems. In doing so, the thesis demonstrates the breadth of work which can be achieved using a Persistent Programming Language, and the ease with which these various projects can be implemented. Further, the thesis derives the beginnings of a methodology for using such a language and analyses how PS-algol could be improved. In doing so, the work aims to put the Persistent Programming Paradigm on a firm basis following significant use and experimentation