When digital computers first became available many of us expected that they would have an early and significant impact on engineering design. This did not happen, and the reasons why it did not are worth examining. For one thing, the early computers were not nearly large enough; engineers do not build things out of spheres and parallelograms as mathematicians do, and quite a lot of storage space is needed to describe a typical problem in engineering design. Unfortunately, the big computers when they came were so expensive that it was not considered economic to allow users to handle them personally, and batch-processing techniques were introduced. The result was to create a barrier between the computer and the design engineer, and to make it impossible for him to get results of any kind without a delay amounting to a few hours at the very best, and often to much more. Emphasis in fact, was put on the efficiency with which the central processor of the computer was used and no regard at all was paid to the efficiency with which the users—in this case programmers and design engineers—worked. We are on the threshold of a development which, there is every reason to hope, will change the situation radically.

The engineering designer may not always be aware of the paramount importance of being able to sketch his ideas as a means of communicating his plans.

An example illustrates the point that drawings, graphs or diagrams convey concepts more easily than a table of figures. Table I gives a series of X and Y values describing a well-known contour, but it is not easy to determine the shape except by plotting the points. One can appreciate at a glance at Fig. 1, however, that it is a well-known shape.

Although the drawing is very graphic, the term graphic will also be taken to include alphabetical and numerical records for the purpose of this paper.

The up-to-date engineer or scientist would not challenge the growing importance of the modern large-scale digital computer in research and development. Yet, while we all agree that computers are important and useful, we disagree from one discipline to another and often within the same discipline on just what the computer contributes or can possibly contribute to research and development endeavours. The difference in approach to computer usage is more one of philosophy than of technique. Indications are given from our recent work that some philosophies of computer utilisation may be more fruitful than others. The availability of the large-scale digital computer has not only facilitated, but directly inspired and shaped the matrix methods for analysis of continuous media and structures, both in the linear and non-linear regimes. Without the availability of a modern computer all this work is literally meaningless.

Some of the unusual factors influencing the choice of digital computer control systems for jet engines are discussed first. The conditions differ in various practical respects from conventional process control applications. Having several identical plants in a typical aircraft and the need for flexibility to keep pace with engine development are of particular importance. Approaching digital control with the traditional process control background may be restrictive in selecting potential applications for the technique.

Techniques which can be used with computers to simplify and speed up three aspects of the realisation of a practical system are described briefly. The first of these is the measurement of the actual plant characteristics needed in the analysis of the control system requirements. The second is a method of obtaining good response in the synthesis of a non-linear control function. The third is a technique for programme writing and for simulation used in the verification of the control system behaviour.

The techniques described are not in themselves necessarily confined to use with digital control systems and should have value in general use.

In this final presentation in this symposium I hope to relate some of the present and future developments in computer applications to the everyday world of an engineering design office. The first engineering industry to make extensive use of the computer in design was, I believe, the aircraft industry. In the late 1940's and early 50's digital and analogue computers revolutionised the treatment of aeroelasticity and since that time the spread of computer methods of analysis and design has been rapid. Until recently these developments have been erratic and badly co-ordinated and even today there is no aircraft firm to my knowledge, either in this country or the USA, which realises more than a small fraction of the potential efficiency in design which could be provided by the computers at its disposal. In the past two or three years this fact has become more widely appreciated and I believe that the really significant steps forward are being taken at this moment. The previous papers have shown clearly the possible extent of the computer revolution in the easily foreseeable future. I hope to illustrate, through a specific example, the influence of this revolution on the activities and on the men employed in an aircraft design office.

The nature of handling criteria and their uses in relation to the prediction of aircraft handling characteristics are discussed. A brief history is given of the development of ideas on handling, and a survey is made of criteria for modern aircraft, looking particularly at the physical factors affecting the pilot, as an indication of the basis of the criteria. The novel handling features arising with new aircraft species like slender wings and VTOL are considered. The scope of flight simulators and variable stability aircraft for systematic study of handling is discussed and the potentialities of theoretical study by means of servo analysis techniques considered. Finally, possible lines of development in relation to the assessment of aircraft handling characteristics are suggested.

The Centennial Year is over, and the activities of the Society are now projecting towards the second century of its existence. A number of new Committees have been formed, notably the Activities Committee and the Management Studies Group Steering Committee, and it is possible that during the on-coming year one or more additions to the number of Groups will be made.

During 1961 preliminary consideration was given by the Weapons Research Establishment to the possibility of measuring the structure of the upper atmosphere by pressure probes on sounding rockets. A similar technique had been used with success in the USA. It was proposed that velocity be measured and that an incidence meter be used to measure pitot and static pressures, from which atmospheric pressure and density could be calculated. There was available at the time comprehensive calibration information on hemispherical-headed incidence meters at Mach numbers below 3. Since the sounding rocket would be flying within the Mach number range 3 to 6 there was a need to extend the existing calibrations to higher Mach numbers. At first it seemed that the results of Baer would do this, but a closer examination showed that his results did not agree with the incidence meter calibrations.

Even at small angles of incidence, the flow separates from the sharp leading edges of a slender wing. These flow separations usually lead to the formation of two free vortex layers, joined to the leading edges of the wing and rolling up to form spiral-shaped vortex sheets above the upper surface of the wing. This vortex formation is illustrated schematically in Fig. 1. The streamlines on the vortex sheet follow helical paths. Smoke injected near the wing apex for flow visualisation remains concentrated close to the axis of the core of the vortex sheet.