---

The Future Isn't What It Used To Be
or
Delphi Revisited

---

Alan Carrick

---

Whatever happened to the chart recorder, and the voice controlled analytical instrument?
Did the world chess champion really give best to a computer programme in 1987?
Why isn't crystal ball gazing so popular these days?

If the answer to all these questions (and many others) lies with the computer, how far did we foresee that in 1978?

Despite the computer, the future remains fascinating. Since the earliest times man has tried to predict the course of coming events, and the computer has given both precision and acceptability to prediction, from weather forecasts to the fractal nature of chaos itself. In March 1978, some ten years or so after I had built my first computer of sorts, it was still unusual to have unrestricted access to a general purpose computer for everyday laboratory tasks, and the notion of the personal computer was just beginning to take hold. PCs, that is IBM PCs themselves, were still in the future, at least as far as the world outside Yorktown Heights was concerned: the Berlin Wall had still more than a decade of existence left. Thus when, at a meeting of the British Mass Spectroscopy Group, I made a 15 minute unscheduled digression from the planned business of the meeting ("Capillary Column GCMS Techniques" - the digression was I think with the connivance of the session chairman) into the business of trying to predict the future, it was both regarded as reasonable to do so, and also reasonable to do it as a pencil and paper exercise, with results worked out and reported afterwards. How long, I wonder, would we tolerate such an exercise now, and how long would we wait for feed back in these days of Internet connectivity and instant response regardless of distance or time zone?

The idea of a "Delphic Poll" - for so the exercise was presented - admittedly stretching the definition, is as old as the oracle itself, which (who?) made its name by a consensus type forecast (there were three "full time" members of the Oracle team), predicting the future usually ambiguously, for politicians, statesmen, and those who could pay. The Rand Corporation developed and refined the technique in the 1960s, extending the application to normative as well as exploratory forecasting, but still with the same overall objectives as the inquirers of the original oracle - control of the future. The MSG 1978 exercise had a similar marketing "hidden agenda", but with some greater degree of openness than the strategic planning of Rand. Twenty questions were posed at the meeting and covered aspects of the design of analytical instrumentation of all types, but most of the questions had a fairly obvious bias towards the use of mass spectrometry as an analytical tool. They fell into roughly three categories: those concerned with the future of analytical techniques in general, those concerned specifically with instrumentation, and those looking to the future of technology as a whole. Some composite questions, of course, dealt with aspects of more than one of these categories at once. More than a hundred questionnaires were returned from the audience out of a recorded attendance of one hundred and twenty seven, a reasonably representative cross section of skills and experience.

Perhaps the most interesting point from a 1995 view is that nineteen out of the twenty postulated "events" were predicted to have taken place by now! Voice controlled instrumentation was the only "advance" predicted to occur, in terms of a median opinion, after 1995, and despite "intelligent" telephone inquiry systems such as those recently in the press for not being able to recognise an English accent as opposed to an American one, this prediction seems to fairly accurate. Of course the key to accuracy of prediction lies in the interpretation of the postulated technical capabilities and indeed of the questions themselves. Taking voice control as a case in point, clearly the technical capability exists, and has for some time. The question is whether the need exists, that is whether there is a market for such devices. One practical application of voice control was thought to be in electron microscopy, where the darkened rooms and multiplicity of controls surrounding instruments of ten years ago might make talking to your instrument seem sensible. But of course the key change is that SEM output devices now have screens which can be viewed in ambient light - computers in other words - and these are the norm, and so the need assumed by the question has been removed.

Although some attempt was made to take into account ambiguity and imprecision in the questions at the "processing" stage, both by the stylised graphical representation of the results and by use of simple statistical devices to focus on the consensus, it is remarkable just how many of the 1978 postulates which seemed reasonable then have been superseded. Instrument fault diagnosis, and particularly self diagnosis (Question 16), was indeed a hot topic: in more recent years, "Expert System" projects covering checks on the correct functioning of the instrument as a whole have been attempted. Both approaches have been superseded by simplicity - simpler instruments and more reliable instruments. The trend towards standardisation of methods and mechanisms and total quality management in instrument manufacture will surely remove the need for the kind of self correcting fault diagnosis envisaged by the original question.

Comparing the three categories of question, it is interesting to note that the area in which the participants' expertise was probably greatest was forecast not just more accurately, but more conservatively. People - or at least scientists, it seems - are more inclined to believe that technology in general will advance faster than it does - or should that be "faster than the marketplace requires?" Perhaps this trend could be laid as much at the door of "Star Trek" as of "Tomorrow's World". Prediction of general analytical procedures seems to have been if anything worse, in that advances came in quite different directions - in mass spectrometry, electron impact ionisation is still very much the norm (Question 9), but the newcomer threatening to take over is now laser desorption, which was not even mentioned in 1978. Legal battles are indeed won and lost on the results of computer supported analysis (Question 19), but today's unquestioned legal "big gun" (DNA profiling) is in an entirely different class from the simple chemical identification of GC-MS, although as many athletes know to their cost, this is powerful enough.

The future indeed isn't what it used to be. 1978's future looks on reflection somewhat quaint and naive compared with the future of today, but perhaps that was ever so. From a particular standpoint, one prediction was depressingly accurate. Question 17 suggested the unthinkable - that magnetic sector mass spectrometers would disappear altogether. Half of the participants thought that it would happen, probably in 1993 and be complete by 1995. In 1994 at least one company (not unknown to the author) shut down its magnetic sector business. Who says predictions don't work?

---

"Megatrends", John Naisbitt, Macdonald & Co., London, 1984 (ISBN 0 7088 2508 7)

"Profiles of the Future", Arthur C. Clarke, (Revised 1973), Victor Gollancz Ltd., London, (ISBN 0 330 23619 9)

"Prophecy and Progress", Krishnan Kumar, Allen Lane, London, 1978 (ISBN 0 7139 1146 8)

"The Third Wave", Alvin Toffler, William Collins Sons & Co., London, 1980 (ISBN 0 330 26337 4)

This document created and maintained by Acolyte Science
For comments, suggestions, or contributions,
E-mail to 100572.753@CompuServe.com
© Acolyte Science, 1995
Last Update, 10-Nov-95