A TALK TO THE TORONTO SECTION OF THE CANADIAN INFORMATION PROCESSING SOCIETY
SEPTEMBER 11TH, 1979
I was invited to make some comments about the computer industry as an objective outside observer. I am indeed a defrocked supplier of computer services. Therefore, while it is true that I no longer need to defend any particular part of the computer industry, I have found a group whose point of view should be put forward. I am now a user of in-house computers and computer services.
Users I have quickly found do not spend much time talking about the declining costs of cycles, the rapidly changing technology, the advantages of distributed computing or other matters that were a way of life for me for nearly twenty-five years. Users are concerned with such mundane matters as cost effectiveness, responsiveness to business needs and simply getting the job done.
You probably expect an old timer from the computer field to come back and berate all of you for not doing things the way he used to do them. In fact just the opposite is true. I am finding to my concern that we are still doing things exactly the way we used to do them and, believe me, the good old days were not that great.
This has led me to a series of observations about some fundamental problems that the computer industry still has which may limit its ability to expand in the decades to come.
I will build these observations around four general topics:
• people and their approach to computing problems
• money and where to look for it for data processing projects
• ideas and where to look for them
• the environment for a healthy computer industry.
PEOPLE AND PROBLEMS
The public service in Canada is more versatile than many give them credit for. I was amused a year or so ago when Dr. Peter Robinson was claiming that we were losing job opportunities in the computer field by having so much computer work processed in the United States or elsewhere. From the same office, Hans Brune had written a Paper indicating that the biggest problem facing the computer field was the lack of adequately trained people. If Hans was correct, one would wonder why we would care particularly if we borrowed some talent from elsewhere. In any case, both were concentrating on the people side of the computer industry which is certainly the correct place to look for any limiting factors.
The Silt Report produced by SHARE a couple of years ago also zeroed in on this problem. The lack of people productivity was apparently the major reason for IBM scrapping its next generation of machines a couple of years ago. As a user I can now assure you that it is a very real problem.
It is not just that there are not enough people to sustain a healthy long range growth rate in the computer industry. The people are simply not producing in an efficient manner. I will give you an example and then suggest some possible solutions to this problem.
When I arrived at Premier Cablevision I was advised that they had underway the development of what was called a Management Information System. I had plenty of other things to do than get involved with a system that I was informed was in the late stages of implementation. I had heard some disquieting rumours that in fact the system had taken longer to implement than had been expected and was rather seriously over budget. When I finally took a look at the system I was dismayed to find that about every classic error one could think of had been made. All the problems I thought we had cured years ago were still being perpetuated by a new generation. For example:
• I could not locate any well documented set of system specifications that had been agreed to by management and the data processing group.
• The system was being developed from scratch, although there were at least half a dozen systems of which I was aware that could be obtained on a package basis that would have been at least adequate.
• The system had been designed to handle everything rather than recognizing that a good system should aim at handling perhaps 95% of the transactions, with manual exception routines for those odd situations that are simply not worth automating.
• There appeared to be some fundamental lacks of understanding of the requirements of the business as opposed to the requirements of the computer system.
• The response time for this on-line system was proving to be woefully inadequate and management was being asked to add more horse power in the form of additional memory, disc drives, etc.
• It appeared that there had been only minimal attempts to straighten out the systems problems as opposed to attempting to automate whatever management presented to the computer group, e.g. we had all kinds of odd rates that had developed over many years. The correct initial approach would have been to try to rationalize the rate structure and then automate.
I am not singling out Premier because I know that at least one other major cable system went through exactly the same procedure and then ended up scrapping their entire system. When will we learn?
After at least a quarter of a century I am finding that we are much more knowledgeable in the computer field but not much smarter. Somehow we have missed an opportunity to pass on a lot of common sense that would save the next generation some of the agony described above.
MOVING TOWARD A SOLUTION
For years I pushed for some form of certification or other professional standards in the computer field. It is probably too late now as the number of practitioners has grown so large that one could easily end up with a quarter of a million grandfathers under some phase-in plan. I still, however, would like to recommend to the field that certified computer auditors of some kind should be available to industry. I have in mind professional people of proven academic and hands-on experience who could be brought in by potential users to audit the approach being proposed by their data processing divisions. This would cover all the usual needs of proper specifications, conversion planning, back-up, editing, documentation and other standard techniques.
I realise that some of this is available from the major consulting firms but, unlike accounting auditors, these people are often too indulgent with those who are paying their fees, and do not have the nerve to tell the company or the data processing division it is on the wrong track or is headed for trouble.
Secondly, I believe that we must be able to produce something equivalent to an engineering manual for data processing. Engineers over the years have at least laid down enough guidelines and rules of thumb that most bridges don’t fall down, (except in the Northwestern United States where in fact they do fall down fairly regularly!). I know that programming has always been thought of as an art, but surely by this time we could have turned it into more of a science.
Thirdly, we need to do a better job of training systems analysts. For some reason we seem to have concentrated on training technically competent programmers but still are not providing to young graduates enough understanding of business requirements. In fact we do not seem to be able to pass on much in the way of just commonsense approaches to mundane problems. Possibly the use of case studies taken from real business situations could help in some of our universities and community colleges.
We need to learn such simple things as, “You never quote on a job. You only quote on a quote,” i.e. the first step always is to ensure you have done your systems homework and then quote- on the time and cost of implementing such an approach. At least then the implementation estimates should be reasonably accurate. I could go on, but I think you see this problem from the user’s standpoint. We cannot leave computer systems development as an art. We cannot mass produce meaningful art and this will be a limiting factor on how fast the computer industry can develop.
A possible limiting factor in the future of the industry is the lack of available capital to automate. This may sound strange, knowing the absolute necessity of increasing the productivity of industry in Canada. Also with the increasing wage demands that we can expect in the next few years, the desire to automate should be very high.
The problem is that we are not necessarily tackling the right problems for our country. Canada is not and is not likely to be a major manufacturing force in world terms. We are a country with a predominance of relatively small businesses. We are further very oriented towards service industries.
If small, service-oriented industries are where people will be employed, then that’s where the efforts of the computer industry should go. I predict a trend in business towards both more big organizations and more small organizations. (Does that sound familiar? The same comment is made about CPU sizes.) The very large firms, e.g. banks, life insurance companies, etc. are now highly automated. The market here while still growing will be limited in the future. As growth in employment in these large industries lessens, the number of people employed in smaller industries or even craft industries will increase.
What do these smaller organizations need? Most of their needs are simple.
• They require standard, simple to use cash flow analyses.
• They need easy-to-use budgeting packages.
Small businesses, however, also need education on the use of such packages. And yet this in an expensive process given the relatively low value of some of the sales in this area. Surely with our experience in self-teach texts, coupled with video tape packages, simple education processes could be devised to train small business in the fundamentals of accounting and then provide them with the tools on an in-house computer or on a service bureau to allow them to implement such standard procedures. It is amazing the number of business failures that take place because of a lack of understanding of the most mundane business principles.
There is a huge market of tens of thousands of medium to small businesses in Canada. The money is there but we will only pry it loose if we stop just chasing the very large businesses in Canada.
Gordon Thompson of Bell Northern Research in a recent paper produced for the Institute for Research on Public Policy commented that he felt the biggest restraint on the development of the computer industry was our inability to view information in new roles. We are still looking at automation in terms of replacement, substitution or expense saving and are not really looking at new ways of doing new things.
The computer industry must be alert to the real needs of business and look for new ideas that will help the businessman in some relatively fundamental areas. It has occurred to me that for a forward looking industry we spend a lot of time developing systems that only look backwards.
A businessman can do very little about the past, unlike a newly elected politician who will immediately rewrite history to show what a miserable job his predecessors had done. Yet if you look at most of the current computer systems we develop, all we produce is historical data. There is very little in the way of trend analysis. Certainly unless a corporation is a very large one there is little in the way of corporate modelling or simulation of alternate futures the corporation could consider.
Such systems need not be complex, but you would be amazed at how few companies even plot in some graphic form the trends of the most fundamental aspects of their revenue and costs. They are often in trouble before they realize it.
There is nothing new about this suggested need, but I believe there are many other such needs in industry. We are not well equipped in the computer industry to research these needs. I suggest some joint projects between our universities and community colleges with the Association for Independent Business, local Boards of Trade or other such organizations to see if we can come up with the ideas needed to open new markets. A lack of new ideas could be a major constraint on the industry.
THE ENVIRONMENT IN CANADA
A final possible constraint may be less obvious to you than it is to me in the broadcasting and communications business. I am now working in a heavily regulated industry and can tell you with no hesitation that over regulation can be the most devastating constraint on any industry. As a company, Premier is now pouring money into the United States and elsewhere in areas such as Pay Television because we are constrained from using our normal entrepreneurial instincts to bring about such services in Canada. In the computer industry you are lucky as you are essentially unregulated, but I wince every time I see an electronics firm going to the Federal government asking for some kind of assistance, or a computer services firm going to ask for R&D money to develop new projects. When you ask the government for something there is a price you will pay. That price may well be increasing government involvement, and that price is very high.
WHERE TO FROM HERE
When a football team is not playing well the usual advice is to go back to the fundamentals of blocking and tackling. Much of my advice from a user standpoint is similar. The need to formalize what we already know how to do, and then make sure we follow our own fundamentals.
If the industry is to continue to develop rapidly in the years to come we must put behind us some of the still present old problems, lessen our obsession with new technology and increase our attention to what the user really needs and will pay for.