Dp departments and the micro - by G. Luxford
A user view - Dp departments and the micro - by G. Luxford published in an unknown magazine:
We have (magazine unknown) in the past published a number of articles about the impact of the micro and the response it requires from the dp department. Each of them was written by somebody intimately involved with computing. Here we publish an article written from the rather different slant of a user who believes that most dp departments are obstructive and unhelpful about micros, and need to make fundamental change to both their attitude and their role.
"The advent of low cost microcomputers has now made computing facilities available in areas where they were previously excluded. The concern of this article is where these low cost microcomputers are used in medium to large company organisations; in which computing facilities have been available for some considerable time in the form of large centralised facilities, normally operated by a data processing department. The point of view given is that of a committed microcomputer user who became totally frustrated by the limitations of the central computer facilities.
In his case, technical expertise was available to permit the use of microcomputers, but many other frustrated computer users are not so fortunate and will need much more sympathetic assistance if they are to realise the opportunities available.
There is, in theory, no real need to have microcomputers in a large company, since all computing needs could be met by the central services. The fact is that microcomputers are increasingly being seen in various departments of large organisations and there must be reasons for these changes.
Cost may be an important factor, but one would expect the new technology to have an equal impact on the large central machine, which should be able to offer more cost effective computing.
With the cost of computers themselves being so low, the cost of peripherals is now a dominant factor, such that if there has to be a departmental peripheral away from the central computer, then adding micro computing power now becomes a minor hardware cost.
It seems, though, that low hardware cost alone is not the only reason for the progress of the microcomputer. In many applications on a large central computer, the cost of program development and the cost of running programs is now very high, because the machines are just so large and clumsy to use. There is also the fact that microcomputers are no longer so small themselves. Many of the modern 100K microcomputers with 20 megabyte Winchester backing stores can swallow some quite hefty computing problems at an economic cost.
Another important factor is software development. Because the cost of microcomputer software can be spread over thousands of applications, very high quality general purpose programs are now available at a fraction of former costs. But even with the software, it is not just a matter of cost; microcomputer software is different from mainframe software.
The dynamics of microcomputer software development are such that the end result is just so much easier to use and more interactive than the mainframe equivalent. A classic example is the development of the electronic worksheet, VisiCalc being one of the first of these. With such software, very complex interactive problems can be solved in a fraction of the time it would otherwise take just to specify the problem, and that includes entering the instructions into the machine and formatting the printed output.
The display screens of microcomputers have graphical facilities, together with lower case letters. These are facilities not normally available from a mainframe computer terminal. This all helps to make the output far more interesting and meaningful to the user.
Problems with mainframes
My company started using microcomputers in Research and Development about two and a half years ago. This was because of the difficulty of obtaining satisfactory technical computing, despite the presence of the company's large mainframe computer just across the road. To be fair, the data processing department were very helpful, but it took months to install and debug one Fortran program of about 1500 cards. This program was already operational on another similar computer.
Life became particularly difficult when the introduction of the Fortran compiler, necessary for this scientific work, caused four major crashes of the IBM 370 at peak processing time, with the result that all computer facilities were lost to the company for some time on each occasion. As a result, all Fortran applications were relegated to evening shifts until the compiler bug was fixed some months later.
It was also found that when the central computer was in heavy demand, any peripheral applications, such as scientific work, were pushed to one side, with a three day wait on one occasion. This made sense in terms of company objectives, but it did not help the individuals expected to meet prescribed time scales.
In truth, the particular job in question could not have been run on any microcomputer then available, but it brought about a resolution to keep off the mainframe computer any application which could otherwise be implemented on a microcomputer.
Even with a background of computing on large mainframe computers and minicomputers, it took a year of intensive effort to get microcomputers into perspective. It was not just a case of learning a new computer language, but one of coming to terms with how to use the limited self contained computing power immediately to hand.
With such machines, there is no constraint, other than capacity, on what one can do. It was a liberation from the many restraints of centralised computing. To write a program, one just switched on and started using the computer as the planning media, in which each element of a program could easily be tested as it was developed.
The gulf between the concepts of microcomputers and large centralised computers is analogous to the difference between the private car and the public railway system. I would not care to forecast the future of large central computers, but one can but consider the history of any centralised service facilities once cost effective individual alternatives become available. People, being what they are, want to have control over their own situations and are prepared to go to great lengths to achieve this.
Having already made the transition to microcomputers, it is possible to see the problems facing those who still adhere to the concept of the large central computer for every application. One can but point out that a microcomputer of the current generation now offers power equivalent to the average central computer of only five years ago, but at a fraction of the cost and with far more effective software.
The difficulty facing many data processing departments seems to be that they have well established systems in which a great deal of money and effort has been invested. No data processing manager is likely to admit that much of the equipment or software he has is really no longer needed, and that it may be much more effective for many applications to give each department their own personal equipment.
Many users of central computer facilities are increasingly finding that their needs cannot be met at an economical cost and in a reasonable time scale. As a result, users are turning to whatever alternative may be available. Regrettably, instead of being given the help and support they need to set up an alternative option, which will reduce the unnecessary overload on the central facilities, the users' requests are often obstructed by the all powerful computer veto. This is normally vested with the data processing department by the directors; who neither understand or care about the essential differences between microcomputers and the large central facilities.
Lack of co-operation
What actually happens to a departmental manager when he goes to ask for assistance from the central computer system?
The first problem is that he must find the data processing department. It is probably tucked away in some obscure corner of the company behind some inconspicuous door. When the data processing sanctuary is located, there is then the problem of gaining entry through an array of magnetic locks and sealed doors to see the appropriate personage in the inner sanctum.
Having got to the person responsible, he is then expected to communicate fixed requirements in definitive technical terms. After putting the request, the response could be something like: "Well yes, we would love to help you, but you have come at an awkward time. You see we are a bit busy at the moment and we are right in the middle of the annual stock review. We are completely occupied with setting up facilities for the new section and you see we are very short of programmers and our very big computer is already running at 110 per cent capacity and the company won't sanction the £1 million needed for a bigger and better computer. Anyway, we must first sort out 99 other requests, which are far more important than your toilet roll stock records. Leave it with us, anyway, and we will see if we can get round to it next year, that is if something more important doesn't occur in the meantime".
So what does the manager with the toilet roll stock control problem do, up to his eyes in paper work, or desperately trying to reduce costs to a level decreed by his superiors? Perhaps he may go back to his central computer department with an alternative request:
"Look, I know you chaps are very busy with all those very important things you have to do, trying to run this monstrous computer, but I have heard of these things called micro-something or other, which everyone tells me can be very useful. Perhaps they would do my job for me and save your chaps a lot of effort. Perhaps you could give me just the tiniest bit of help to sort something out."
Well, what kind of response is such a request likely to get? "You don't want to bother with those toys. Look, leave such matters to the experts, we know that we are doing. We have 15 years experience at this business of baffling people like you with technical jargon. OK, so you may have to wait a while, but your problems are not really so important and so what if we lose a few customers or hire a few more people in the mean time?"
The next step
Now what does the departmental manager do about this toilet roll stock records? He still has his problem to solve and seems no nearer to finding a solution.
Well, there are a number of options open:
• Give up and carry on with the old system
• Call in an external expert
• Find a way to obtain an independent computer
• Get advice or assistance from another department which has already used microcomputers.
Regrettably, most managers just have to give up and carry on as before. There is no real need to worry - after all they did try and the experts said it couldn't be done. In any case, without the experience required or sympathetic support, there is little chance of success in a potentially hostile environment. He is probably wise not to press the matter any further.
Some managers may try to obtain external help, but without the sanction of the data processing department, they are on their own in shark infested waters. If they sink, no one will come to their rescue. Before following such a route, the wise manager would carefully study any guides available on the best route to take.
Those with previous computing experience and a technical background may stand a chance with a specific limited microcomputer application. It may be quite possible for them to go ahead, purchase a suitable computer and get on with the job. The problem is that as soon as any capital request mentioning "computer" appears for approval, it is diverted straight to the "computer experts" of the company in the data processing department, with the obvious consequences. Some companies seem to spend a fortune in time and effort trying to stop unauthorised expenditure of a few £s.
Given an embargo on any computing equipment and enough desperation, it can become a real challenge to defeat the bureaucracy of the company. With a computer called a "programmable desk-top calculator" or just a "typewriter" and with a printer called a "recorder" or "data logger" some way will be found to circumvent the system.
If a manager is inexperienced with computers but is determined to establish a microcomputer application, then he may be able to find someone else in the company who has already blazed a trail. He can be advised on how it was done and what the major pitfalls are. One big advantage is that the other manager will at least be sympathetic to the situation.
It is, of course, quite legitimate for a data processing department to say, "Why does anyone need microcomputers anyway?" Well, here are just a few of the reasons:
Cost effectiveness. Many effective computer applications can now be carried out on equipment which costs less than the price of connecting a dumb terminal to the central computer. Program development and standard software is very much cheaper.
Greater flexibility. With modern software, microcomputers can easily be adapted to required applications.
Very interactive. Gives immediate response and makes finding errors and making corrections much simpler.
Independence. Control over one’s own problems with no priority conflict with other departments.
Less daunting to use. More human in scale and within the department.
Easier to learn. An effective way to gain computer experience.
Simplicity of use. Much easier to implement an application compared with a large mainframe computer.
More accessible. Not remote, within physical and mental reach.
Better reliability. Mass production of low-cost units means that equipment must be reliable. Faulty units are easily exchanged with "off the shelf" replacements or in-house spares.
Transportable Can be moved around particularly if on a trolley.
Data collection. Low cost microcomputers can be directly connected to data collection equipment with little risk to the company system.
Requires only a normal electricity supply - if necessary, a voltage regulator.
Costs £3,000 to £10,000 for an effective system.
Low cost combined with flexibility of application means the cost is quickly recovered.
Low financial risk:
Low capital cost with a gradual transition from manual methods and adaptability of the equipment makes risks of capital very low and minimises disruption of departmental activity.
Immediate delivery. Equipment is available from stock for most items.
Getting microcomputers in by the back door is obviously not the best way to progress. This will inevitably lead to an incoherent and unplanned hotchpotch of miscellaneous machines. But, given no other way, that is what will happen.
It is essential that, in any large company, there is recognition of the fact that the era of the centralised computer monopoly is over. The large centralised system simply can no longer offer all the facilities required for modern computing needs.
Computing requirements are now just too diffuse and diverse to justify the cost and complexity of a large central machine for every application. Personal computers are now so powerful and so cheap and easy to use that it is only common sense to use them for suitable applications.
Change of attitude
The first and most important requirement is acceptance by the data processing department that microcomputers have a useful role to play as both stand-alone facilities and as intelligent links to the central computer.
The data processing department must dispel the air of mystery surrounding computing. They must open up the computing world to the rest of the company, as far as is practical. It possible, the data processing department should be moved into the centre of the company, so there is much more contact with other departments and the problems they face. The data processing manager should be accessible to anyone who needs to see him, in the same way as any other manager.
There must be a much greater effort on the part of data processing departments to see computing requirements from the perspective of the user. If this requires that computer programmers go and work in Sales, or any other department, for a limited period, then so be it. In this way, there will be greater opportunity for everyone to exchange ideas. Computing will have to be much more user orientated, responding to user requirements, rather than being arranged for the convenience of the data processing department.
If the data processing department is to gain credibility as the company "experts" on microcomputers, then they must be seen to be applying such facilities in their own department to improve their own efficiency. Before they can teach others, they must first learn themselves; learn how to use microcomputers effectively and learn to understand the essential differences between "driving a passenger train and "driving a private car".
There must be a realisation that microcomputers are different to mainframe computers and will fill a different role in the company computer facilities. The microcomputer function will be to absorb much of the trivia of computing, passing on to the central computer only information of common company interest.
The worst mistake would be to treat microcomputers as small mainframe computers. Any data processing member that talks of timesharing on a microcomputer should promptly be removed from any position concerned with equipment selection! They have obviously not understood the philosophy required.
Change of role
The data processing department must accept that their role is changing. We will always need the large central computer as the keeper, and administrator, of the primary company data store and for the large well defined processing tasks in which high speed data coordination is required. But a company will also need the small microcomputer for the local "domestic" situation and as an intelligent link into the central computer facilities.
If any data processing department thinks that their microcomputer responsibilities extend no further than a decree on which machines departments may use, then they must think again, that is if they wish to retain responsibility for the computing facilities of the company.
Companies will need a section of trained microcomputer people willing and able to go out into the company and sympathetically help departments to select, justify, purchase and then set up "domestic" computing facilities. It will take up time and effort to "hold the hand" of each department in turn and lead them through the difficult early stages of setting up new microcomputer facilities, for which, in the end, the department itself must be responsible.
In any department there is always at least one person quite capable of taking on the responsibility for the microcomputer, providing he is given the right encouragement and help. The people being trained will not be computer experts and so technical jargon will have to be avoided.
While the users may not be computer experts, their ability to use microcomputers should not be underestimated. They will have the advantage of not being constrained by previous conceptual limits and they will have the incentive and knowledge required to implement their particular requirements.
There may be fear and even resentment by some staff of the introduction of computers into their department, so good personal relations will be important. There will be no air-conditioned sanctuary in which the new breed of computer specialist can hide away from dissatisfied users.
If there are already people in the company with appreciable microcomputer experience, then any previous animosity between them and the data processing department should be set aside and their experience harnessed in setting up any new microcomputer facilities.
The data processing department should not apply unrealistic standards to microcomputer applications. Risks must be sensibly balanced against commercial reality. Many users are not critically concerned with absolute accuracy, but are far more concerned about obtaining a reasonable evaluation of unexpected situations very quickly.
Users will expect their systems to be flexible and able to adapt to changing circumstances, not least of all because as they learn they will think of new ways of applying their equipment.
A company and its data processing department must realise that introducing microcomputers will absorb considerable amounts of both time and effort. Adequate arrangements must therefore be made for both of these resources. If necessary, recruitment and retraining of computer staff must be undertaken, so they can then in turn go into the company to set up the local systems and train the users.
This article has attempted to show that companies are turning to microcomputers not just as a cost effective alternative to the central computer facilities, but also because the microcomputer has many inherent advantages. Apart from the positive benefits of microcomputers, they are also being taken up because of frustration with the poor service obtained from many central computer facilities for ad hoc problems.
The monopoly of control by a data processing department over all computing facilities is a stifling influence on the effective development of micro-computer applications.
Data processing departments will have to change their views about microcomputers or hand over the responsibility to someone else. If they are serious about the use of microcomputers, they must demonstrate a more open attitude to computing, using microcomputers in their own applications before going out to set up the required facilities in other departments."