In Search of Techne

Ch 3.4: The Computer in the 1970s

(c) Roy Johnston 2006

(comments to rjtechne@iol.ie)

Introduction

The principal innovative productive service covered in this chapter is 'computing'; the present-day boom in mini- and micro-computing applications is foreshadowed, and two of the prime participants identified (Digital and Wang). There is an excursion into mainframe networking, with a political dimension (in the all-Ireland context). The use of the computer in industrial process control is touched upon.

This is an area which in the present Irish context has begun to 'take off'. I get the impression that if I were writing the column now I would have considerably more to say about innovative computer applications originating in Ireland. (RJ circa 1985.)

Note that some of the footnotes in this section need to be tidied up, due to an earlier drafting having suffered loss of content. RJ 30/05/06.)

October 10 1973

Two weeks ago I had occasion to visit Paris to look at the SICOB exhibition which caters for those interested in the technology of information processing. In the company of some financial writers I had the chance to look at a new system being launched by IBM, directed at the banking and financial market.

I also took the opportunity of looking at some visual display equipment associated with mini-computers, or 'intelligent terminals' as they are known in the trade jargon. These units come very cheaply, the problem is how to get people to use them imaginatively and to leap outside the boundaries of current practice with the new power which is available.

To return to the IBM system.. those who have observed airline 'real-time reservations systems'(1) at work... will immediately see the relevance of the extension of this principle into banking.

In the former system the sales-clerk has a keyboard and a screen. Flights can be displayed. reservations made and cancelled, using a central file which is instantly updated for each transaction.

This is worthwhile especially for short-haul airlines, as the profit is in the marginal seat. If there were a lag in recording a cancellation, a new booking might be wrongly rejected. In translating this principle into banking, however, IBM have taken into account the dissimilarities. In banking there is a basic three-day lag in clearing-house activity. Therefore there is no advantage in having a centralised 'instant update'(2). Also, the base of the credit-structure is the local knowledge of the bank-manager. So they have devised a small-scale system, based on the mini-computer, which operates autonomously at branch level. Thus the manager is not downgraded by over-centralisation.

The terminal enables the bank-clerk to answer within seconds any question relating to your account condition(3)... Various specialised terminals enable statements to be printed, cash to be dispensed, savings books to be updated etc...

All this is, of course, labour-saving. Another way of looking at it is that it enables more transactions to be carried out per unit of staff time, liberating staff for higher things, like- doing discounted cash flow analyses of investment projects for customers... The unions will need to watch to see that there is no downgrading. The question of job enrichment vs de-skilling needs close scrutiny.

The price per installation ranges from about £5000 to about £50,000, depending on the size of the branch, which looks cheap as computer systems go....

There are also leasing arrangements. The latter are to be discouraged on ethical grounds, as I understand that leased equipment at the end of its 'useful life' (ie if it has been superseded by a new generation of equipment) is withdrawn from the market and physically smashed(4).

Bought equipment, on the other. hand, can be sold to a broker who will find a home for it; even if the manufacturer no longer supports it, ingenious people in colleges of technology, even in remote and undeveloped parts of the world, can happily keep old computers going, learning the black arts of computer technology in the process. Smashing good equipment increases the cost of this learning process, squanders the world's non-renewable resources and leaves us the poorer in available assets.

Returning to the IBM banking system... it appears likely that the small size of the package will give IBM a keen marketing advantage. A bank can put in one or two such systems on a trial basis, without having to decide to standardise on one supplier.

On the other hand, the position of the customer is strengthened, in that the firm has to learn to deal with a variety of suppliers; if the basic unit of the purchase decision is small, the firm need not be at the mercy of a single supplier. Know-how 'in house' can be built up about a range of manufacturers and equipment...

This impression ofa buyers' market is strengthened by the SICOB exhibition itself. The outstanding impression is the wide availability, high performance and low cost of small interactive systems with visual displays.

I took note of two such systems, the Digital and the Wang. The former was on display in the form of a system (priced at less than £10,000) set up with a simulation of a moon-landing, for potential buyers to play with. One could control, by means of a 'light pen' acting on the screen, the thrust and orientation of one's space-ship, while the machine kept track of the residual fuel, and displayed your position and orientation relative to a rugged landscape. This had a queue of enthusiasts watching it. The implication is that if you can have a toy enabling you to emulate at this level of sophistication on your desk, it should not be too difficult to simulate the future course of your business in the economic landscape. The bottleneck here is in the provision of people who know how to adapt programming sophistication in this direction.

The Wang, at about £4000, has a keyboard with the operators of a high-level language ('Basic', looked down on by some as a 'poor man's Fortran', but in fact a powerful language in its own right) available on it directly (5). In other words, instead of a low-level machine language and an elaborate interpretation/compilation procedure for interacting with a high-level language, they have made the machine 'think' in a high-level language to begin with. This gives it a considerable reserve of computing power; the whole job literally sits on the desk.

Thus for a small firm the initial step into computing is now quite small, with a wide choice of manufacturer. The small unit can usually be interfaced with a large 'mainframe' subsequently, should the need arise...

The problem now is how to get people to use this power imaginatively, to understand qualitatively new abilities that are given to them by this remarkable equipment. If only the manufactureres would stop, for a while, developing new stuff, and give us users a chance to adapt our thinking and practice(6)...

The tragedy is that at present most of the applications of these techniques are either trivial or anti-social(7).

May 1 1974.

Last week Dr F J Smith, of the Queens University Belfast computer centre, spoke to a meeting of the Scientific Computing Group of the Irish Computer Society (known as the SCICOM Group). Dr Smith made a convincing case for establishing a single large computer centre to supply a computing utility to the whole of Ireland, supplying a number of (particularly university and college) computer centres with back-up capacity using data transmission lines. It would be able to do large and/or long jobs cheaply, support large files, do jobs requiring special peripherals or special software packages.

It would be linked into an international large-scale data-transmission utility network(8), making available (for example) up-to-date economic statistics of prospective national markets to exporters, and enabling jobs to be transferred from one centre to another by mutual arrangement, taking up slack capacity.

Such a centre would enable software to be developed for export; at present, Dr Smith stressed, we export the graduates to foreign software-houses and then buy back the software.

The present data-link structure is unco-ordinated; it shows the traditional features of all Irish phenomena as they have developed under the influence of the Partition structure: there are six links from Belfast to Britain, four from Dublin to Britain, six from the provinces of the Republic to Dublin, one from Dublin to Belfast and one from Cork to Britain.

Dr Smith would replace this system with a co-ordinated system centred round a maxi-computer centre somwhere between Dublin and Belfast (he suggests Rostrevor for aesthetic reasons); this would have high-speed lines noth-south connecting the two major Irish centres, and high-speed lines east-west connecting into the world network. Provincial centres would connect in by telephone lines of appropriate capacity.

The stimulus which made the Northern people take this up so strongly was....the production of a British plan for maxi-computer communication centres which simply forgot about Northern Ireland, because it was basically too small. Together, however, the two parts of Ireland have about as much computing business as one of the proposed British maxi-regions, so that an all-Ireland project makes good economic and technical sense(15). The sums of money involved are of the order of a capital cost of 2.5 million and an annual cost of half a million. This would be considerably cheaper than to allow each university centre independently to upgrade.... all would upgrade to maxi-status with the aid of the network.

I also spoke, on the same occasion, to Bob McLaughlin, who is director of a small firm spun off by the QUB computer people. This arose out of the spare-time activities of some electronic engineers, who hooked up an IBM 1130 to a Technicon autoanalyser and a Coulter counter, giving an automatic analytical system for blood-samples. They are now a limited company, with the blessing of the QUB authorities, and are projecting a £100,000 turnover this year, employing 16 people and supplying the UK market.

It is beginning to break through to the university authorities that this type of activity is not disreputable.

The IBM 1130 is, in a sense, a precursor of the present generation of mini-computers. Once the software is well-defined, it becomes possible to do the job with a 'dedicated micro' at a fraction of the cost. One sacrifices flexibility of programming in return for low cost and speed.

The original computers of the 1940s had wired-in programmes; the wheel has turned full circle.

February 25 1975

The idea of computers communicating with each other over the telephone not so long ago was sci-fic; now it is commonplace. Nor is it just technical fireworks for the enthusiasts, it is economically interesting, especially for firms which, in the early days of computers, bought one for a status symbol and then found that they had an expensive department requiring highly-paid and mobile staff. It is also of interest for firms which up to now have been bureau-users.

The extent of the interest was displayed at a one-day conference on February 8 organised by the SCICOM Group. This was originally planned for an attendance of about 60 people; within two days...they had to close it off at over 100 participants. Most were from government, local authority, university and applied research institutes (in roughly equal numbers); there were however only a handful of industrial users, and as few as three people from the Colleges of Technology, illustrating the Cinderella status of the latter.. The Regional Colleges, which are supposed to be training the technicians and technologists of the future, only have access to computing in two cases out of eight.

The common ground among the people attending was the problem of dealing with a multi-centred system, as seen by large-scale computer-users.

The Northern Ireland contingent were mostly university people and constituted nearly half of the university participation.

The seminar itself consisted of an opening session covering the basic technology of communications between computers, followed by an outline of the experience of a major user (in this case the Revenue Commission), followed by a session in which spokesmen of the post offices of the Republic and the UK gave their angles on the matter.

In the lunch-interval there was time to inspect an exhibition of equipment which had been hooked up to parent computers elsewhere using the St Martins House telephones.... The technical session was dominated by the QUB people.... Dr FJ Smith... Bob McLaughlin and Dr FC Monds, who outlined the non-post-office options which exist for communications over short and medium distances (infra-red, lasers, radio)......

Dr IN Hooton of Harwell (originally the Atomic Energy research centre in Britain, now a general-purpose applied-science research institute) outlined the principles of CAMAC, a standardised code, instigated by organised users, whereby items of equipment from various manufacturers can be made to talk to each other, despite the desire of manufacturers to impose their own hardware for all purposes by the use of private codes. A competent user of equipment can therefore build a system with items from various manufacturers, provided of course he is able to maintain the system himself. CAMAC has been developed by 19 major national centres, in the USA, the UK, the USSR and elsewhere. It is in process of becoming an EEC standard.

...Mr T Mulhern of the Revenue Commissioneers outlined the background and status of the revenue system, which files data on over a million people, and can directly access personal tax files from any one of 13 inspectors' offices from one of 21 VDUs, each of which is equipped with a security cassette which limits access from a particular VDU to a particular area of the file..... The system handles up to 2.5 enquiries per second....

The post-office people.... outlined their plans rather tentatively; it does not yet amount to big business for them (2% of revenue in the UK, 0.1% of the users in the Republic, regrettably non-comparable statistics). The main growth is in 200-baud lines for use with VDUs. The rental for a 100 km 200-baud line is 440 pounds per annum. Thus to rent a line to a friendly mainframe, which might spend 1% of its time looking at your problems, is a relatively cheap operation compared to owning your own hardware.

Exhibitors included the Dublin County Council (which is on-line to an ICL 1900 at the Sugar Co.), Riomhaire Teo (which has a large Sperry Univac at Furbo near Galway, with high-speed lines to Dublin and Manchester), Honeywell, UCD, QUB and the IIRS. The QUB people had a radio-link to home-base. Others were using acoustic couplers into the phone system.

Currently the IIRS is 'bench-marking', ie shopping around measuring the performance of a standard programme on various competitive systems.

Riomhaire Teo is sponsored by Gaeltarra Eireann; I understand that it has got some business with the educational authorities in the UK for on-line applications in schools.....

October 28 1975

Digital, who have a major production unit in Galway, have supplied a PDP 11/45 to service an international research-orientated computer network in Vienna.

The International Institute for Applied Systems Analysis (IIASA) was set up on 1972 as a result of efforts by the late Lyndon Johnson in the field of east-west detente.

It consists of 70 scientists from the USA and USSR working together for suggest guidelines for the solution of world-size problems (eg resources, especially energy, pollution and the population explosion). The head of the Institute is Dr A V Butrimenko,of the USSR Academy of Sciences.

November 18 1975

Coras Tractala have a service available to exporters which is interesting.....as an example of a type of system which could usefully be reproduced in other contexts.

A firm wishing to use the system registers with the Market Research Division....and supplies a profile defining its areas of interest, both as regards products and markets.

The service then supplies a weekly bulletin matching the firm's profile, culled from the world literature. This is, in fact, a personalised newspaper service. It has replaced a bulletin that CTT used to produce which had got so big that no-one would read it.

How is it done? Not surprisingly, there lurks a computer behind the scenes, in this case that of Cara, the Aer Lingus subsidiary. The system design was done by Cara, in close association with CTT personnel. The design is good because it allows for the use of human intelligence in an area where that commodity is superior to the stupid efficiency of the computer. A pure computer system would have loaded the files with garbage. In the Cara/CTT system, the weekly updating of the files is done by sectoral specialists in CTT, each of whom filters the literature for material of interest to the aggregated profile of all the firms in the sector. This is a typical 'judgment and experience' type of job.

Once on file, however, the material can be filtered by the computer using a simple logical net based on the profile of each firm. A statistical abstract is also kept, giving the amount of activity by firm, by product, by market etc. This enables adaptive feedback to take place between CTT and the firms involved. Profiles can regularly be revised in order to strike the right rate of flow of relevant information.

This type of information system is likely to become much more general in the business world, ultimately replacing directories and other printed look-up devices with annual update. It is already commonplace in science and technology in the 'technical information' field.

November 25 1975

In a previous feature(16) I have described the printed circuit production process... one of the steps was the translation of an electronic circuit into a board layout. In Craigavon this step was sub-contracted out to another specialist unit in the Andus group.

It is feasible for an independent firm to establish itself in this specialist field; it is a high-technology service involving conversational graphics with an appropriate computer. Software has been developed specially for this purpose, within the generic field of 'computer-aided design'.

A leading firm in this field is Redec Software ltd, in Tewkesbury, England. From this a specialist subsidiary has emerged, known as Cybergraphics; this is registered in Ireland as Derritron Teo.

It is located, courageously, in Ballyferriter, Co Kerry. The managing director is RJ Hewlett, who was previously with Redec specialising in interactive graphics development, on the technical side. The financial director is BH Dando, who looks after the Redec interest, and the chairman is Mr P Hegard, who is Norwegian. The Gaeltarra interest is looked after by MJ MaNamara..

This firm will be a relatively small employer of high-grade labour. ....Its main problem, I suggest, will be in holding high-grade staff in a 'culture shock' situation. Perhaps there are by now enough technically competent people around who would know how to relate to one of the few remaining viable Gaeltacht communities without disrupting it irrevocably, or without themselves being frozen out. The key question will be how they face the problem of educating their children...(17).

January 13 1976

Datagraphic Europe contends that a direct linkage from computer to microfilm is here to stay, and backs this up by extensive European training and service facilities.

Datagraphic Europe is a subsidiary of a US parent company in San Diego, California. The technology involved is attractive, in that it involves a break from the biggest single source of computer down-time, the line-printer, with its mechanical wear and tear. An optical system which reduces a VDU display to 105mm microfilm involves one mechanical movement per frame, rather than one per line; the movements are in the tradition of the well-tried cine-camera technology. A data-transfer speed of 3.7 pages per second is claimed for the tape-to-film unit.

While this represents a breakthrough in data-handling technology, and can be very useful if integrated into a well-designed system, there is a danger that it will become a 'technological fix' for data-systems which are overloaded with paper, most of which is printed with low-level junk. This will be replaced by boxes of microfilm, also loaded with junk, which is rarely if ever referenced. The problem is swept under the carpet.

Few computer system designers have faced the problem of how to use the computer intelligently, to boil down the masses of low-grade data into relevant abstracted and structured information, which matches the perception-patterns of management.

However, there are many applications such as banks, insurance companies etc, where there exist massive archiving problems. In such cases a direct VDU to microfilm link would make a big hole in the paper and storage bill.

May 11 1976

The SCICOM Group ran a conference on May 1 which looked at the prospects for Ireland developing as a source of exportable computer systems (hardware and software).

On the high-technology frontier, Frank Imbusch (UCG Physics Department) reviewed the potential of optical communications systems, which the combination of light-emitting diodes (LEDs) and low-loss glass fibre optics has rendered interesting. Performances of the order of gigabits per second over distances of the order of kilometers are now feasible, with up to 50 gbits per second for short distances, using single-mode fibres. This level of performance may become an important means of matching the development of integrated circuits towards micro-miniaturisation. The performance of the apocryphal 'computer in a matchbox' is now limited more by the problems of interconnecting between the elements, rather than by the integrated circuits themselves. Fibre-optical systems, as well as being compact, are immune from pick-up of stray electrical signals.

The opportunities for technological entrepreneurship in Ireland are not in fibre-optics itself, but in the development of imaginative applications in useful systems, making use of available hardware....

Nippon Electric chose to use this event to announce their....4K ROM, which occupies a 9mm square and consumes 400mW.... Also Mr V Clay of Digital outlined the technical tricks whereby the PDP 11/70 makes optimal use of storage, giving the impression of a main memory of the order of megabytes...

....R J Hewlitt (Cybergraphics) is offering a computer-aided design service for PCBs..... PJ Stenning of Qeleq (Dundalk)(6) gave a cri-de coeur from the systems development industry: will the producers of devices please slow down and give users a chance to develop marketable systems around their devices, without being leapfrogged!

Bob McLoughlin... outlined....basic medical data-reporting systems.... there is also a profitable side-line in election analysis. In a situation where people don't know what questions to ask, he suggests that one should offer an unstructured data-bank with fast access. Patterns of questions will then define themselves, suggesting subsequent system development steps....

Tom McGovern (System Dynamics, the only Irish software-house) outlined the history of the firm(18) and its current problems: it costs of the order of £100,000 to bring a software package to the market. The snag is lag-time in awareness: by the time they have identified a market, someone else is already in.... If the home market were more accessible, it would help. Decision-makers in Ireland tend not to believe that an Irish software-house can deliver; they depend therefore on the mainframe suppliers for software packages, or on packages developed abroad(19).

Subsequent to the conference I had the opportunity of looking at a financial planning package developed by David Algeo (Craig Gardner). This provides a flexible, conversational-mode budget development procedure, adaptable to the needs of a specific company....

The SCICOM Group intends to draft a memorandum to the Minister suggesting modes of possible State support for a computer systems industry to develop....within the interstices of the (mainly foreign-owned) hardware and component industry....

July 13 1976

(Arising from the May 1 SCICOM seminar)....a working group was set up for the purpose of formulating proposals for the Government as to how best to promote an indigenous computer systems industry.....specialising in software...using other peoples' hardware.

The world software market is worth £6B annually and is growing at 25% per annum. Independent software houses employing 1000 people or less account for 5% of the market, but this share is growing at 40% annually.

The budget of a software-house would be of the order of 50% salaries, 30% marketing and overheads and 20% support facilities, including computers. There already exists a pool of skilled personnel in Ireland.

Obstacles to the development of a native software industry are:

1. Lack of credibility in the home market; users prefer what they consider to be the safety of reliance on major foreign firms, despite the experience that they send junior personnel to deal with what they regard as a minor peripheral market.

2. Lack of Government contracts: there is no Government policy in the matter.

3. Cost of access to major markets: this is a geographical factor which could be overcome by a travel subsidy.

4. Lack of feedback from EEC policy committees: Ireland is entitled to a share of EEC software contracts; the civil servants who work on these committees are seldom alive to the needs of Irish specialist contractors.. There are no channels whereby such civil servants can be briefed, apart from some which are beginning to develop under the stimulus of the NSC, initially in the fields of energy and environment... Most EEC committees however are serviced by career civil servants(20) who are technologically illiterate, and innocent of any commercial entrepreneurship in the national interest, unlike their continental counterparts, who bat consistently on the side of the major European-based multinationals.

5. Lack of proper financial support: this is currently directed by the IDA towards training grants and accommodation, thus favouring foreign firms with established markets, rather than towards marketing, which is what Irish software houses need most.

Subsequent to the seminar, the working group (which included a strong and competent Belfast component) came up with a series of recommendations:

*The Government should commit itself to an active encouragement of an indigenous Irish computer industry, and formulate policies to ensure its growth.

*A major survey of the present Irish computer industry should be made, and its potential in the world market assessed.

*The Government should contract, whenever possible, its computer systems from Irish companies.

*The fact that the software industry produces a 'product' should be recognised, and this product given the same status as traditional manufactured goods.

*The consultancy side of the computer industry should be given the same status as engineering consultancy. In particular, the contracting of computer consultants abroad should qualify for the same grants and tax reliefs as consulting engineers.

*Grants in the form of venture capital are most appropriate to the industry.

*In recognition of the fact that two to three years may elapse between the award of a contract to a software company and payment, and in order to increase the credibility of the company, the Government should underwrite viable Irish companies for the duration of a major contract.

*The marketing of products of the Irish computer industry should be supported.

*Co-operation and joint development of products between the universities and industry should be encouraged.

*There should be cross-border co-operation, particularly on contracts involving large sums of money, and which would benefit from the combined expertise of specialists working in both parts of the country.

*To ensure an adequate supply of talent to such an industry, the Government should make a start by commissioning a report dealing with computer-related subjects in the educational system.

July 20 1976

Most people in business are aware of the computer as a device which handles payroll, stock control and other routine chores. Few are alive to its use in planning......

This field still counts as a development area, and is being explored by IBM in its scientific centres, under the general heading 'advanced computer applications'. The centre at Peterlee, Co Durham, is one of 12 such, which are scattered throughout the world. Each centre contains some 10 to 30 IBM scientists, who form the nuclei of interdisciplinary teams involving researchers from outside bodies, including universities. Typically, a university staff member might spend a year at an IBM centre on an IBM Fellowship.

Typical 'advanced applications' developments are: how to stop Venice from sinking (at the Venice centre); air pollution (Palo Alto, California); medical X-ray image enhancement (Heidelberg); medical diagnostic systems, water resource management (Haifa). There are other centres at Cambridge (Mass.), Los Angeles, Madrid, Pisa, Bari, Tokyo and Mexico City.

The Peterlee centre currently is specialising in systems of interest to local and regional authorities, as well as systems of interest to second-level education. They are also going one step more basic, and attempting to understand formally the process whereby researchers build computer-based models of reality.

Work initiated at Peterlee and alread in current use includes:

1. A financial planning model for the Borough of Teeside.

2. A data-base system to enable the Swedish Government to study the effects of nuclear power-stations on Baltic-sea fish.

3. An urban management system for the Greater London Council.

4. An interactive career guidance system for schools.

Dr Michael Dowsey has developed what amounts to a set of management games for use in schools, in the age-group 11 to 16. These are implemented with a computer-terminal located in the school. Of the three games developed, the most exciting is 'XPLOR', the North-sea oil and gas exploration game. We ran this at the seminar, the participants dividing up into competing teams for the purpose.

Each team consists of a chairman, financial controller, scientist and rig manager. The chairman leads the team, helping it to decide (a) where to drill (b) what sort of rig to use (c) whether to hire or buy the rig.

The scientist has access to seismic survey maps and has to interpret them, giving advice on where to drill. The rig manager plans the purchase of hire of rigs to meet the needs of the drilling programme, within the technical and financial constraints. The financial controller has to estimate the costs of hiring and moving rigs, predict the need for overdraft (if he does this and borrows in advance, he gets a better rate), and generally keep the team in touch with financial reality.

You plan your drilling programme, feed it in and get the results, taking into account what your competitors have done without your knowledge, including suitably randomised weather effects.

This has the fascination of 'monopoly' and is much closer to reality. It also teaches team work, and how to spread risk. It is competitive, in a situation where co-operative skill is more important than chance, and therefore educationally sound. The kids love it....

We have, of course, got here the expertise and to spare to do all these things, and more.... Corporate planning models were in continuous use in Aer Lingus in the 60s. Local authority models have been developed in AFF, directed mainly towards traffic flow; CIE also have expertise in this field. School-orientated work is developing thanks to the efforts of voluntary enthusiasts. There is an input-output model of the Irish economy in the ESRI. Census tapes are available fro project-oriented analysis, keeping confidentiality. The trouble is that all this is dispersed; it lacks continuity of effort and common experience; there is no strategic planning of the development of computer models of relevance to Irish needs, no centre of responsibility with power to allocate resources and set priorities.

Do we have to depend on the vision of the research teams of a transnational corporation for our advance computing development strategy, or do we as a national State take steps to act independently? It is a question of political will.

The Dublin Transportation Study was done in the old cheap-energy environment. The models were built, the conclusions drawn and the teams dispersed. The consequences of the decisions based on those models are still with us, in the form of the Urban Motorway Plan, which is being resisted by residents all over Dublin. there are no plans for a re-think, to my knowledge, or for the establishment of an urban transportation model for use in adaptive mode in an ongoing situation. A centre of know-how has been kept going by CIE. Energy costs have quadrupled. Does this not call for an emergency re-look at the Dublin transportation problem, with a project team drawn from CIE, Foras Forbartha and the Universities?

July 27 1976

(Returning to Peterlee)..... As well as a model of the Scottish economy at a detailed level....there is under development a more long-term strategic type of model, at a higher level of abstraction, and with non-linear dynamic interactions. The specific version being developed relates to the economy of the north of England, and uses techniques developed by Professor Jay Forrester of MIT.. These techniques have been used for highly abstract world economic system models (on the initiative of the Club of Rome), giving results which are controversial and useful, in that while they are not quantitatively reliable, they have made people look more closely at the question of long-term resource management. At the level of macro-economic studies of regions, or indeed small national economies, there is every indication that Forrester-type models are useful tools, which planners ignore at their peril.

The North of England system dynamic model team consists of 20 people, drawn from IBM (Peterlee), central and local government. To date, eleven technical reports have been published, covering topics such as population and employment trends, fluctuations etc, with their underlying reasons.

The primary inputs are: (a) past sales (b) public expenditure (c) technology (d) earnings. These inputs work their way through a series of loops with intermediate variables such as business expectations, planned output, labour demand, output per man, employment, investment, price levels, value added, stocks, savings, borrowings etc; eventually there come out estimates of profits, migrations, consumer expenditure etc, which can form inputs to the next cycle.

I have not heard of any macro-economic modelling of this type going on in Ireland; the norm appears to be the more static input-output type of analysis.... At the time of the debate over the EEC referemdum, Raymond Crotty did some dynamic model-building, rather in the Forrester spirit, which predicted some unpalatable effects, most of which have already come true, despite the rudimentary nature of the model. This suggests that if one has any feel for the dynamics of the interactions of even a few key macro-economic variables, it is possible to be qualitatively predictive to a surprising degree.

There is in UCD an embryonic centre of system-dynamic expertise, in that Philip O'Kane has been using Forester techniques in the Department of Civil Engineering, under Professor Jim Dooge(21). The method lends itself admirably to the modelling of water-resource systems, in which field Professor Dooge has world standing.

At a more marketable level of development...is the Interactive Planning System, developed by Dr Barry Aldred. This is aimed specifically at the urban planners ans managers of urban systems....

....IBM has a research project studying the methods of the researchers...led by Dr Sandy Gourlay in association with the Sussex Science Policy Research Unit...

A model consists of of a set of relationships connecting a set of variables.... By a matrix representation....it is usually possible to get non-zero elements to bunch along the diagonal....imposing a natural order on the variables....permitting the model sometimes to be split into sub-models..... the craft of the model-builder is made scientific and generalisable.... potential for stimulating pure mathematicians creatively towards practical but interesting applications of advanced mathematical ideas of structures and relationships....(4).

NOTES

These notes give some clue about the scope of the missing material!

1. Her first name, Ada, has been borrowed as a name for a high-level language.

2. The key invention, the 'flip-flop' or bi-stable circuit, emerged in Cambridge in the 1930s, in the service of nuclear physics.

3. See Chapter 3.6 for some insights into OR.

4. Dr Paddy Doyle's book 'Every Object is a System', published in 1976, provides a 'cognitive basis for system description' with roots in modern abstract algebra.

5. The writer in 1963 successfully used an analogue computer (belonging to Solartron) to solve a set of linear differential equations which can be set upto describe the fermentation of sugar to alcohol, with growth of the organism limited by the available nitrogen. Such a system can predict the dynamic behaviour of a continuous fermentation system subject to transients.

6. But see Chapter 3.1 on 25/9/74, as well as associated note (10). Qeleq tried to make the transition into digital 'dedicated micro' technology, but was 'leapfrogged'.

7. In the event the Irish banks put the initial effort into streamlining the clearing-house operation, so this argument ceased to hold.

8. Perhaps as a result of the above, the Irish banks have in the end gone for centralised mainframe systems, to the detriment of the level of service available at local branch level.

9. This melancholy task, I understand, was carried out by Hammond Lane Fountries, now defunct.

10. This design philosophy, more recently, is the basis of the success of the Sinclair.

11. This speed of development is in some cases visibly counter-productive, as people hold off buying until they see what the next generation of equipment will do!

12. The principal culprit here is the the US military establishment, which is in the lead as regards military technology. If given its own way, it will destroy us all accidentally, with a 'bug' in the early-warning system, coupled with a 'launch-on-warning' strategy.

13. Now in the Central Bank, associated with the development of the national economic model.

14. For example the Euronet system.

15. It also of course makes political sense. Dr Smith's crusade was fuelled by post-Sunningdale euphoria, vith the real possibility on the horizon of a 'Council of Ireland' with resources.

16. See Chapter 3.1 on 24/6/75.

17. This firm did not survive as a Gaeltarra enterprise. The cultural/anthropological problems involved in providing an economic base for the survival of the remaining Irish-language traditional communities have never been effectively understood.

18. Founded in or about 1968 by spin-off of key staff from IBM (Ireland).

19. The writer had occasion to supply a DCF package adapted to Irish tax-laws, designed to user specification, in or about 1971.

20. I have heard horror-stories of similar nature in relation to food standards and the EEC, but regrettably have been unable to embody them in the 'Food' chapter in an appropriate form.

21. For a brief period in 1982 Minister for Foreign Affairs.

[To Irish Times Column Index] [Techne Associates]

Some navigational notes:

A highlighted number brings up a footnote or a reference. A highlighted word hotlinks to another document (chapter, appendix, table of contents, whatever). In general, if you click on the 'Back' button it will bring to to the point of departure in the document from which you came.

Copyright Dr Roy Johnston 1999