Cover Image
close this bookBlending of New and Traditional Technologies - Case Studies (ILO - WEP, 1984, 312 p.)
close this folderPART 2: CASE STUDIES
View the documentChapter 3. Application of microcomputers to Portugal’s agricultural management*
View the documentChapter 4. Off-line uses of microcomputers in selected developing countries*
View the documentChapter 5. The use of personal computers in Italian biogas plants*
View the documentChapter 6. Microelectronics in textile production: A family firm (United Kingdom) and cottage industry with AVL looms (United States)
View the documentChapter 7. Microelectronics in small/medium enterprises in the United Kingdom*
View the documentChapter 8. Integration of old and new technologies in the Italian (Prato) textile industry*
View the documentChapter 9. The use of numerically controlled machines on traditional lathes: The Brazilian capital goods industry*
View the documentChapter 10. Electronic load-controlled mini-hydroelectric projects: Experiences from Colombia, Sri Lanka and Thailand*
View the documentChapter 11. The application of biotechnology to metal extraction: The case of the Andean countries*
View the documentChapter 12. Cloning of Palm Oil Trees in Malaysia*
View the documentChapter 13. Technological Change in Palm Oil in Costa Rica*
View the documentChapter 14. Biotechnology applications to some African fermented foods*
View the documentChapter 15. Use of satellite remote-sensing techniques in West Africa*
View the documentChapter 16. India’s rural educational television broadcasting via satellites*
View the documentChapter 17. New construction materials for developing countries*
View the documentChapter 18. Photovoltaic solar-powered pump irrigation in Pakistan*
View the documentChapter 19. Photovoltaic power supply to a village in Upper Volta*

Chapter 4. Off-line uses of microcomputers in selected developing countries*

* Contributed by the ILO.

UNTIL RECENTLY, INSTITUTIONS and research scientists in developing countries have been handicapped in the collection, processing and analysis of data. They were faced with two extremes: labour-intensive hand-processing on the one hand and capital-intensive mainframe computers on the other. Although it is argued that labour-abundant developing countries should generally use labour-intensive methods, some inherent disadvantages of these methods make them unsuitable for the processing of large quantities of data. First, a great deal of time is required for the processing and analysis of data with the result that by the time the information obtained is published it is already obsolete. Second, there is a high risk of errors each time data are collected or transcribed. Third, the baseline data are generally inadequately used.

The mainframe computer overcame the above disadvantages but had its own shortcomings. Its high purchase and upkeep costs and requirements for special handling (air-conditioned rooms, specified power supply facilities and trained personnel) have posed problems for users in developing countries with scarce financial resources, unreliable and unstable power supplies and limited skilled manpower. In addition, since mainframe computer facilities have to be centralised, access to them from data collection points is not always easy due to limited or non-existent communication links.

As noted in Chapter 1, recent advances in microelectronics have led to a revolution in computer technology. The microcomputer is capable of performing fairly large-scale data collection, processing and analysis functions. It can also operate under more adverse conditions (compared to the mainframe computer). The advantages of the microcomputer over the mainframe computer are summarised in the following passage which refers to its installation in Kenya:1

“The IBM Model 30 mainframe was introduced in the early 1960s. It required special handling - an air-conditioned room about 18 feet square which housed the Central Processing Unit (CPU), the control console, a printer and a desk, for a key-punch operator. The CPU alone (the ‘brain’ of the machine) was five feet high and six feet wide - and had to be water-cooled to prevent overheating. The CPU of a micro-computer is inscribed in a silicon chip smaller than a fingernail.”

The mainframe computer which cost US$280,000 (in 1960 prices) is now replaced by an Apple II plus computer costing only US$5,000. Developments in software (programs of instruction in computer language) for these micro-computers have made it possible for inexperienced users to learn the basic operations involved quite quickly.

A recent survey by the United States Agency for International Development (USAID) on the use of microcomputers on its projects and those of the Department of Agriculture (USDA), identified 76 separate applications in 33 developing countries (five in Asia, 16 in Africa, nine in Latin America and the Caribbean and four in the Near East). The total number of microcomputers involved was 112 of which 43 were Apples, 15 TRS-80s (Radio Shack), 11 Northstars and 43 others. The majority of software uses were off-the-shelf; most general uses being word processing, electronic spreadsheets and data-base management. In some countries (Nepal and Tanzania for example) networks of computer users have been formed including users working for USAID, USDA, FAO and the World Bank.

Microcomputers can be used either on-line or off-line. In on-line uses the microcomputer receives inputs from a system through analogue to digital converters and the output is fed back to that system through digital to analogue converters automatically. For instance, this is the case when a microcomputer is used in a power system to control the voltage level. The voltage at a particular point is “sampled” and converted to a digital signal which is fed to the microcomputer which can then compare this value with a standard one. If there is a deviation from prescribed limits, a signal is sent to a voltage controller to reduce the voltage.

Off-line uses involve the non-system application of microcomputers. They include uses in data collection, processing, storage and analysis as well as mathematical modelling of a range of processes and computer-aided designs.

Some concrete off-line uses of microcomputers in developing countries are:

- storage, processing and analysis of survey data in rural development projects;

- management in planning and agricultural institutions;

- data analysis and mathematical modelling in research laboratories (for example, modelling of traditional technologies like open-fire cooking and wind-mill water-pumping); and

- use in educational institutions for teaching, and in libraries, for information storage, documentation and word processing.2

In this chapter, four case studies are presented with the aim of:

i) highlighting lessons learnt from the applications; and ii) determining the scope and shortcomings of microcomputer use in traditional activities in developing countries. The first study examines the processing of data from a rural survey undertaken as part of a rural development project. The second study stresses the usefulness of forming user groups to ensure maximum exchange of information on microcomputers in developing countries where local expertise is limited. The third study focuses on the use of microcomputers for financial planning in agricultural ministries while the fourth examines use of microcomputers in the analysis of wind data and in the modelling of wind-mills for developing countries by Volunteers in Technical Assistance (VITA).

I. USE OF MICROCOMPUTERS IN FARM MANAGEMENT SURVEYS (NIGERIA)3

The Agricultural Projects Monitoring, Evaluation and Planning Unit (APMEPU) is a specialist unit of the Federal Government of Nigeria. Created in 1975, it monitors and evaluates a number of agricultural development projects (ADPS) throughout Nigeria. Evaluation is done by means of surveys of smallholder farmers in the ADPS. Data processing is thus critical to the Unit’s performance.

Originally, data were collected by trained enumerators, checked at the project headquarters, and then sent to APMEPU where they were punched into cards for transmission to a centralised mainframe computer located at the Ahmadu Bello University (Zaria) 80 kilometres away. Software problems, power failures, voltage surges, inadequate standby facilities and poor maintenance resulted in APEMU’S failure to perform substantive analysis of the data which were eventually processed overseas. This led to the Unit’s decision to purchase a minicomputer with a one megabyte central processor to be operated with two megabyte discs and three tape drives. It was proposed to locate microcomputers at remote project sites to enable data to be put on floppy discs and transported to the central microcomputer. The first microcomputer was installed in August 1981; since then five projects have microcomputer installations involving a total of 17 machines. The use of microcomputers has eliminated the need for punched cards; transport of floppy discs instead of bulky questionnaires is much more practical.

Choice of Microcomputer

The choice of microcomputer was governed by a number of factors, e.g. flexibility and ease of operation, robustness of the power supply unit and availability of a wide range of software. In a recent survey performed by the University of Jos, 50 microcomputers were studied. It was concluded that Apple II offered the greatest flexibility for a wide range of potential uses. In addition, the modular nature of Apple II makes repair and maintenance simpler as spare modules can be kept for replacement in case of malfunctioning. The modular characteristic of the Apple means that the whole machine does not have to be sent for repairs in case of a fault.

The microcomputers are to be used for data collection, word processing and simple data analyses. The hardware chosen is the standard off-the-shelf equipment with no special modifications. The software consists of programmes in BASIC on five one-quarter inch diskettes. BASIC language was chosen in preference to higher-level languages like FORTRAN and PASCAL because of its flexible data-entry screen, the possibility of rapid program development, its speed of debugging and its excellent character-handling with the use of string manipulation functions.

The Agronomic Survey

This survey involves the collection of information on crop areas, production, agronomic practices, and use of purchased inputs on farmers’ plots. Starting from October 1981, the backlog of data for seven projects was entered into microcomputers and was available for analysis by February 1982. The use of microcomputers has also made it possible for data from 1981/82 season to be reported much earlier than in previous years.

Staff of APMEPU evaluation departments assists in the specification of the basic system as well as in the improvements and modifications based on actual field use. The system was selected to enable operation by staff with varying levels of skill; no project required programmer supervision in entering the 1981/82 data.

Data entry is usually done by a single operator for each survey. During entry data are validated by comparing coded values with permitted ranges, and monitoring logic within the form. When discrepancies are identified, the faulty questionnaire is flagged for subsequent editing and correction. After data entry two reports are produced; “field-plot report” and the “household report” which present the major items of the survey in a concise format. This type of presentation gives a bird’s-eye-view of the survey data and enables further errors to be spotted. In addition, simple reports can be produced in time even at this stage.

The Household Survey

The household survey has a complex structure unlike the agronomic survey which has only one form per study unit per year. Each household is visited once a week and two interviews are recorded on one form. Questions are asked about household income and expenditures, consumption, and labour utilisation on farm and non-farm activities. The large volume of data necessitates that results be produced on a period-by-period basis.

Software

The preparation of tables for the agronomic analysis is done by the Visicalc program which enables the user to enter and manipulate data in a tabular form. Tables obtained from Visicalc can be included in reports and transfer of data from Visicalc to Visiplot enables their graphical representation to be obtained. The Wordstar package was used for word processing. This considerably simplified report preparation. Survey packages have also been prepared by APMEPU.

Training

For the Agronomic Analysis Programme, a two-week training course is offered to beginners. These courses, offered by APMEPU, instruct personnel on the use of microcomputers, data entry and report preparation. During the course, data are entered to ensure that sufficient problems are encountered, and instruction on preventive maintenance is given.

In addition, specialised training is given in the use of software packages. Period of training for Visicalc is three days, for Visiplot/Apple plot, one day, and for Wordstar, five days.

Hardware Supply and Maintenance

The equipment was bought from a single agent in Lagos at a very high price. However, the dealer has been unable to meet the demands for computer systems and his supplies have been somewhat erratic. APMEPU was able to install and modify the equipment in the field; it has assumed complete responsibility for maintenance. A dealer’s “kit” including a complete set of replacement boards and spares for each of the chips is kept. Maintenance of peripheral equipment is done by referring to the usually comprehensive reference manuals or by testing individual boards, or by exchange with a similar unit. Repair is done by replacement of faulty parts by spares; equipment which cannot be repaired is sent to the United Kingdom either as a complete item, or as a specific component if the fault is localised.

Lessons from the Study

Two main lessons from this case study are that:

(i) decentralised microcomputer system can considerably speed up data reporting from rural development projects; and

(ii) the required training period of only two weeks demonstrates the effectiveness of the task-oriented method of training.

II. APPLICATION OF MICROCOMPUTERS IN DEVELOPMENT PROJECTS (NEPAL)4

A computer user group - Nepal Association of Microcomputer Advocates for Support and Technical Assistance (NAMASTE) - was formed in January 1981. The main objectives of this association are:

(i) to promote the appropriate utilisation of microcomputer technology for development in Nepal and other countries;

(ii) to serve as a forum for exchange of software, information and ideas;

(iii) to provide a pool of hardware knowledge and spare parts to aid maintenance and repair;

(iv) to foster donor support for the above three purposes; and

(v) to facilitate information exchange with other (non-Apple) systems in Nepal.

Within two years of its formation, the membership of NAMASTE has grown from three to over 35 users. Microcomputer applications are shown in Table 4.1 below. A total of ten development projects made use of microcomputers.

Table 4.1. Programme Applications of Microcomputers in Nepal

Programme

Applications

1. Family Planning and Maternal Child Health Project

1a. Monthly reporting system of 10-20 services provided by 1500 family planning clinics nationwide

1b. Monitoring of target versus actual reports

2. Health Planning Unit

2a. Sales monitoring

2b. Financial planning

3. Rapti Development Project

3a. Monthly financial analysis of 54 offices of 21 line items

4. Community Forest Development Project

4a. Socio-economic survey data analysis

4b. Project projected versus actual on: finances, nurseries constructed. forest planted, and transferred to village control

5. Resource Conservation and Utilisation Project

5a. Procurement tracking system

5b. Payroll and personnel records

Source: D. Ingle, N. Berge and M. Teisan, Acquiring and Using Microcomputers in Agricultural Development: A Manager’s Guide, United States Department of Agriculture, Washington, DC, 1983.

Software

The following four major types of programs were used for project management: Visicalc (the electronic spreadsheet), statistical packages, word processing, and several customs-designed programs. Visicalc was used for financial accounting, resource monitoring and report generation. The data-base management software package which was ordered with the hardware was hardly used.

Hardware

Of the 35 machines in use, 18 were in development projects financed by international agencies. In all but two cases, the Apple II plus computer consisting of 64 bytes of memory, two 5-inch disc drives, various printers and battery-powered back-up systems were used.

One major problem affecting microcomputer purchase in Nepal was the recent establishment of a National Computer Centre consisting of two mainframe computers. This centre which is in charge of all computing in the government agencies also had the responsibility for approving the acquisition of microcomputers by these agencies. The centre is confident of its ability to meet the needs of government agencies, and as a result, approval for microcomputers is difficult to obtain. Many of the machines in use are purchased as “accounting” machines or “typewriters”.

Training

An expert was sent by USAID to train project personnel. The course which was initially planned to last for six weeks in practice lasted for one year during which period only Visicalc was learnt. Of the 22 participants who started the course only six completed it.

Lessons from the Study

The following three lessons can be drawn from the above case study:

(i) the formation of a local group of users of new technology (microcomputers) was helpful in dealing with issues related to its application and maintenance;

(ii) although the Visicalc program is very flexible, it should not be used to replace special-function programs like the Data-Base Management (DBM) system. The latter was hardly used in Nepal despite its availability and the fact that it is more efficient in analysing some data.

(iii) training should be oriented to specific tasks rather than teaching of programming skills.

III. MICROCOMPUTERS AND FINANCIAL MANAGEMENT IN MINISTRIES OF AGRICULTURE AND LIVESTOCK DEVELOPMENT (KENYA)5

A number of problems in financial management were observed in Kenyan Ministries of Agriculture (MOA) and Livestock Development (MLD). In budget preparation, no effort was made by the ministries to evaluate budget items; neither were priorities established. This led to unrealistic budget requests being made to the Treasury. Expenditure monitoring and reporting suffered from problems of lack of focus, untimeliness and inaccuracy resulting in the inability of these ministries to spend their allocations, and accumulation of unclaimed bills for the donor agencies.6 This situation threatened to reduce government budget allocations for MOA.

The above problems necessitated a search for ways and means to improve the performance of the two ministries in formulating realistic budgets and in monitoring expenditure effectively. In 1979, as part of a USAID-funded project to explore the application of microcomputers in food security systems in developing countries an expert was sent to Kenya on a one-month visit to examine whether microcomputer technology could improve information management in MOA and MLD. The purpose of this visit was to explore the use of microcomputers in:

(i) improving the ability to evaluate consistency and implications of crop forecasts;

(ii) increasing the capacity to assess the status of food security within the country;

(iii) providing more systematic and timely data on budget expenditure, personnel, vehicle control and payroll administration;

(iv) developing increased capacity to provide information on the status of project implementation; and

(v) developing a word-processing capability that would reduce the time required to prepare project feasibility studies.

In 1981 it was recognised that the need for computers in ministry-wide budgeting and financing was most immediate. Accordingly, the use of a microcomputer was recommended.

Hardware and Software in Use

The type of microcomputer recommended was an Apple II with 64K, a CP/M card, two disc drives and a matrix printer costing US$5,000. The choice of Apple was governed by the availability of software, its high level of resistance to fluctuating power supplies, the presence of several other similar facilities for maintenance and repair in the country, and the possibility of expansion of its capacity by using expansion boards. The ministries bought one expansion board of 128K capacity. A member of the USAID technical assistance team was assigned to the Management Support Unit (MSU) of the ministries on a full-time basis to assist with the installation of the microcomputer and with the training of Kenyans in its use. The software in use included Visicalc, Wordstar (word processing) and dBase II.

Microcomputer Applications

Budget preparation. In 1981, the microcomputer was used to prepare the ministries’ budget requests for the 1982/83 financial year. The initial request from the Ministry of Agriculture for recurrent expenditure was US$83 million as against a treasury ceiling of US$23 million, while the development budget request was US$71 million as against a ceiling of US$30 million. It was decided to do in-house reviews with the aid of the microcomputer instead of passing such requests to the Treasury for arbitrary cuts.

Using the Visicalc, a five-page summary format (as opposed to 100 stencilled pages produced in the past) based on all program subdivisions of the two budgets, was prepared. It presented 1981/82 budget proposals, expenditure for that year and 1982/83 forward budget figures. This information was used to evaluate budget items in the light of expenditures for the current year; it eventually led to a substantial reduction in the budget requests. The microcomputer was taken to a number of meetings in which the proposals were discussed and decisions were entered into it as soon as they were taken.

Expenditure reporting. A Visicalc summary template was created for each sub-heading of the recurrent and development budgets for MOA and data for field expenditures were entered into the microcomputer as soon as they were received.

The final report is available as soon as the last figure is entered since the machine can reproduce up-to-date details on demand. This significantly diminished many of the shortcomings of the traditional methods of expenditure reporting. Thus, reports were produced much more quickly, their accuracy improved and human errors in arithmetic operations eliminated. In addition, since the figures were entered by the data processor himself who understood their meaning, errors were reduced and less professional time was required in proof-reading and correction of the work. Finally, the reports became more focused and therefore usable by the decision-makers. The Visicalc allowed for some flexibility in the modification of the format in which data were presented. However, it was not adequate for reassembling the data along different analytical dimensions for use by the different units of the MOA. This flexibility can be achieved only by the dBase system which is much more complicated to use.

The timely availability of expenditure reports made their use possible in the current year’s budgeting exercise as well as in addressing management problems of over and under-expenditure.

Training

The background of the personnel trained to enter data in the machine ranged from secondary (high) school-leavers to university degree-holders. A higher secondary school mathematics (advanced level) seemed to be required for programming the software packages for specific applications. A woman trained in programming was now able to train lower level personnel: she was able to make minor modifications to the dBase programme. Plans are underway to train a mathematics graduate with at least a Bachelor of Science degree in producing original software programs.

One can conclude from the foregoing that few barriers exist in the adoption of a new technology which fulfils a genuine need. Moreover, micro-computers can serve as a tool for improving financial management in government institutions which can considerably reduce costs and delays in decision-making. Finally, mastery of dBase requires much more training than Visicalc.

IV. ANALYSIS OF WIND DATA AND MODELLING OF WINDMILLS: VITA EXPERIENCE IN MICROCOMPUTER USE7

The determination of the feasibility of using wind power in any location and the choice of windmill to serve specific needs require a detailed knowledge of wind behaviour at a particular site. Therefore, collection and analysis of wind data constitute an important task for research scientists. The types of information required are:

(i) yearly pattern as defined by monthly wind speeds;
(ii) wind histograms for each month of the year;
(iii) measure of day-to-day wind variations;
(iv) daily fluctuation of winds;
(v) duration of calm periods;
(vi) wind direction; and
(vii) maximum wind speeds.

The above information can be used for assessing the wind resource, designing wind systems and water storage facilities, and determining optimal windmill sizing.

Data Collection

Wind measuring devices are of three types, namely, anemometer totalisers, non-programmable wind recording devices and programmable recorders (data acquisition systems). Programmable recorders have been made possible by recent advances in microelectronics technology. One type consists of wind-specific recorder which feeds data into a microcomputer for processing. The data can be stored in digital form in cassettes and are fed into the computer for processing with user-written computer programmes. Once in the computer, histograms, averages, variations and graphic representations can be obtained. Another method uses the microcomputer memory chips to store the data. The chips are known as EPROM (Erasable Programmable Read Only Memory). In general, cassette tapes can store more data than EPROM but cassettes are more vulnerable to hot climates. VITA is now using programmable recording systems in Djibouti.

The climatronics weather data collection system in Djibouti uses batteries charged by solar panels. A programmable microprocessor is used for data analysis. It uses magnetic data storage cassettes and has a duplicate paper chart recorder. Plans are underway to supply to Djibouti an Apple computer for data analysis.

Data Analysis and Modelling

VITA has developed a number of programs for analysis of wind data. The first such program is the Weibull wind distribution program which can produce a “Weibull” curve from an observed wind histogram. This curve is a statistical model which can be used to predict wind patterns in more detail than with histograms. The program also compares theoretical and observed values of mean wind speeds, mean power, energy pattern factor and histograms.

Another program calculates water outputs to be expected from an American multi-blade windmill. It creates a probable output curve which can be used to predict water output at any given wind speeds using the above machine. Inputs required for this program are: size and type of machine, the pumping head, pump characteristics and the wind regime. The program gives water flow rates for any instantaneous value of wind speeds as well as total water output per day or month.

The third program computes the water storage capacity required. This information is very important, given the high cost of storage and distribution systems. The inputs required for this program are: a windmill output curve, 30 daily windpseed values, tank size and consumption level. The program highlights possible shortages in addition to determining the storage capacity.

The fourth program uses a life-cycle costing method to determine the annualised unit cost of each cubic metre of water pumped over the life of the windmill. Inputs required are: windmill life, total initial capital investment, operating and maintenance costs, daily water output, salvage rate of the equipment, the discount rate and the length of the project.

While programs for water pumping predictions have to be tested extensively, the initial results seem encouraging. Projections have been compared to actual measured values in Honduras. Agreement was within 6 per cent over a wide range of daily wind speeds, as is shown in Figure 4.1. VITA uses an Apple II computer and the above programs to analyse the potential of windmills in any given location.

VITA has also used computer simulations to do comparative analysis of new and conventional (both wind and diesel-powered) methods of water-pumping.

V. CONCLUSIONS

While all of the applications described in this chapter are possible with mainframe computers, the cost of such equipment, the requirements for special accommodation with controlled temperature and humidity, sensitivity to power supply irregularities and the need to have highly skilled personnel for maintenance and repair, made them unsuitable for use in developing countries. The recent development of microelectronics has enabled the functions of the mainframe computer to be performed by inexpensive, easy-to-use microcomputers which are less sensitive to climatic and power supply conditions.

For a coordinated approach to the use of computers in developing countries it is recommended that needs assessment be done to determine the issues and problems involved in rapid and accurate collection, organisation and analysis of data. The needs should be assessed together with the available expertise. Although many of the operations described in this chapter required clerks with only secondary education, more complicated operations might demand a higher level of expertise.

Once the needs are assessed, an appropriate technology choice should be made. A wide range of hardware and software exists, but some expertise is needed to assess the systems that would be suitable for concrete needs.

In order to minimise new technological dependence of developing countries, adequate emphasis should be given to the development of local expertise in the use, maintenance, trouble-shooting and repair of microcomputers.

While it appears that microcomputers can be useful in various applications in developing countries, there are also a number of undesirable effects which must be considered before deciding on these applications. One of these effects are the employment implications of the use of microcomputers. Although no detailed study on employment consequences exists, a widespread use of microcomputers is likely to create some degree of unemployment. In the developed countries it is often argued that any employment loss would be offset by growing employment in the microelectronics industry. This argument cannot be valid for developing countries which import microcomputers. Thus, it is desirable to consider employment implications as one of the factors in the choice of microcomputer technology. Over-dependence on microcomputers results in a tendency to use them for solutions to every problem encountered, a fact which leads to inappropriate uses of the equipment. For example, in word processing, the decision to use computers must be carefully considered if clerks are not to be displaced.


Figure 4.1: Microcomputer Printout of Predicted and Actual Values of Water Pumped and Different Mean Wind Speeds in Honduras (The white dots are the actual performance observed in Gravel Bay; the predicted curve is obtained from the computer model.)

Another effect of microcomputer use which must be guarded against is that on health. Steps need to be taken to safeguard against occupational stress like the effects of the use of visual display units on the eyes.

There is no doubt that microcomputers can serve as a useful tool in the development process. However, their widespread application requires a coordinated and methodological approach if their negative effects are to be avoided.

NOTES AND REFERENCES

1. “To each his own computer”, in Newsweek, Atlantic Edition, Winterthur, Switzerland, February 22, 1982.

2. Many teaching institutions in developing countries who could not afford the cost of acquisition and/or maintenance of mainframe computers can possess microcomputers and successfully introduce students to computer technology.

3. J. Benett and D. Poate: “The use of microcomputers in farm management surveys”, in M.T. Weber et al., Microcomputers and programmable calculators for agricultural research in developing countries, MSU International Development Papers: Working Paper No. 5, Michigan State University, East Lansing, 1983.

4. M.D. Ingle, N. Berge and M. Teisan: Acquiring and using microcomputers in agricultural development: A manager’s guide, United States Department of Agriculture, Washington DC, April, 1983.

5. See T.C. Pickney, J.M. Cohen, and D.K. Leonard: Microcomputers and financial management in development ministries: Experience from Kenya, Development Discussion Paper No. 137, Harvard Institute for International Development, Cambridge, Mass. 1982, and D. Ingle et al., ibid.

6. The government typically pays bills for donor-assisted projects and claims reimbursement from the donors. At the end of 1981 it was estimated that a sum of US$155 million was outstanding.

7. Volunteers in Technical Assistance (VITA), study drawn from VITA News, Special Issue on Wind power for developing nations, Arlington, Virginia, United States, July, 1983.