1968

Most Appropriate Computer Systems for UIA

Detailed Examination

- / -


Section of Report of a Preliminary Investigation of the Possibility of Using Computer Data Processing Methods (1968): a summary of the various parts of this report, and details of its contents (with links to the various parts), are provided separately


The previous section examined each of the possible systems in order to make a preliminary selection of those which warranted further examination.  In this section detailed solutions to the problem of publication production, semi-commercial processing, and research processing are discussed.  These solutions are based on the contacts made with computer companies regarding each problem.  The conclusions indicate the further steps that need to be taken before definite decisions can be taken on whether the systems should be implemented.

Production of Publications

The discussion in the previous section showed that the best approach to the use of computer typesetting facilities was initially to concentrate on the separate production of the Yearbook, Calendar and Bibliography. This was investigated by making contact with the only known firm in Europe to be engaged in computer typesetting on a service bureau basis. The firm is Rocappi, Ltd based in London and using International Computers and Tabulators Ltd (I.C.T.) equipment. On the basis of its experience with this firm, l.C.T. had built up suitable software for use on all its machines in the 1900 series. No other computer manufacturer appeared to have produced software packages for this application, although International Business Machines Inc. (I.B.M.) has experience in producing programs for particular newspaper applications. l.C.T. does not currently have any service bureau machines in Belgium.

Description of Program (I) The first contacts were established to determine whether the Calendar and Bibliography computer typesetting economies would be comparable with the current cost of producing these publications using conventional methods.  No attempt was therefore made to refine the problem by combining different operations or by doing any indexing automatically. Production of the Yearbook was not discussed at this stage. The definition of the program as accepted by Rocappi specified a January 1958 starting date.

Comparative Costs: The offer made by Rocappi on the basis of the above program description compared favorably with the four offers received from conventional Belgian printers.  It was decided however that since the use of typesetting had not been fully explored and that moving some operations to England required further thought, the Rocappi offer would not be accepted.  The latter had been made before the December devaluation of sterling which would have made the offer even more interesting. An additional factor was that one of the Belgian printers had introduced an exceptionally low offer.

Description of Program (II): In the light of the above experience, further contacts were made with Rocappi, I.C.T. and I.B.M. to explore the possibility of combining the production of the three publications together with an increased use of the research facility offered with the use of computer typesetting. A detailed list of the facilities discussed is given in Exhibit 28. This summarizes some of the facilities which were mentioned earlier.

Results of Discussions

a)   Rocappi.  This company is currently using an I.C.T. 15OO computer and hopes to convert within the next year to an I.C.T. 1900 with disk units.  The company did not have a detailed description of the software it uses and it appears that each case is dealt with individually.

The Rocappi representatives were very enthusiastic about the combination of the three publications into one integrated system with research facilities. All the facilities that could be required by the UIA are technically feasible.  There are, however, three reservations.

  • As was explained earlier, the more programming sophistication necessary in a given application, the higher is the cost. The UIA program description did not enable them to establish criteria for selecting which facilities should be favoured and which should be eliminated.
  • The second difficulty was the relatively low monthly volumes of data processing which would tend to make the production of the monthly calendar and bibliography uneconomic, as well as the quarterly Yearbook supplement.  (This is probably one reason for which the major competing calendars are produced on a two or three monthly basis.)
  • The third difficulty was the lack of information on which questions the UIA wanted answered. The conclusion was that the UIA should prepare a more detailed program description giving indications of the relative importance of each facility so that Rocappi could optimize between the costs and the number of processing options to be included. Once this had been done, detailed costs could be given for the implementation of the system. At this stage, they would send over a systems analyst to determine how the system could best be implemented.

b)   I.C.T.  As mentioned above, this company made use of its experience with Rocappi to produce a complete set of software for the I.C.T. 1900 series. A detailed manual was supplied indicating all the processing options available. Features which would be important to a UIA system are summarized in Exhibit 29.  The company is based in England and is only just beginning to enter the European continental market. The representative did not think that they would have a service bureau in Belgium in the next three years, although it might be possible to hire time on one of their machines which had been sold to a customer in Antwerp. They could not offer any programming services.

The conclusion was that no further action could be taken unless the UIA was prepared to make use of their service bureaux in England.

c)   I.B.M.  This company has extensive service bureaux facilities in Brussels and has already got computer typesetting systems in operation here in several newspapers. The company was very enthusiastic about the possibilities of the combined system. The same reservations were raised as in Rocappi's case. The conclusion was, as with Rocappi, that the UIA would prepare a more detailed program specification indicating the relative importance of different processing options.

d)   Siemens.  This company has extensive service bureaux facilities in Brussels and also has computer typesetting system in operation in Europe.  The company has not yet been contacted.  This should be done if it is decided to continue the discussions with I.B.M. and Rocappi.

Conclusion on Publication Production

The production of the Yearbook of International Organizations using computer typesetting methods is the most fruitful possibility. This would enable specialized directories to be produced at no editorial cost and would make it difficult for specialized organizations to split the UIA market for Yearbook information.  It would also leave the UIA in a position to decide whether the Yearbook should be made into an annual publication in order to compete effectively with the Europa Yearbook.

Preliminary conclusions from discussions with the companies contacted indicate that monthly and even quarterly processing of data would prove uneconomic. It is not yet clear whether the convenience of the typeBetting facility and the savings in editorial cost will be important enough to warrant obtaining funds for the switch from conventional methods of Yearbook processing.

No detailed conclusions are possible at this stage.  It would be more convenient to operate with a company based in Brussels, but there may be an important saving in making use of a service bureau in England where costs are up to 50% lower in some cases. Printing costs are also lower.  It might be possible to take advantage of a telex link across the Channel to avoid the postage delays.  This would have the advantage of preparing the UIA for forthcoming national and international data networks linking libraries and other centres.  If this approach is adopted, care must be taken that the telex code used is an accepted standard (see van der Wolk, L.J. Ml8;  United States of America Standard Institute,M17). A disadvantage of this approach is that no international standard has yet been developed and when it is, it will disrupt systems already in operation.

It is clear that the UIA must decide on the relative importance it wishes to place on different publications, survey subjects, and questions. It must also decide on the volume of the requests it plans to cater for. No information is available on the number of requests that are likely to be received once a system is running smoothly.

No study has been made of the possibility of using computers without tne typesetting feature, to prepare bibliography and calendar texts, together with indexes for all three publications.  The computer printout is then copied and reduced photographically for offset reproduction (see Ap. Exhibit 38c). This has not been investigated because it is not the UIA policy to use cheap production techniques for its major publications. Some of UIA's major competitors (even government supported libraries) use these methods. It is therefore possible that they might be more suitable solutions to the publication production problem. The more funds expended on the quality of presentation of a publication, the less funds available to improve and maintain the quality of the contents, unless the publication must be sold on basis of presentation rather than content enough to warrant obtaining funds for the switch from conventional methods of Yearbook processing.

No detailed conclusions are possible at this stage.  It would be more convenient to operate with a company based in Brussels, but there may be an important saving in making use of a service bureau in England where costs are up to 50% lower in some cases. Printing costs are also lower. It might be possible to take advantage of a telex link across the Channel to avoid the postage delays.  This would have the advantage of preparing the UIA for forthcoming national and international data networks linking libraries and other centres. If this approach is adopted, care must be taken that the telex code used is an accepted standard (see van der Wolk, L.J. Ml8; United States of America Standard Institute, M17). A disadvantage of this approach is that no international standard has yet been developed and when it is, it will disrupt systems already in operation.

It is clear that the UIA must decide on the relative importance it wishes to place on different publications, survey subjects, and questions.  It must also decide on the volume of the requests it plans to cater for. No information is available on the number of requests that are likely to be received once a system is running smoothly.

No study has been made of the possibility of using computers without the typesetting feature, to prepare bibliography and calendar texts, together with indexes for all three publications.  The computer printout is then copied and reduced photographically for offset reproduction (see App. Exhibit 38c). This has not been investigated because it is not the UIA policy to use cheap production techniques for its major publications. Some of UIA's major competitors (even government supported libraries) use these methods. It is therefore possible that they might be more suitable solutions to the publication production problem. The more funds expended on the quality of presentation of a publication, the less funds available to improve and maintain the quality of the contents, unless the publication must be sold on basis of presentation rather than content.

Semi-Commercial Problems of the UIA

The discussion in the previous sections showed that there were apparently two feasible approaches to the solution of the UIA routine semi-commercial problems.  These were either to attempt to undertake a computer solution by itself, or alternatively, to attempt to pool together with several other organizations in order to reduce costs. In either case it was possible to visualize using either classical equipment (sorters, tabulators, etc) or full scale computers. The problem in the former case was the lack of flexibility in designing programs, so that it would be unlikely as a solution in conunction with other organizations.

In order to cover both possibilities contacts were established with computer companies, commercial computer service bureaux and also with service bureaux using classical equipment.  The UIA problem was defined in as much detail as possible and initial offers were requested for preparing programs and for the monthly processing cost.

To prepare the way for discussions with other organizations which might be willing to collaborate with the UIA, the solution for which offers were invited was presented in a generalized manner. The idea was that the UIA problems would be generalized so that a program could be designed which would be sufficiently flexible to handle the data processing of other organizations.

In parallel with these contacts, offers were invited from service bureaux specializing in classical equipment.  In this case, because of the lack of flexibility, the offer was made specific to the UIA problems.

Description of Generalized Program: An international non-governmental organization was assumed to have two basic types of semi-commercial problem, namely addressing from a mailing list and procedures associated with handling financial transactions. In a commercial organization, these problems would be treated by many separate programs and might not even be within an integrated system.

It was assumed that the set-up costs and processing costs which would be incurred by treating each problem separately would make such an approach impractical. This is an expensive technique when dealing with low volumes of information and is only justified when a large proportion of time is spent in processing relative to the amount of time spent in setting up the machine. In an effort to avoid the high initial cost of preparing a number of programs, and the regular supplementary cost of setting up the machine for each job, it was decided to investigate a radical and somewhat unconventional solution.

The solution was to perform all addressing (publicity, magazine wrappers, etc.) and all financial transactions (invoicing, reminders, monthly accounts, etc.) in one operation involving only one set-up period.  The computer would therefore be set-up only once to deal with all organizations in the group. This would be done on a monthly basis in order to allow a sufficient volume of information to accumulate between runs. A further object of the system was to reduce the amount of time spent by NGO personnel on routine administrative operations.  In small organizations the executive staff often have to spend considerable time on administrative work rather than on directly developing the activities of their organization. By streamlining the processing of administrative data, it was hoped to assist in further increasing the efficiency of small organizations as well as contributing indirectly towards increasing their effectiveness.

In detail, the envisaged solution performs the following functions listed in Exhibit 30.

A general program specification is given in Exhibit 31. This Exhibit also includes data on the volume of UIA data processing. It was not possible to go into further detail because each computer service bureau employs different equipment and has to adapt such a specification to the facilities at its disposal.

Details of System Operation: The system envisaged would operate as follows.  The customers, contacts, or members of each organization using the system would be grouped on a main address tape within each organization in sequence.  (Where tape is mentioned, use of a disk unit is also possible.) This tape would contain for each customer or member, the name, address, and a code of about 40 columns giving information on the type of organization, location of headquarters and other details.  The suggested coding for the UIA file is given in Exhibit 33.  Different codes could be employed by other organizations.

During the course of each month, orders would be received by the collaborating organizations. These orders would be converted into codes on a standard sheet. The sheet could contain codes for:  number of customer/ member, publication ordered, number ordered, currency in which invoice is to be prepared, discount to be given, standard additional remarks to be added to the invoice (see Exhibit 34). Any non-standard comments to be included on the invoice in lieu of a separate letter, could also be noted.  In a similar way payments from the customer/members would be received, either as membership fees or for publications purchased. These payments could be converted into codes on a second standard sheet. Since the two sheets do not contain much information, and the limit for input by computer on punched cards is 80 columns, it should be possible to combine these two sheets to cover both orders and payments. The codes for payments could include:  number of customer/member, amount paid, currency (see Exhibit 34).  The customer/member number would be obtained from a single card file, preferably of the rotating wheel type. During the course of the month, information would also be received on changes of address and new organizations. This would be indicated on a third sheet (see Exhibit 38).

At the end of the month prior to the run, the various sheets would be grouped together.  Decisions would be taken on what groups of organisations should require address labels printed for publicity purposes or (questionnaires (see section on General Mailing below). Separate sheets would be prepared indicating the appropriate call codes of these groups, e.g. all university libraries, all European based international organizations, etc. (see Exhibit 35).

The various sheets ('punch documents') from all the organizations using the system, would then be sent to a service bureau which would prepare cards from them. These cards would then be sorted into customer/ member order within the participating organization code. Any standard cards required at the head of the card pack for a particular organization would be added. In the UIA's case this could include a set of about 100 cards giving the names of the UIA publications and their current prices in various currencies (see Exhibit 36), and a currency conversion rate card (see Exhibit 37). These cards save using permanent storage space in computer memory and also allow changes to be made from time to time without alteration to the program.

Note that in the UIA's case it would probably be an advantage to hire a card punch.  This would mean that data could be punched directly onto the card without requiring preparation of an intermediate punch document. If the Flexowriter system was used (as for the suggested publication production system), a typed duplicate of the contents of the card is prepared as the card is punched.  This avoids the need to use a verifier.

Monthly Computer Run: The cards are now ready for the monthly run. The previous month's updated address tape for the participating organizations is set-up on the computer tape decks. The computer first checks the code of the participating organization whose customer/member data is being processed It can then bring into memory any special program instructions for that data. The standard data cards at the head of its pack are read into memory (see Exhibits 35, 36 and 37). The computer then checks through each customer/member.

In the UIA's case the procedure would operate as follows. The computer has a record of which groups of addresses are required that month (from cards as in Exhibit 35).  It therefore examines the codes of each organization on the main tape to see whether they correspond to these groups. If they do, it then checks to see whether any modifications cards (Exhibit 38) have been included for that organization in the card pack. If there are, it makes the modification and stores the address on another tape ready for printing at a later stage. At the same time it copies the old address tape onto a new tape, incorporating all the necessary modifications.  Depending on the equipment available in the computer system, it is possible to split up the groups of addresses for general mailings so that the addresses when printed out will be conveniently in packages for the dispatch departments.

Order Processing: Following the modifications/additions card for each customer/member in the months card input may be an orders or payments card (Exhibit 34). Alternatively, this card may occur separately, but in any case all the cards are identified by a customer/member number which is matched against the same number on the address tape.

Each participating organization may wish to send to its customer/members, documents which may include invoices, receipts, reminders, statements of account, etc.  If the computer had to be stopped every time the processing for the next participating organization was reached in order to change the pre-printed stationery giving letterhead information, the system would be impracticable. If it was also required that separate documents for each organization with the indication 'invoice', etc. should be employed, the system would be totally impractible. The solution to this problem was to employ a generalized pre-printed document which could be used for all the accounting purposes and also for gentle reminders to members to renew their subscriptions. For a given participating organization, the computer would in each case print out its name and address, together with the name and address of the organization to which the document was to be sent. According to the use to which the document was to be put (invoice, receipt, statement of account), an appropriate indication would be printed out. An example of such a document is given in Exhibit 32.  It can be adopted to use with window envelopes to avoid the present system of retyping the address. In the UIA's case the procedure would be as follows.

Any orders for a given customer/member would first be processed. The computer would print-out 'invoice' as indicated in Exhibit 32 and then go on to supply the various lines of the invoice in the normal manner. Because the UIA has some publications with titles in two languages, the appropriate text for a given publication (supplied via cards as in Exhibit 36) must be selected according to the language of the client as indicated in the code on the address tape. Discounts (indicated via cards as in Exhibit 34) and other texts would be added. Whilst processing the order(s) for a given customer/member details of the invoice total, date and number would be stored on magnetic tape. This tape would serve as the basis for preparing the debtors ledger required by Belgian law for tax purposes.

Since the UIA has some difficulty in collecting payment for publications sold, it is an advantage to combine a statement of account with the invoice.  'Statement of account' would therefore be printed on the same document (see Exhibit 32) if there had been any orders, or would be printed at the head of the document if there had been no orders. Unpaid invoices would then be listed, with perhaps an indication of the number of times reminders had been sent for a given invoice. Payments would be indicated and adjusted against the invoice tape.  Since payments may not necessarily correspond to a given invoice, no attempt would be made to link the two. Invoices would be considered paid once a credit had been received for an amount greater than the oldest unpaid. The layout could be approximately as shown in Exhibit 32.

Receipts: If a receipt was expressly requested, as is sometimes the case, the computer could then go on to print out 'receipt' and the appropriate details.

Copies: There is no difficulty in producing copies of documents on a computer. Stationery is available which includes a sheet of carbon paper between two sheets of continuous preprinted stationery, or more copies (up to about 5) can be obtained by using more sheets of paper in the normal way.  Such stationery, particular for larger numbers of copies, is relatively expensive. There are two methods of avoiding excessive expenditure on copies.  The simplest is to ensure that instead of producing a copy through a carbon, the computer prints out two versions of the same text, next to one another, on to what is essentially the same sheet of continuous stationery perforated down the middle.  The second version could be pre-overprinted with the word 'copy'. This system is ideal for an original and one copy.  Theoretically, the computer could go on to print out on the next sheet a further two copies and so on. This practice is not recommended because of the waste of computer time on the relatively slow printing operation. It also uses up computer memory. A compromise is possible, where memory space is available, by supplying sufficient sheets of carbon to give duplicates in the normal manner and printing two invoices at the same time ('two-up'), as explained above. For those exceptions requiring extra invoices, the computer goes on to print a further set.

It is clearly in the interests of the organization to reduce the number of copies to the absolute minimum.  In the UIA's case invoice copies are currently required as follows (see Exhibit 10). Two copies to the customer, one tax copy in invoice number order, one customer record copy in geographical location order, and an occasional supplementary copy for Yearbook sales. This could be reduced to the two customer copies only which would require only single sheet stationery printed two-up, since all the other information would be stored on magnetic tape. It is preferable to have single carbon copy stationery giving two spare copies to be filed to cope with exceptions, correspondence and customer requests for invoice copies.

Publication Dispatch Labels: In the current system each order generally involves retyping the customer name and address on the number of labels required to dispatch the packages ordered. This is time consuming and leads to errors. It would be much more satisfactory if the computer could arrange for package labels to be printed which would also contain instructions to the packager on the contents required.

This is quite feasible. As each line of the invoice is prepared referring to different publications ordered, the computer prepares a label. The number of labels will clearly bear a close relation to the quantity of a given publication.  It will not be the same since some publications go together into one package.  Clearly single items in succeeding lines may mean that a label is prepared which is unneccessary for a package, since the publications can be combined, but in this case the label will serve as an instruction to the packager and can then be destroyed.

The labels cannot be printed out whilst the invoice is being prepared since the pre-printed stationery is on the printer. Nor can they be printed out at the end of the processing for a given participating organization, since the system is designed to go straight on to the data for the next organization. The suggested procedure is that the label data should be stored on magnetic tape until all the data for each participating organization has been dealt with. Each would then have such labels on the storage tape, but they would be in participating organization order.

General Mailings: When all financial procedures were finished, the computer would request a change of stationery.  Special label stationery would then be . mounted. Normally this may have four labels in a row ('four-up') which are printed simultaneously to save computer time. These labels may either be identical or different. The computer would then print out all the labels arising from the invoice detail lines.  It could be instructed to print out the name of the participating organization at the head of its set of labels to facilitate dispatch of the labels to each packaging department.

In the UIA's case there is a further complication. Much of the UIA publication sales go via booksellers. In such cases the invoice must be sent to the bookseller, although the latter may give instructions to send the publications to a given address.  The latter address has to appear on the invoice with some phrase like 'Order dispatched to:'. The address has to be used on the package labels.  This address is valuable as representing a future customer, but because of the arrangement with the bookseller no definite policy has been established as to whether publicity material should be regularly sent. On the assumption, that the address should be retained, and used, at a later date, and since it must in any case be supplied to the computer to prepare labels, the following procedure is suggested. The name and address would be punched onto cards (see Exhibit 38) which would be positioned behind the order card.  The address would only be stored for the immediate production of labels. The punched cards would be kept for later use in publicity campaigns.

Monthly Accounts and Statistics: Using manual methods it is not always practicable to collect together all information on monthly activity which would be useful as a guide for future decisions.  In many cases the only figures available are those prepared at the end of the year. These do not always present a true picture of the organization's activities since the end of the financial year may or may not represent a peak or low point in income or expenditure.  Figures throughout the year can be compared with equivalent months in previous years to evaluate progress. One. of the great advantage of computers is the facility with which statistical and accounting information can be prepared.

A detailed list of suggested information on monthly activities is given in Exhibit 30.  This includes monthly totals of invoices (number, amount), payments, unpaid invoices and their delay, number of each publication group sold (amount paid, discount, UIA gross profit, tax). It would be possible to arrange each month for counts to be made of any particular group of customer/members on the basis of their code grouping (using cards as in Exhibit 35).

Exceptions: The following exceptions arise during the course of monthly operations: - bad debts must be eliminated from the file and so indicated on the code of the organization (via cards as in Exhibit 34). Bad risk organizations could have an invoice sent with a special note to indicate that the publication would be sent on receipt of payment.

  • some invoices require that the amount should be added in writing as in cheques (instructions via cards as in Exhibit 34)
  • recalling information on a particular organization, to check the status, subscription, etc. (instructions via cards as in Exhibit 38).
  • supplying copies of unpaid invoices on request (manually)
  • invoices for special services (e.g. research work), corporate membership invoices, etc. (processed manually)
  • multiple subscription copies required by some organizations. This is covered by organization code (see Exhibit 34)
  • addressing letters for special mailing. This could either be done by sticking labels onto the letter, or by doing a special computer run. This would be more expensive.

Description of UIA Specific Program: In order to cover the possibility that the UIA semi-commercial problems might better be resolved by use of classical equipment of a less flexible nature, preliminary contacts were established with service bureaux specializing in this field.

The UIA problem was presented using the generalized program specification (Exhibit 31) with the comment that the flexibility required by the accommodation of other organizations in the system be ignored. It was stressed that the system should be made specific to the UIA. problem.

Manual Costs: The manual costs can be divided into two groups, namely addressing costs in regular mailings mainly associated with the use of the addressograph system, and the accounting costs. The procedures by which addresses were run and the addressograph system maintained, were itemized on the basis of the preliminary study in Exhibit 10. An estimate was obtained for the time taken to perform each operation. It was not practicable to time operations for psychological reasons within the organization. The estimate obtained was that given by the person responsible for a particular task together with comments byother persons who were associated with the task or effected by it. The estimates are considered to be a fair approximation (since a range is given) but are likely to be slightly lower than the actual time taken when allowance is made for setting up to perform a particular task or resting during the course of performing the task.

The same procedure was adopted for accounting tasks. The total cost is estimated at Bfr 7,300 to Bfr 18,700. The manual costs for performing the operations are listed in Exhibit 40.  These results even though expressed as a range are in good agreement with an estimate of the costs obtained by apportioning the total time spent during a month on the job of addressing and the job of invoicing. This estimate was done entirely separately and gave the range as being Bfr 10,000 to Bfr 18,000.

Costs of Service Bureau Computer:

When the generalized program specification was discussed with each service bureau, it was emphasized that a preliminary estimate was required of the cost of implementing a system that would deal with the UIA problems.  The system could be adapted to other organizations but their effect on the program was not to be taken into account. The costs given are therefore preliminary estimates for the treatment of UIA data only.

Estimates were obtained from three computer manufacturers which run service bureaux, one independent non-profit service bureau, and from one programming service bureau (which only gave estimates for the programming cost).

1. Computer Manufacturer Service Bureaux

a) Honeywell: Punch and verify card file (14,000) Bfr 100,000 Preparation and testing of program Bfr    60,000 (includes creation of tape file, monthly update provision, label feature, invoicing statement of account and reminder feature, creation of debtors ledger tape) Monthly processing office hours   either  Bfr  5,000 other periods  or  Bfr  3,500 These prices do not include the design of the tape file, documents, etc. (which would be done by the UIA).

b) I.B.M.: Punch and verify card file (14,000) Bfr 100,000 Preparation and testing program Bfr 129,282 (includes creation of file and the recpiired features in program) Monthly processing Bfr 11.581 (includes labels, sales ledger, sort of payments, payments ledger, customer ledger). I.B.M. submitted a detailed proposal in which each cost was split into fixed and variable portions. This permits an estimate of the cost if different size files are prepared, particularly if other organizations participate in the system. The breakdown is given in Exhibit 41.

c) N.C.R. (National Cash Register) Punch and verify card file (14,000) Bfr 100,000 Preparation and testing program Bfr 39,000 Monthly processing Bfr 11,864

2. Interburo: This organization was specially created to perform the data processing for a number of mutual and social security companies. The Board of Directors is composed of representatives of organizations using its facilities for the right to which they pay a nominal membership fee of Bfr 2,000 per year.  Data processing was stated as being charged at cost plus 10%.

Despite lengthy discussions with the Director of this organization, he was not prepared to commit himself to specific costs. He was particularly interested in the possibility of the UIA grouping other organizations.

3. Programma: This commercial organization specializes in systems analysis and the production of computer programs to be run on standard service bureau machines. They do not possess such a machine themselves but test their programs during time hired on customers' machines. Punch and verify card file (14,000) Bfr 100,000 Preparation and testing of program Bfr 50,000 Monthly processing Bfr 5,000

Costs on Classical Equipment or Card Computers

These costs only apply to UIA data and do not include the possibility that other organizations might participate in the system. Estimates were obtained from two service bureaux.

1.  C.A.M. This organization was originally a purely commercial organization but converted to a cooperative in 1956.

  • Punch and verify card file (14,000) Bfr 100,000
  • Preparation and testing of program  Bfr 40,000
  • Monthly processing Bfr  9,000

2. Fiduciare Comptable Fiscale et Juridique, S.A., Brussels: Punch and verify card file (14,000) Bfr 56,000 (this figure does not include a separate card for a code on each customer/member) Preparation and testing of program Bfr 50,000 Monthly processing Bfr 10,200 (6,200 plus 4,000 for statistical and accounting data)

Comparison of Costs between Manual and Computer Systems

The first point to note in the comparison of the costs between the service bureau using classical equipment and computers, is that both the programming costs and the monthly processing costs are of "the same order as those quoted by Honeywell and N.C.R.  Since the classical equipment and card computer systems are severely handicapped in their treatment of relatively complex processing required in the UIA case, these will not be considered in the comparison between the manual and computer systems.  In particular, it is difficult on these systems to process the invoices effectively.

The range of figures for the cost of monthly processing is therefore: manual: Bfr 7,000 19,000 computer: Bfr 5,000 12,000 for an investment of Bfr 140,000 230,000 possible saving: Bfr 2,000 14,000 by use of the computer system possible cost increase: Bfr    0 5,000 by use of computer system Since these figures are only preliminary estimates, it is better to ignore the possible direct financial advantages and consider whether the UIA is prepared to invest approximately Bfr 200,000 to increase the general efficiency of its treatment of commercial data processing. Although it is tempting to take into account the possible savings, the majority of organizations which install or use computer systems report that direct financial saving are in fact usually negligeable on a given operation. They do however report that the major benefits arise indirectly from improved control and effectiveness of operations.

Implementation of Semi-Commercial Processing via a Service Bureau

Implementation of the system would not be very difficult. All usefully address data in the files would have to be resorted and the information recorded on specially designed punch documents (based on Exhibit 38). Each address would have to be coded on an attached punch document based on Exhibit 33.

The existing system and the new system could be run in parallel for one or two months using temporary staff on the existing system.

Advantages of Computer Semi-Commercial Processing for the UIA

  • avoids errors which arise in manipulation of addressograph plates
  • increases the number of groups of addresses which can be used making it easier to contact a particular type of address either by field of interest or by purchasing history
  • facilitates the organization of advertising campaigns
  • reduces invoicing errors
  • reduces routine work
  • reduces dependence on trained personnel in an organization where personnel turnover has always been high
  • avoids the problem of finding and paying reliable persons willing to perform the unpleasant and rather complex task of manipulating addressograph plates and indicators
  • forces the UIA to adopt well-defined procedures.

Disadvantages of Computer Semi-Commercial Processing for the UIA

  • the delay required to allow orders and other processing to accumulate to an economic volume
  • the difficulty and cost of designing and operating satisfactorily an invoicing system with many exceptions
  • the rigid definition of the UIA operations at a time when the production program is not stabilized
  • many of the problems of the current system are due to poor control of operations. A computer system would not improve the situation and would probably make it worse unless an attempt was made to change current methods. If this was done the current system would probably prove satisfactory.

Conclusion on Semi-Commercial Processing

The proposed system would satisfactorily perform all the processing required, particularly if the number of exceptions could be reduced. It is difficult to determine whether the system would give better results than an improved manual system.  Since it is necessary to improve the control and methods of the addressing, invoicing and accounting, it would be tetter to do this before making any decision on the adoption of computer methods. Once an improved manual system is in operation, the advantages of using a service bureau may then be less significant.

Research Processing

A study has recently teen published on the construction of a general model for the automatic processing of documentary data. This study concluded that the non-mechanical stage of information storage, namely indexing, represented between a third and a half of storage costs. Also that mechanization of indexing will not substantially effect the total cost of automatic documentation.

On the basis of this statement, it is clear that if the UIA plans to produce the Yearbook using computer typesetting methods, it would be an advantage to avoid increasing the cost of data preparation by attempting complex indexing/coding of each entry. The expensive non-mechanical coding for research purposes (as opposed to simple index keywords for the Yearbook) should be done in a systematic manner for a whole group of entries. This could best be done by producing a deck of coded cards for all the organizations in the Yearbook and possibly their meetings.  These could be quickly and cheaply sorted according to any combination of characteristics for research purposes.

The important point is the range of questions which the UIA wishes to be prepared to answer and the precision required, in the answers. The more questions the UIA wishes to be in a position to answer, the higher the initial and operating costs will be. The ability, to answer such questions will stimulate commercial inquiries but there is no indication of the size of the market for this sort of information.

Initially it would be an advantage to use a code card in which only a few items of information are punched. These could be used to produce the standard Yearbook statistics. As contracts were obtained for further details, these could be punched into the free positions on the card. In this way the research ability would be built up on the basis of market requirements.

Conclusion on Research Processing: The cost of using coded cards and standard search programs can be limited to the time spent in coding and the cost of perforating the cards. The processing cost can be kept to a minimum by using standard programs (either classical sort equipment or a tape sort or a computer). These need only be used when the research program is required. The combination of the research problem with computer typesetting would increase initial and processing costs without any guaranteed improvement


creative commons license
this work is licenced under a creative commons licence.