The question raised is whether the project was cost-effective? That question cannot be answered without considering the whole Common Base Programme and what has been achieved over the period. About 27 man years of centrally funded effort has been put in at RAL together with an additional 8.5 man years of effort provided directly by the Engineering Board from other programmes.
What have been the benefits achieved on top of the effort provided by ICL?
The breakdown of 27 man years is approximately:
| Management | 3.5 |
| Assessment | 3.5 |
| User Support | 5.5 |
| Hardware/Operations | 6.5 |
| Systems Development | 7.0 |
| Communications | 1.0 |
with the additional EB effort providing further systems and communications effort:
| Systems Development | 6.0 |
| Communications | 2.5 |
The exact breakdown is not clear but in general the additional effort provided by other projects was to increase the level of software available in particular areas.
The Science Board PERQs other than DL were:
Systems 42-44 Systems 61-70
of the systems delivered. All were delivered during the first half of 1982.
Effectively, in the period to March 1985, the costs were:
The original purchase price of the Science Board systems were:
List Price SERC Price
42-44 $52,000 $41,600
61-70 £26,600 £19,950
This represented a saving in purchase price amounting to £85K as a result of the Common Base Policy discounts agreed between SERC and ICL. While almost certainly some discounts could have been achieved by bulk purchase, it is not unreasonable to assume (looking at the discounts available from other suppliers), that the Science Board probably received a discount of at least £56K or £4K per system as a result of the Memorandum of Understanding.
These systems were upgraded as a result of negotiations with ICL to at least 0.5 Mbyte systems at no cost to the Board. The cost of this upgrade would have been £2.75K.
Finally, the 16K WCS upgrade was achieved at a cost of £1.4K instead of the list price of £3.2K, a further saving of £1.8K.
Thus during the 3 years, the Science Board received hardware discounts amounting to over £8.5K per system. Thus, per year, the Science Board received a discount in hardware costs amounting to £2.8K per year. There is also some saving from the cost of the PERQ2 awarded to Dr Holbrook.
The maintenance agreement with ICL, as part of the Memorandum of Understanding, produced a discount of £0.25K per system per year.
As a result of the Memorandum of Understanding with ICL, Science Board users have received software from ICL free of charge and benefited from the software added to the system by the joint developments. If the Common Base had not been in place, software would have had to be purchased from ICL. There are two possible approaches:
Over the last few years. the new releases of products have been approximately one per year. The main items of software are:
£
PNX 1582
Bell F77 1412
ERCC F77 1582
PASCAL 1520
GKS 1415
TOTAL 7511
Assuming the site continues to receive releases, the maintenance contract is attractive, reducing the £7.5K cost to £2629 + £535 - £3164.
The cost of the products such as SPY and WW developed as part of the collaboration would be conservatively worth £2K without the maintenance contract and £700 with it.
Consequently, the savings per year in software terms is of the order of £4K per system per year. It is worth noting that the PNXSR release for PERQ1s is unavailable as a standard ICL product and could not be obtained from them.
Even using the most conservative method of buying software from ICL and maintaining a single primary licence for the whole of SERC, if ICL had allowed this, the cost of software for PERQs per year would have been around £200K. The Common Base costs can nearly be justified on this alone.
A more difficult assessment is to decide the worth per year of the PNXSR upgrade. Without it, the Science Board PERQs had a life of about 4 years before becoming obsolete. The PNXSR release extends their life to about 6 years.
Taking the reduced hardware cost of £20.3K per system, this effectively gives a saving of approximately £1.5K per year in hardware costs over the 6 year period. Alternatively, it can be seen as a saving of £5K per year in the next two years as a result of not needing to purchase new systems.
The total savings per year to the Science Board in real money have been:
£K pa
Hardware 2.8
Maintenance 0.25
Software 4.0
PNXSR 1.5
TOTAL 8.55
In addition to these direct savings, the Science Board has benefited from:
RAL would assert that there are clear benefits that can accrue from the Science Board continuing to provide funds to allow their systems to be part of the Common Base. A decision to remove them from the Common Base will have the following effect:
The total savings in the Common Base Project amount, approximately, over the period 1982-85 to:
£K pa
Hardware 1000
Maintenance 100
Software 600
Upgrades 90
16K WCS 250
Extension to PERQ1 Life 1000
TOTAL 3040
The cost of the programme funded centrally was:
£K pa
Staff Costs 600
Recurrent 300
TOTAL 900
Thus on a straightforward cost benefit analysis, the Project had a substantial saving in funds to SERC. Even so, there were a great deal of indirect benefits that should also be included:
The UK now has a thriving Single User System industry with at least 3 suppliers who were connected to the programme in some way. ICL are a major force in this area and it would not have occurred without SERC's initiative. The origins of Whitechape1 come from Logica and QMC's involvement in the Common Base Project. High Level Hardware were kept in touch with developments and encouraged to put a high quality display on their system. Both hardware and software consultancy was supplied to High Level Hardware to enable them to produce a product.
Of countries outside the USA (and possibly including the USA), the level of awareness in the UK is easily the greatest. We have a breadth of experience in single user systems that should provide a launch pad for competitive products in this area.
Despite the obvious setbacks in the programme, RAL still believe that the project, considering the widest UK involvement, was a great success and cost effective to SERC in simple monetary terms.
Taking a narrow SERC view, an alternative would have been to continue using multi user systems for at least another two years and start the Common Base programme later. While attractive, it does not take account of the fact that DCS had grants that needed equipment of this type at the start of the programme.
Starting the project two years later would have meant a diverse set of equipment being purchased for research which would have had similar problems to the ICL PERQ but with little muscle or support to influence the company's direction. The later assessment has shown that when the PERQ was chosen there were no alternatives and nearly 4 years later there is still no system that is significantly better all round.
For the future, the UK has two very competitive low cost products in the ICL PERQ3 and the Whitechapel MG-1. This would not have been achieved without the Common Base Project.
Currently the Science Board pays the Common Base project £5K per annum to provide maintenance and support of the systems as part of the Engineering Board Common Base under the agreement with ICL. Continued payment of this amount will provide, inclusive of VAT:
Total is £7518.
The alternatives are:
Particular points need to be made:
Dr Dodson's PERQs which are still running POS get less benefit from the Common Base. With 3 systems on a single site, a good contract with one primary and two secondary licences could be obtained from ICL which might be cheaper than the SERC cost and more beneficial to Dr Dodson.
It is recommended that the Science Board look at the long term needs of Dr Dodson. If he intends to remain with POS, an alternative could be considered.