For some time, it has been apparent that a document defining the Policy of the Computing Division would be helpful for staff within the Division. The Group Leaders, together with all staff at the PSO level in the Division, have met on a number of occasions to consider drafts of a document defining the Computing Division Policy and another giving background to it.
Both documents have been agreed for issue within the Division to staff of SSO grade and above or equivalent. All members of the Computing Division should be aware of the policy of the Division and its influences, and it is the responsibility of line managers to ensure that staff in the grades not receiving these documents are made aware of those aspects which affect them. For the present, these documents are not for publication outside the Division.
It is planned to review the policy documents at six-monthly intervals. Comments on the content of the documents and upon the policy itself should be brought to the attention of Group Leaders.
E B Fossey
27 May 1982
The purpose of the SERC is defined in Paragraph 2 of the SERC Charter as follows:
2. (1) The objects for which the Council are established and incorporated are as follows:
The contribution of Computing Division to these objectives is to provide computing facilities for SERC Laboratories, SERC supported researchers and other Research Councils. In doing so, part of its resources will be devoted to research aimed at improving the quality and techniques of computing for the benefit of SERC, and of the universities and the computing world in general.
The RAL Computing Division is the major source of computing expertise in the Council and it should play a major part in setting the quality, standards and direction of computing in the Council as a whole.
A rapid expansion in the use of information technology is expected in the next decade. In order to play a full part in this expansion, the Division is expected to increase in size. No arbitrary constraints on such increases will be applied. The size of the Division will be defined by the functions it carries out.
The Division will continue to involve the universities in its activities wherever practicable. Contracts for specific projects will be placed with universities when manpower limitations prevent them being done at RAL, and when such contracts would be of benefit to RAL or the university.
The policy will be reviewed and updated at least twice a year.
The remaining sections detail the major functions of the Division. The percentages of effort quoted are expected to refer in practice to the manpower requirement.
3.1 A comprehensive computing service should be provided for the RAL site, SERC researchers in universities, other SERC establishments, and other Research Councils. The various activities that contribute to the maintenance and development of this service will constitute at least 50% of the Division's activities. The total direct effort should not drop appreciably below the current level.
As technology allows savings to be made in Operations, effort will be directed into positive user support and consultancy. Thus, increased efficiency will result in an improved service to users, not the same service with lower staffing levels.
3.2 The Division is responsible for communications on site, and for ensuring that all SERC users have access to its facilities.
3.3 The Division will pursue an active policy of standardisation in hardware and unaltered proprietary software wherever it is sensible to do so.
4.1 The Division will actively support SERC's research programmes in computer science and related subjects (currently through support of the Information Engineering Committee). This should form at least 20% of its activities.
4.2 Specific efforts will be made to ensure a close relationship and mutual benefit between these activities and the service side of the Division.
4.1 The Division will support Board or other funded work on a contract basis. This should form about 10% of its activities.
4.2 The projects will be accepted if there is some relevance to the other work in the Division, or if they are potentially of value to the future work of the Division.
5.1 The Division has a role in improving the quality of computing throughout the Council. Education (primarily advanced training) and consultancy not specifically related to the service should occupy about 10% of the Division's activities.
5.2 All staff are to be encouraged to spend a tenth of their allocated time in professional development.
7.1 The Division undertakes research and development activities related to the needs of SERC in general and the RAL site in particular. This is separately recognised and accounted and should be at least at a level of about 10%.
7.2 Where appropriate, staff are encouraged to develop cases for research to be funded through formal SERC mechanisms. Otherwise it is funded within the Division through a charge on all projects.
8.1 The Division will ensure that its members conform to the highest professional standards. The effectiveness of the Division in meeting its objectives will be monitored, and continual improvements will be made in the methods by which projects are managed and implemented.
8.2 Members of the Division will be expected to conform to ethical standards of the highest level. Those defined by the British Computer Society are taken as a model.
Members of the Division are expected to document all work and are encouraged to publicise appropriate work by producing reports and submitting papers to journals and conferences.
This document is intended to accompany the statement of Computing Division policy. It explains the policy and interprets it in the context of the current divisional activities. It is aimed at helping those working in the Division to understand the goals being aimed at.
Clearly we do not have the right to define our policy in its entirety. Our policy is constrained by that of the laboratory and of the SERC itself. We can be instructed to take on certain work. We have complement and financial constraints, although unlike commercial organisations we are not required to cover future costs by current income (i.e. 'make a profit'). However there is no doubt that the way we use our energies in pressing forward on certain fronts has a great effect on the eventual distribution of funding. A diversification of funding sources has already begun and will continue. Much of our work is now paid for by the Boards, DOI has invested in the Office Automation projects, and JNT are funded jointly with the Computer Board.
In this changing environment, we Must be aware of our objectives, long term strategies and short term plans. The policy document is the beginning of this process.
The general comments at the beginning establish Computing Division's major role in computing, not just in respect to its own users or to the laboratory, but to the whole of SERC and the community it supports.
There are many advantages in having a maximum divisional size of 150 or so - it is then quite easy for everyone to know each other and the management can be fairly free from bureaucracy.
However such an artificial constraint may have a strong effect on what we are willing to take on, and it was decided not to impose any such constraint. Thus an increase in size to 200 or more is not unlikely.
The effort levels associated with the various functions refer to a proportion of the total direct staff, not necessarily the financial resources.
'Service' refers to the role currently performed by User Interface, and most of Systems and Operations groups, plus the Graphics section. It consists of providing and developing computing facilities - hardware, software, means of access, advice, etc. to users in general, with funding supplied centrally. There is, of course, a large service element in most of our Contract work (see section 5) as well, but these projects are funded on behalf of specific groups of users and we do not necessarily have the same freedom of action as with centrally funded services.
The division originally existed solely to provide services, and the interest and motivation of many of its staff derive from this role. A large continuing service role will give a stability of income and a demanding environment to validate technological developments. The service function should occupy not less that 50% of the effort. Although this is considerably less than now, the actual numbers are not expected to decrease because of the expected increase in the size of the Division. Thus the current service levels must not be allowed to decrease but should improve.
Strategies will be defined in the following areas.
The need to keep abreast of technology and to meet the requirements of a wide range of scientific users dictates that the service must be based on a plurality of hardware and software architectures. The service will probably continue to be offered on a mixture of central facilities, distributed minis and personal computers. The latter will be used for general computing and highly interactive applications, while central computers will provide high speed number crunching and large scale data processing. The Division must be active in ensuring that it can provide for the extremes of use demanded by the SERC scientific user.
Storage is going to be difficult to distribute in the quantities required in SERC. As an example, the Central Computer tape library contains over 60,000 tapes, many full of data recorded at 6250 bpi. The Computing Division should regard itself as having responsibility for storage of information within SERC when asked. We should be more active in doing R&D in new technologies such as video discs.
Communications within the RAL site to the Division's facilities and communication around the site should be regarded as a Computing Division responsibility. The future responsibility for the CB/SERC network is still not clear. However the Division has to ensure that its users can access the facilities at the required rates. It has acquired great expertise in wide area networking, and should continue to have a very strong influence.
Within the constraints implied by the above requirements, the Division is committed to minimize support costs by pursuing an active policy of standardisation in hardware and unaltered proprietary software wherever possible.
Support is open-ended - almost any level can be defined as useful. We should adopt the minimum day-to-day level of user and system support consistent with achieving a defined standard of service acceptable to the user. However; in addition we should have an aggressive approach to consultative support - assisting professionals to map their problem to the hardware and software available, without being locked into it. The aim is for SERC to do more cost-effective science by getting the most efficient use out of its own staff and equipment, and out of the scientists it supports. Some way must be found to monitor the effectiveness of this.
Software support at the moment has insufficient emphasis. Long term, the raison d'etre of the Division will be its software expertise rather than its hardware. We need to take a more dynamic role in supporting applications software like graphics, text processing, file handling, databases, software tools, data management etc.
Crucial pieces of software that are fundamental to the computing service provided to the users must be supported centrally and not by an outside agent unless we have a contractual commitment from the other party.
We should continue to actively support SERC's research programmes in computer science and related subjects. The support of those areas of University research which fit with the Division's other activities are of particular importance. For example:
Information Engineering in the wide sense will be an important ingredient of future computing. We should be actively involved in such areas as databases, knowledge based systems and office automation because of their interest and involvement in projects throughout the Division.
SERC supports computing research in the expectation that it will lead to developments useful to the computing world in general. Specific links must be maintained between the service side of the Division and the research-support side in order to ensure mutual awareness and benefit.
We currently support some projects on a contractor/customer basis. Examples are Starlink and Grants & Awards. Contracts are funded by other Divisions, a specific Board or an outside agency for us to do a job - typically for an identified group of users. Such projects can be interesting, a good discipline, and the Division can gain in status.
However there have been problems where funding is inadequate, or the work had no relevance to the rest of the programme. Contracts should thus be considered carefully.
We should attempt to get representation on the relevant funding body if we accept projects of this type.
Given that we do support these sort of projects, we should attempt to be involved with all the Boards of SERC. This should be taken into account when considering projects of this type.
This section covers advanced computer training of SERC personnel in general - not training of our own users to use our facilities, which is covered in the Service role.
The Computing Division at RAL is SERC's main repository of expertise in computing matters. We should expect and be prepared to give help to other Divisions and Laboratories on computing matters. This should not be regarded as a chore but as part of our function. This is particularly true for the smaller Laboratories and Central Office where the computing expertise is not high.
If anybody is responsible for organisation of computer teaching in SERC, it must be the Computing Division. It is not sensible to suggest that the function can be effectively performed by Central Training. On the other hand, it does require a positive decision on the part of SERC and ourselves which will require manpower and the recruitment of staff interested in coordinating such activities. Most disciplines in SERC recognise that part of their role is educating users in their field. For some reason, the average computer programmer feels that it is outside his responsibility. We should be active in making new members of staff aware that at some stage of their career this may be required of them.
The Division should look sympathetically at R&D activities not completely directed at the immediate goals of the Division's programme but ones of relevance to the overall structure and development of the Division's role in the future.
If we embark on such programmes, they need to be agreed by the policy making body within the Division. The amount of manpower available for such activities will be necessarily small. However, even a small amount of effort here is likely to raise the quality of expertise in the Division as a whole, particularly if the R&D activities are slotted into the Divisional structure rather than being a self contained adjunct.
It is vital that we arrange for a formal set of divisional activities that blend together naturally so that new ideas and techniques plus fundamental knowledge can intermix with user requirements and operational experience. Everyone must see where and why activities fit together.
R&D activities should ensure that the functions of long term importance to the Division are adequately represented in the Division's portfolio.
Such research may be supported by a charge on all the funded projects, but it is obviously an advantage to solicit support through standard SERC mechanisms, and this will be encouraged.
The Division is, to a large extent, not consistent and often unprofessional in its management of programming activities. It does not currently have a defined methodology for the development of new systems and their maintenance afterwards.
Change control is not consistently applied across the Division nor do Group Leaders and senior staff have a total commitment to high quality control of change.
A working party in the Division should define a methodology for implementing and maintaining software systems that is consistent across the Division.
For a Division of our size, the number of reports and publications produced is very low. This has a number of secondary effects. The expertise of the Division and its external status is diminished. The promotion prospects of staff are decreased - they are in competition with scientific personnel who are conditioned to publish every minor event. Also, the dissemination of information is often low.
Staff should be clear that it is their responsibility to produce reports or to publish in appropriate journals. Management will regard time spent on such activities as important.
