Science Faculty =============== The Science Faculty have been heavy computer users for many years and most have used EMAS at some stage in the last decade. These departments group themselves naturally into three section - allowing for some large and diverse departments like Physics to appear in more than one section. Group A: Researchers into aspects of Computing. These include Artificial Intelligence, Artificial Intelligence Application Institute, Computer Science, Mechanical and Electrical Engineering, Centre for Speech Technology and Cognitive Science. This group is primarily based on Workstations linked by LAN and has the most advanced equipment in the University Group B: The Collaborators These include School of Agriculture,Genetics, Molecular Biology, Astronomy, Statistics and Physics (part). The needs of these groups are determined by collaboration outside the University or by other external pressures. Apart from some IBM PC use (Statistics) these are based on DEC Vaxes with VMS. Group C: The Timesharers These include Chemistry, Chemical Engineering, Fire and Civil Engineering Physics (part), Meteorology, Geology,Geophysics, Maths and Zoology. These have been heavy users of central timesharing (mainly Emas) and continue to do so though some workstations and departmental machines are appearing. All these departments make some use of micros for word processing, departmental administration of teaching. In general the amount of micro use is small and the comments and conclusions correspond to those found in more heavily micro orientated faculties like Social Science. These concerns will not be documented again in this section. a) Current Use of Computers in Research All science departments are heavy users of computing in various ways 1) Group A This group is primarily based around a LAN equipped with workstations (in practise these are mainly Suns). The LAN has at least one Network File Store running under Sun NFS Protocols and the LAN normally supports the X-Windows remote window management scheme. The operating system is Unix and the main languages are C, Prolog and Lisp. The LAN provides enough sharing for the whole network to resemble a single system. The users are well satisfied with the services provided especially the large high resolution bit mapped screens which enable facilities to be offered that are not available on a central timesharing system. Package software was not an issue although we noted in passing that Engineering users of the finite element package PAFEC had retreated from workstations to a time sharing mainframe because of performance problems. 2) Group B This group was constrained to to use a particular system (almost always Vax/VMS) by its collaboration or by the need to utilize specialized networks like the Astronomers Starlink. Much specialized software was needed and normally made available via various collaborations. Their need to use particular hardware had precluded much comparison with other systems although the shortage of CPU and Disc space on various Vaxes were a recurring theme. The most demanding use was DNA matching based on the Wisconsin package and its large associated Database, which is of interest to most of the life science departments. The demands of this package on both disc space and CPU are clearly exponential and if unconstrained, will soon consume all spare Vax time throughout the University. 3) Group C These people had used EMAS for a long time and were not unhappy. They appreciated the long term stability, managed file and archive stores, good diagnostics and friendly interface. Much of their work is in Fortran, often home written or adapted and problems with portability were classed as minor. The group were forward looking and realized that they would have to move to a different system sooner or later but were concerned that they might suffer a large upheaval for little or no real benefit. There was criticism of Unix partly because of its steep learning curve and also for the effect on scarce departmental funds of running workstations. The free nature of EMAS was a substantial inducement to continue with it. b) Computers In Teaching This large and highly numerate group of departments agreed, almost to a man, that developments in Computers in Teaching would not relieve any part of their teaching loads in the next few years. Artifical Intelligence were particularly sceptical of some of the claims made. Computers would be involved in course work and programming was and would continue to be essential for many students. Some departments used, and most looked forward to, Electronic Mail as the prime method of communicating with students and receiving work for assessment. c) Computers in Administration All departments had at least a micro, some much more, for departmental administration. For some this was just maintaining class lists while did budgeting and accounting also. All were highly critical of the accessibility of the DPU machine and the amount paper and skilled time needed to resolve differences in budgets, grant etc. Pleas for proper integration of Office systems were frequent and heartfelt. d) Likely Developments All departments see a continued substantial expansion in demand that will be constrained only by funding limitations. In particular the researchers see continued computer language developments that will require substantial increases in CPU power. The engineers see large expansion in the use of graphic displays and similar demand for image processing whilst the continued work on DNA will transform the life sciences into voracious consumers of computing power. The parallel machines are of great interest to theoretical physicists and as programming the curious machines better understood it is expected that other disciplines will start to use them e) Improvements desired Almost all departments desire more power, more storage and faster communications provided in ways that are familar without any transitional disruptions. Group C also requested adequate planning aids and especially time to complete any move from EMAS. Meteorology felt that ten years should be allowed, others rather less. A number of other desiderata appeared:- 1) Long term data store There is a growing requirement, currently mainly satisfied via the EMAS archive store, for long term secure data storage capable of handling gigabytes of data for periods up to a hundred years. 2) Office Environment Everyone felt that a proper office environment with consistent access across the whole University is essential for departmental administration and offered the prospect of substantial savings of time and possibly money. This requires full access to DPU and Faculty offices as well as links to academic computing. It is highly desirable that access should be from the same work station/micro as the departmental officers used for their academic work. 3) High Powered Box Some central super workstation (mainframe?) of more than 100 Mips to which large jobs could be offloaded from a personal workstation. 4)Backup Support Some form of assistance to departments in securing the data on their distributed file store; this was expressed as a technical requirement but could equally be construed as a support issue. f) Funding The science faculty is by far the best funded and few of the agonies seen in other faculties were apparent. There were many machines used for less than four hours per day to be seen. Indeed under used capacity in distributed work stations is already a valuable asset if the software to utilize it ever became available. Some of Group C raised the question of value for money and were not entirely convinced that distributed computing was reasonably competitive with EMAS or an EMAS like shared service. These questioned how staff convenience should should be valued and whether the University should not go for the most economical solution rather than throw "Money and Unix" at its computing problems. g) Support The new support arrangements were generally welcomed although Forestry strongly preferred the old Advisory service. What was repeated requested, in various forms, was help with the labour intensive parts of running a departmental machine:- Backup, negotiating and obtaining maintenance, software licensing and general trouble shooting. From these discussions it emerged that the staff officially allocated to departmental computing were much too few - from 25 to 50 percent of the staff EUCS would allocate in similar circumstances. The deficit was partly made up by dedication (eg the man who stays to 4am Saturday morning to complete the Friday weekly backup of his departments Vax) but mostly by stealing research time from others. h) Special needs The prime special need is for a method of doing DNA matching as efficiently as possible. The data and and requirement are both growing so fast that it will be impossible to satisfy on an likely expansion of Vaxes. There are clearly ground for hoping that an interdisciplinary team Biologists and Computer Scientist a greatly improved method might emerge. The commercial significance of this might be large . The second special need is parallel processing which is increasingly being seen as the key to the next technological leap forward. Edinburgh is currently well placed but further expensive and specialized equipment will be needed to stay up with the leaders. i) General Observations The overall impression is of a well funded faculty pressing on in a wise and energetic manner. Computing is accepted and used in all areas. The contrast with other faculties is marked.