There have been no changes in staff this quarter. However, next quarter will see the arrival of our other two sandwich students, Gary Symes and Mike Davies who will be working with Mike Elder and Jean Crow respectively. It will also, sadly , see the departure of Peter Sutterlin after his year's attachment - although we shall see him again briefly in November when we organise a workshop in conjunction with COGEODATA on communications between databases. We congratulate Mike Chiu and Pella Machin on gaining their promotions from 1 June.
The present staff in the group are:
| S Backer | Sandwich Student | (SB) |
| D A Byfield | Secretary | (DAB) |
| M F Chiu | Attached to FR80 | (MFC) |
| J E Crow | (JEC) | |
| M Elder | (ME) | |
| E M Gill | (EMG) | |
| M F Guest | (MFG) | |
| P Kent | (PK) | |
| J W E Lewis | ½ attached to FR80 | (JWEL) |
| P A Machin | (PAM) | |
| V R Saunders | (VRS) | |
| P F Smith | (PFS) | |
| B Stokoe | (BS) | |
| P G Sutterlin | (PGS) | |
| A J H Walter | (AJHW) |
A procedure due to Davidson has been incorporated into the MUNICH-CI program, the aim being to produce highly excited state wavefunctions in as automatic a fashion as possible.
CI calculations on various states of the systems HeO and H3 have been performed, in conjunction with members of Professor J N Murrell's group, with the aim of producing potential energy surfaces for use in scattering theory calculations.
Work has commenced (in collaboration with Dr B T Sutcliffe, York) on a configuration selector for the CI program. The method is based on perturbation theory.
We have started on work aimed at incorporating restart facilities in the CI program, partially based on code obtained from Drs Diercksen and Kraemer (Max Planck Institut , Munich).
The major work has been the design and implementation of an infinitesimal perturbal theory program for the calculation of NMR spin-spin coupling constants including the Fermi contact, spin dipolar and nuclear dipole-electron orbital terms (plus cross terms). This work is now complete.
Coding the the Boys localization program has commenced, but the program is not operational.
Design work has been completed for a '3-density matrix' SCF program. This should prove of use for the study of certain excited states which are becoming increasingly more experimentally accessible (and therefore theoretically more important).
This meeting took place at the beginning of the quarter at the Queen's College. Oxford and appeared to run very smoothly apart from a breakdown of an overhead projector.
Feedback indicates that the participants enjoyed the conference and that it was useful to them. It has been decided to publish the proceedings of the Conference and we are now awaiting one or two outstanding manuscripts.
At the end of the last quarter an application was made to the Physics committee for. guaranteed time on the 360/195 by Professor N H March, Dr P J Grout (ICST) and JWEL, to investigate systems exhibiting pyroelectricity. This application has been successful and work should start on the project during the next quarter.
The lattice models of two dimensional liquid crystals are now fairly well-understood, but at the moment one of the thermodynamic parameters is inconsistent and we are trying to find out why.
The results for a sample three dimensional liquid crystal are being written up.
This heading covers the ILL project. The preliminary Monte Carlo calculations are now complete and the pair distribution functions have been evaluated. The structure factor has been evaluated by Fourier Transform methods, but extrapolation will be required in order to study the desired region of reciprocal space. Fortunately an extrapolation method is available.
With the arrival of the FR80, SB has begun to make movies from the data provided by A Rahman. We have not progressed very far because of skew incompatibilities between the 195 and the FR80, making the 195 tapes unreadable.
The Rahman data is also about to be used by Professor Beeby's group at Leicester to calculate the Placek correction fer the neutron scattering of water.
The service continues well and a further 13 data sets of Weissenberg films have been scanned. This is a bit up on the average for a quarter's work. One particular set of films is worth mentioning - they had previously been measured by eye giving a crystallographic agreement factor of 14% for the solved structure. After scanning on the microdensitometer, refinement reduced this agreement factor to 6.7%.
Users are slow to report their results, but the solutions of the structures which we have heard about are still giving agreement factors in the range 5-8% with a mean or about 6.4%.
The program for scanning precession films had previously been tested roughly, and has now been used to scan both a set of small molecule films and two packs of protein films. The former was scanned successfully with only minor changes to the program. Symmetry agreement factors in the range 2-4% were obtained. Modifications were made to the background calculation in the scanning program to cope with the more varied background on the protein films.
A FORTRAN program was written and checked out on the 1906A to index precession films. This has been used to index both the sets of films which have been scanned, and can now be used routinely.
During this quarter it has become routine to perform some processing of the raw intensity data for users with the SHEL-X program suite. This extension of the scanning service is, we believe, very helpful to users whilst at the same time being economical of 1906A time since the SHEL-X routines are fast and reliable. We have discussed with some XRAY users the possibility of their using the SHEL-X system as an alternative.
A few problems this quarter - the magnetic tape deck gave trouble, and Sintrom who now contract to service it took it away to mend it, which meant we were down for a couple of days.
The scanner has functioned well. It was noticed that the readings obtained.for the standard 2D and 3D filters were no longer as high as they should be. NMP did a sequence of adjustments to correct this. The ALPHA-16 processor board returned from Computer Automation after 3½ months. The standard tests work but the Autoload facility no longer functions. So the board has gone back yet again. The memory board which was also sent to CA in February has not yet reappeared - we are very pleased that we managed to borrow the spare machine from Cambridge!!.
The 1906A still supports both XRAY-70 and XRAY-74. The IBM 370/195 runs the 74 version. Very little work has been done on the system this quarter - with only one new bug appearing and no new development work.
Other universities continue to show interest in our "advanced" 74 version, and we will be sending copies to other ICL users.
Interest is being shown by both Sussex and Oxford universities in using the protein refinement program RS4E.
In both cases the electron density maps are generated on a 1906A (ours or Oxford's) and with a great deal of assistance from MFC we can now read these on the IBM 370/195 as input to the refinement.
The use of the program is still at an experimental stage, but one test job has been run so far for Sussex, and Oxford's use is imminent.
Work has just begun on the display system required before interpretation of such refinements can be made. This is particularly important for the work with the Trisaccharide - Lysotyme complex which is underway.
A project with Dr L Nwoye of Hull University has been started involving digitising strips of 16mm film which display diffraction patterns from frog muscle fibre exposed to the light of a ruby laser. The aim of the experiment is to measure the differences in the diffraction pattern caused by muscle contraction and relaxation. A program has been written and the initial results are encouraging.
The refinement of the phospholipid structure of R Mason and P B Hitchcock has been successfully concluded. The laboriously collected microdensitometer intensity data repaid the efforts involved in writing the programs by reducing the crystal structure R-factor from the original 30% (visually estimated data) to 16.0%. A small error in the original structure was corrected, but otherwise the postulated structure is now confirmed, and with sufficient accuracy to convince the previously sceptical. The off-line processing program is now available for use with problem films, providing we are convinced that the results will justify the effort.
More widespread use of ATS is still inhibited by lack of a suitable typewriter terminal; the one on order still has not arrived. The LEARN ATS program was made available through Paul Thompson of RL towards tbe end of the quarter, and was tried by MBS who found it helpful. Before this was available the basic ATS system was demonstrated to Mr Walkinshaw's secretary from RL; she also seemed quite impressed. We are still getting used to the system, exploring its facilities and using them to prepare documents (we must learn 'house style', though). In the next quarter we hope to get to grips with the ATS program itself as some modifications may be needed if RL phase out their 2314 disk packs. When the typewriter arrives, we hope the secretarial staff will like the system sufficiently to use it for their everyday work; at the moment the 360 management is agitated because ATS is not used more heavily.
Some time (about two-three weeks) was spent preparing a talk on FORTRAN and data handling for bulk data. This went well although the time for the talk was insufficient. The slides for the whole talk have been copied and sent to the attendees.
The (1972) QMC version of the BCPL system has been mounted. A utility program on the lines of ATS/360 has been written on the 1906A. This is a program for formatting text and was written to help in writing the ALPS manual. It includes most facilities of ATS relevant to the production of manuals by a non-interactive program.
Full implementation of PAFEC (Program for Automatic Finite Element Calculations - obtained from Nottingham University) on the 1906A is almost complete.
A "PAFEC" macro has been written to ease user access to the package, reducing excessive line-printer output and giving the user more control over the PAFEC system without requiring a detailed knowledge of the package or the 1906A operating system. A facility to access information about the running of a PAFEC job in the event of a system failure has also been provided.
Provision is made within the "PAFEC" macro for the job start and job finish times to be output to user files, and also to a central file together with additional information about the user job, so that a record can be kept centrally of the usage of PAFEC on the 1906A. It is hoped that this will provide a valuable feed-back to central staff and allow informed policy decisions to be taken on future PAFEC support and development.
Structural drawing on a TEKTRONIX visual display is available; a Hewlett-Packard plotter may also be used but this is not encouraged due to the slowness of this device for drawings of any substance. Production of hardcopy is not yet a user option.
A number of research engineers are now using this package. In the main these users are not inclined to study the rather cumbersome manual provided by Nottingharn, nor to learn the fine details of the GEORGE IV system. A lot of time has therefore been spent in advice on data preparation, use of the package, and general "hand-holding" support.
A routine for the automatic generation of a finite element mesh has been obtained from John Reid of AERE Harwell. In collaboration with John Reid some modifications have been made and a program written to use the routine to generate a mesh for analysing the stress distribution in a carbon fibre-reinforced material. This work formed part of our support to R G C Arridge of Bristol University, who sought advice on how best to tackle this problem.
Other mesh generation programs, TRIM (in use at Rutherford Laboratory by Alan Armstrong) and POTENT (written at Culham Laboratory by Chris Thomas) have also been investigated. They did not, however, prove particularly useful for solving over an irregular region with general boundary conditions.
A problem of free-surface flow over a weir, sent to us by Dr A M Binnie of Cambridge University, is being tackled using John Reid's routine to generate a mesh on which to solve Laplaces' Equation.
A study has been made of currently available FE packages and graphics systems which might be implemented at ACL.
During the information-gathering process we invited a speaker to ACL to talk about the FE packages STRUDL and NASTRAN, visited the CAD Centre at Cambridge and Lloyds Register of Shipping in London, and JEC also visited engineering establishments in the US.
A project has now been set in motion, in collaboration with Basic Software Group, to establish a graphics system on the PDP15 linking to an FE package on tbe 1906A, for interactive data manipulation and results interrogation. The listing of a module, PAMGEN, of LUISA2, a graphics system supplied by Geoff Butlin of Leicester University has been obtained together with documentation, and a listing of I/O routines of W S Atkins finite element package ASAS is also being studied, with a view to implementing these packages at ACL.
A day symposium on Digital Simulation Techniques was held at ACL on 4 April. This was also the Easter meeting of the Computational Physics Group of the Institute of Physics. An abstracts booklet was produced and about 40 people attended from research establishments, universities and polytechnics.
Towards the end of the quarter the first processing of the data was completed. This represents the fulfilment of ACL's original commitment to the S68 project, a commitment which had already expanded: to compensate for the failure of the onboard tape recorder the satellite was kept active for a total of 17 months spread over a period of 2¼ years while the original plans envisaged a single scan of 7½ months. We had to omit from the processing the data acquired during the last month or so of the satellite's activity as it was not moving smoothly then. It also turned out that some other data tapes did not need processing since the data they contained was no use to us as it had been recorded during the hibernation periods. In all 132 tapes have been processed.
Development of the main processing program for the second scan is complete but there is some more work to he done on the subsidiary programs. Also the finishing touches have to be put to the new and greatly extended reference catalogue we will be using. In a month or two these tasks should be complete and the second processing should be under way; in the meantime our CPU use will probably drop sharply from its recent high level.
The main library of object modules has now been put on permanently mounted disk space, which has reduced the number of SETUPs needed and improved turnround for library updating jobs; this has speeded up program development considerably.
It now appears that Dr Thompson will be able to use the 370 from ROE, via the existing RJE station and link at present used to access the 1906A and an onward link to RL will not have to wait for the 4080 to be available as a front~end processor. The facility is almost ready; it will only be inferior to the promised 4080 system in one or two minor respects. We have taken the opportunity to transfer some of the more heavily-used functions, particularly the library updating, to an ELECTRIC-based system since this will make things much easier for remote users and should further improve turnround for these jobs. The source version of the members of the main library will be held in a partitioned data-set on permanently-mounted disk space: they will be read into a temporary ELECTRIC file (one at a time) for editing. A set of programs to list, compile and copy these source versions has been prepared; some are based on software acquired from M F Chiu.
Some more work has been done in program development for Dr Morgan (ROE): he has become much more 'self-supporting' and will presumably become still more so when the 'onward link' to the 370 is available.
June was enlivened by some rather extraordinary happenings. A meeting of the S2/68 Joint Executive Committee due to be held in Liege, was cancelled at four days notice when a telex arrived from Liege stating that the Belgian team working on the project was being disbanded, no more work would be done on the Belgian half of the project and so there was no point in holding the meeting. Professor Wilson and Dr Thompson went to Liege to find out what was happening.
Some of the junior staff working on S2/68 in Liege had been removed from the project for reasons of economy although the whole team was not really being disbanded. In the resulting confusion we know nothing more except that some authority in Belgium has set up a committee of inquiry - perhaps we may see some new faces at the next meeting! The main Belgian project (a new catalogue of bright ultra-violet stars) was very nearly complete anyway.
Most of the commands have now been written and tested. The MERGE command has been abandoned through lack of time. Work is still in progress on the CLPLOT, GPLOT, DUMPBN and RESTBN commands and on the BOON Mark 2 macro.
Work on the 19O6A has been held up by the poor state of the machine, difficulty in getting through to ACL on the telephone, joblimit reached, and slow postal service. There are also problems with lower data, which we hope to overcome by overlaying the program. The manual has been written and typed apart from a table which gives approximate execution times for several of the more important commands. This can be added later.
The program and manual should be ready for release for use at ACL within the next few weeks.
The current state of the ALPS project is:
In the initial version, due for experimental release very soon only a subset of facilities will be provided. The main omissions are
A vast amount of de-bugging still needs to be done, mostly at the level of very fine detail. It seems unlikely that major revisions will be needed in order to provide a full working system.
Progress at the IGS London installation has been slow this quarter because much of our effort has been directed towards trying to convince IGS that G-EXEC should be retained as an in-house system. We felt that the feedback from users and exchange of programs with other geological surveys would be in IGS's interest and ensure that G-EXEC best met the needs of the geological community at large. However, it now appears that the development and maintenance of G-EXEC is to be taken over by the planned NERC unit, probably at Chilton thereby losing the direct contact with the geological users. The IGS Computer Unit in London currently consists of two systems analysts, four programmers and a clerical officer, all of whom carry on their normal work within the framework of G-EXEC. This situation is to cease in the near future; two of the programmers, the clerical officer and one of the analysts, are leaving within a month, the other analyst is being advised to join the NERC unit already mentioned as soon as it comes into existence.
The existing G-EXEC software is being used more extensively than ever within IGS, and several new applications are now in regular use, in particular a sand and gravel resource evaluation suite has part I operational, and a computer-produced turnround document is being used for new data in the Mineral Statistics data bank. The Hydrowell data bank (transferred from Atlas in a new form) is now being loaded entirely by geologists, with no assistance from the computer unit; this is indeed the general rule, as users are getting more confident at handling their own data. The production load of IGS computing is already far more than the Computer Unit could possibly generate by itself, demonstrating the desirability of user participation.
The outstanding news of the quarter is unquestionably Steve Henley's success in implementing G-EXEC in Costa Rica at the Geological Survey, under an ODM contract. He implemented a basically geochemical subset of the full system, selecting about a quarter of the total library to leave permanently available, with the rest for them to add as they required - tape and disk space was at a premium. A new one-pass Controller rather like the PDP 11/45 one was written specially by Steve, and between times he gave courses and was written up in the local press on several occasions during his 6 weeks' stay.
Many requests have been received for G-EXEC, among them from the Dutch Soil Survey and the Swedish Geological Survey, both of which were originally approved hy ACL but have now disappeared into a legal tangle. Other requests have come to IGS directly; we are assured there are many of them, some from commercial organisations. All are awaiting a decision by NERC or SRC, or NRDC or possibly worse still a combined committee of all three.
John Cubitt, the fourth member of the original team, leaves on 4 August to take up a post at Syracuse University in New York State, under our old friend Professor Dan Merriem. John's statistical expertise will be sorely missed, as well as his ability to do an incredible amount of work and still stay cheerful and optimistic. With the present uncertainties, it is difficult to make firm plans; real-time implementation of the Controller and new automatic archiving facilities were planned for this summer and autumn, but we are not yet certain whether these plans will be carried out.
| Brunel Conference Report, 30 April 1975 | J E Crow |
| Note on NASTRAN and STRUDL, 1 May | J E Crow |
| Interactive Graphics for FE Packages, 21 and 30 May | J E Crow |
| Interactive Computing for Engineers - Notes of a visit to CAD Centre, Cambridge, 5 June | J E Crow |
| Interactive Computing for Engineers - Notes of a visit to Lloyds Register, 6 June. | J E Crow |
| Discussion paper on the proposed ACL Engineering Interactive Facility, 20 May | J E Crow |
| Interim Report on visit to US, 25 June | J E Crow |
| "The BOON System - A Comprehensive Technique for Time-Series Analysis" | P Kent (+ Lancaster) |
| SRC Microdensitometer Service at ACL Report for the year ending 31 March 1975 | P A Machin, M Elder |
| SRC Microdensitometer Service: User Note 2 | P A Machin, M Elder |
| A Second Microdensitometer | M Elder |
| Seminar to Oxford Biophysics/Crystallography Group, 17 June 1975 | M Elder |
| Discussions with Professor Mason and Dr Hitchcock, Sussex Univ on 9 May | M Elder |
| Discussions with Professor Mason and Dr Hitchcock, Sussex Univ on 9 May | M Elder, P A Machin |
| Meeting at State House on Crystal Structure Searches on 19 June 1975 | P A Machin |
| Meeting at the Physics Department Oxford on 20 June 1975 to discuss computing facilities for crystal structure search programs. | P A Machin |
| Institute of Phyics meeting on Digital Simulation Techniques held at ACL on 4 April | B Stokoe, J E Crow et al |
| Liquids Conference held in Oxford on 14-16 April | B Stokoe, J W E Lewis et al |
| Users seminar held at ACL on 17 April | B Stokoe + others |
| ALPS Progress Meeting held on 22 April | B Stokoe, P Kent, A J H Walter |
| BOON Progress Meeting held on 23 April | B Stokoe, P Kent |
| Joint Applied Mathematics Seminar held at ACL on 24 April on Boundary Value Problems. Talk given by Dr R England | B Stokoe + others |
| ALLC meeting on Languages, held in Oxford on l May | B Stokoe |
| Meeting to discuss joint ACL/COGEODATA workshop held at ACL on 7 May | B Stokoe |
| Image Processing Meeting held at ACL on 12 June | B Stokoe, P Kent et al |
| ALPS Progress Meeting, held at ACL on 19 June | B Stokoe + others |
| IMA Meeting on Cosmology held at Burlington House on 27 June | B Stokoe |
| Boon meeting to discuss the Mark 2 manual held at ACL on 5 May | P Kent |
| Finite Elements and their Applications held at Brunel on 7-10 April. | J E Crow |
| Talk on NASTRAN and STRUDL held at ACL on 24 April | J E Crow + others |
| Visit to CAD Centre, Cambridge on 3 June | J E Crow, J W E Lewis |
| Visit to Lloyds Register held on 6 June | J E Crow, J W E Lewis |
| Visit to US Engineering establishments + AICA International Symposium on Computer Methods for Partial Differential Equations, Lehigh Univ, Bethlehem, USA from 8-21 June | J E Crow |
| Visit to Southampton to discuss liquid crystals with Dr G Luckhurst; on 24 April | J W E Lewis |
| Visit to Imperial College to have discussions with Prof N March on pyroelectrics on 29 April | J W E Lewis |
| Calculations on water - discussions held with Professor Beeby at Leicester Univ on 14 May | J W E Lewis |
| 13 June - lecture at IBM East Croydon on CSMP | J W E Lewis |
| Discussions with Ray Mountain (NAS) and M Dixon at Reading Univ on liquid Crystals on 18 June | J W E Lewis |
