This quarter saw the departure of Susan Hockey. She has taken up a teaching appointment in the Computing Laboratory at Oxford University, and we wish her every success. The first of our three students, Simon Backer, has arrived for 6 months, and will be working with John Lewis. Next quarter Mel Cooper will be joining us again, this time on a short-term consultancy for 5 weeks. He will be working with Peter Sutterlin and Elizabeth Gill on communications between databases, with a special application to geology.
The present staff in the group are:
| 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) |
The MUNICH-CI program has been updated to incorporate a multi- root function possibility.
The out of core diagonalization procedure of MUNICH-CI has been extensively modified so that the production of highly excited state wavefunctions is now a practical possibility.
We have performed some practical calculations using MUNICH-CI to test the usefulness of molecular orbitals produced by the ATMOL2 MC-SCF programs in a full CI, with encouraging results.
Future work in MUNICH-CI will be focused on providing better restart facilities and on production of a write-up. Thereafter we will concentrate on the production of a PNO-CI program.
We have completed design and implementation work for software to create and generally organise full mode ATMOL3 dumpfiles which are indexed (for easy retrieval of data litems) and contain much material which should make for much cleaner integration of wavefunction analysis programs, CI programs etc. The integrals program has been modified to produce. full-mode ATMOL3 dump files, and the SCF programs are in the process of modification.
The ATMOL3 UHF SCF program is now operational, although software to perform 'spin contamination' annihilation is not yet complete.
Design work is now complete for a new method of performing the Boy's localization method, to be incorporated into the ATMOL3 system. The method requires an inverter for large sparse matrices which are held out of core, and we believe we now have a design capable of performing this task efficiently. Coding of this program will commence next quarter.
The preparation of these proceedings is now well advanced, and publication may be expected during the next quarter.
The work on lattice models of liquid crystals is virtually complete and will soon be written up. A slightly more elaborate model has been constructed and preliminary results look quite encouraging.
These were described in a paper which JWEL was invited to give to the Liquid Crystal Liquids Forum held in Cambridge.
A week was recently spent at ILL, Grenoble discussing the feasibility of some calculations and attending a workshop on Liquid Dynamics. Some investigation was also made of the computing facilities at ILL and there was discussion with the British representative there to try to gauge what we might be called on to do in the neutron beam area. (This area is a Science Board responsibility.).
The trip was most worthwhile and two collaborative projects are under way. This trip is described in more detail in a separate report.
Preparations for the symposium are nearing completion. The programme is complete and the number of overseas speakers and delegates quite impressive.
The end of the quarter marks the end of the first year of operation of the film-scanning service. In the (somewhat biased) judgment of the people running the service, the year has been very successful, and this is confirmed by all the crystallographers to whom we have spoken. A total of 43 data sets were scanned during the year for 27 crystallographers from 22 different universities, polytechnics and institutes. Of these structures 8 have been reported to us as fully solved and refined. The range of crystallographic agreement factors is 5.3 - 7.5% with a mean of 6.4%. A number of other structures have been solved and are being refined. One structure was also solved in parallel with the microdensitometer data using data collected on a 4-circle diffractometer. The two resultant structures showed no differences and the agreement factors were 4.9% (diffractometer) and 5.3% (densitometer).
The program for scanning precession films is now at the stage of final testing before we begin to run it routinely. It is based upon the Weissenberg program, PROXIMA 1.1, with some extra routines added. These routines were tested with a FORTRAN simulation program and shown to work satisfactorily, before they were coded in Alpha-16 assembler. The program was written with protein precession films in mind as well as small molecule films and one of its first tasks will be the scanning of a set of haemoglobin films from the Wellcome Laboratories.
The first problem confronting a crystallographer when he receives an intensity data set from the service is the need to bring individual film packs to a common scale and to combine the readings for equivalent reflections. We have now implemented, with the assistance of their author, the SHEL-X program suite of Dr G Sheldrick, Cambridge, on the 1906A. These programs perform almost all the X-ray structure solving operations of the XRAY system, though somewhat less generally. However, their strong point, which coincides with a weakness in XRAY-74, is their suitability for processing film data, As part of the film-scanning service we shall now be able to sort, scale and merge reflection data and apply Lorenz-polarization and absorption corrections if the crystallographer so desires.
The only problem this quarter involved the Alpha-16 computer. As though to emphasize the discovery that nobody in this country was prepared to take on the routine maintenance of the Alpha-16, the computer began to produce parity errors that were finally traced (NMP, PB) to a component of a memory board. Two boards were sent to Computer Automation for diagnosis and repair, and are still there two months later. Fortunately we were able to borrow a second Alpha-16 from a Cambridge Engineering department and were back on the air within a few days.
We have begun to process some diffuse X-ray scattering patterns from liquid crystals for Professor A J Leadbetter at Exeter. Some routines have been developed to get the scanner output into acceptable format in the 1906A in order to run the SPROGS contouring program on it. Sample contour maps have been obtained, and discussions are now underway to finalise a film-scan-contour system that can be used routinely.
Research at Bristol under Dr Atkins has began on an extensive series of polysaccharide structures. This work will study the conformation and packing of these structures with the aim of gaining information about the various conformers which occur naturally in tissues. X-ray photographs of these polymers are digitized, and the intensities of the resulting diffuse spots estimated. Some sample photographs have been digitised, and a package of FORTRAN programs has been given to Bristol for them to tailor to their specific problem. The package contains microdensitometer output handling routines, the peak search routine from XRAY74, and the hybrid off-line Weissenberg processor developed for the Sussex phospholipid photographs.
What more can one say? The heading is now a statement of fact rather than of intent. Admittedly the graphics routines are not yet hooked up, but it seems better to wait a bit and use the FR80 rather than worry about an SD4020 interface. A macro has been written, the system released and there are already a few users. Only one bug has subsequently turned up, and it proved to have been present in the 360 version as well. XRAY-70 is still available, but can gradually be phased out.
A number of errors have been corrected in the contouring program, and a few improvements made to other routines. CONTRS, and ORTEP - the molecule plotting routine - have been used quite a lot this quarter, possibly indicating a lack of these facilities at the regional centres and individual laboratories. Some interest has been expressed by the London regional centre in implementing XRAY-74 on the CDC 7600.
A protein structure refinement project in close liaison with Professor D C Phillips' group at Oxford is now well underway. The refinement program, RS4E, from Bob Diamond, Cambridge, has been successfully implemented on the 370/195 and is being used to refine a tri-saccharide lysozyme derivative. This program fits a model structure to the observed electron density map and moves the structure systematically to improve the agreement. Some necessary utility programs have been written, and we now need to prepare a graphical display system that will probably be based on overlaying the contoured density maps with structure diagrams.
Diffraction methods can not only be used to study crystalline materials but can also be employed to investigate such less ordered forms as liquids. These systems can he described theoretically in various ways, but one approach has been through the development of integral equations linking the pair potential and triplet functions with the distribution functions (and the structure factors). Some calculations are being carried out in collaboration with Dr J Woodhead-Galloway (Manchester) based on model potentials he has suggested. The calculation is exact for the one-dimensional model and can be performed for the two-dimensional model using an ad hoc model based upon a series of Fourier transforms.
ATS has been in limited use throughout the quarter, the limitations being due mainly to the short tine (0845-1000 each day) for which the system is available. Loading ATS from ELECTRIC has been fairly straightforward although often there are still the delays due to waiting in queues, failure to establish the link from RL etc. Archiving and tidying the files to re using "roll-in" and "roll-out" is quite satisfactory.
The LEARN ATS program has been delivered and awaits clearance by RL before it is made available. We me still awaiting delivery of an improved 2741 typewriter terminal so that documentation can be produced on ATS compatible with ACL house style.
Once the program and the terminal have been installed, it should be possible to introduce the system to the secretarial staff on a regular basis.
The relative advantages and disadvantages of ATS, the ELECTRIC layout facilities and the magnetic card typewriter have been evaluated and are being considered. It is hoped that the system finally decided upon will allow an author to produce clearly formatted error-free manuals and documents of any complexity of either type or printed (via the FR80) typeface with only a single typing of the initial full manuscript. Proof-reading, and insertion of complex equations could be performed by the author, while corrections, aesthetic layout and the preparation of the document for "printing" could be handled by secretarial staff. Rough drafts of the text should be available by listing on the lineprinter.
The Laboratory has continued to assist engineering users with computational problems. A number of engineers have used the finite element package PAFEC at ACL, with SKC's help, and the graphics display feature of this package which was developed at ACL is in great demand as a means of checking input element data.
In addition JEC has been looking at fluid problems, in particular a free-surface weir flow and a viscous pipe flow in irregular geometry.
Liaison has been maintained with other laboratories especially Harwell, where John Reid's work on automatic triangulation of a region with complicated boundaries and internal interfaces has been usefully applied to an ACL user's problem involving a carbon fibre material.
After a holdup (due mainly to program bugs) which started in the previous quarter and lasted until early February, processing went ahead fairly steadily through the rest of the quarter. IV has now taken over completely the running of the programs and liaison with Dr Thompson on development work, which is now mainly concerned with the software for the planned reprocessing of the raw data. The ACL Applications Committee has accepted the recommendation of the S68 Data Committee that 370/195 time be allocated for the project until the end of 1976 to cover the second processing, and a third 3336 disk pack has been acquired to be used in extending the catalogue used for identifying stars.
PFS is no longer working full time on the project. Apart from administrative work he has implemented some graphics programs for Dr Thompson and for Dr Morgan, also of ROE, who has resumed his studies of the characteristics of the background noise in the S68 data. One of these jobs was used as an opportunity to discover something about the POLYGRAPHICS system; when this was about half completed we found out that POLYGRAPHICS would soon be superseded by a 370 version of SPROGS.
In future the main processing program may write most of its lineprinter output to magnetic tape which can then be brought to ACL, read into the 1906A and the contents printed out at ROE via the link. PFS is investigating this and other possibilities.
Progress on ALPS has been slower than expected, mainly because of poor access to the 1906A. As a result of this more code has actually been written, but too little testing has been done. We expect to have to spend time re-writing code which has undiscovered design faults.
The only large area of coding left is in the area of producing tables. DF has joined us to write part of this. AJHW will be spending most time on this area also (but mainly concerned with the recognition and conversion of table definition statements).
It is hoped to have a user version available by the end of July.
After the demise of the NERC Computer Graphics Subcommittee at the end of last quarter the only NERC duties have involved some maintenance work on the SACM contouring package. Interest in this package seems to be picking up on the 195.
The range of statistics programs in G-EXEC has been expanding rapidly this quarter as John Cubitt and Sandra Strachan have added many univariate and multivariate analysis modules from the Kansas Geological Survey collections. These are to be used shortly in the final stages of a large engineering geology survey. Several more utility programs have been added, dealing with the replacement of data descriptions on files and the building of an index of data descriptions. The latest version of the Executive Controller is now partly working, with a more efficient use of disk access as suggested by VRS. The elapsed time for a typical first-stage job is now 30 seconds to 2 minutes compared with 5 to 12 minutes for the earlier version. Modifications have also been made which reduce the elapsed time of second-stage jobs. The other new features of the controller include automatic copying from tape to disk or vice-versa if a file is requested in one medium and is resident on the other. Instructions appear on the end of the first-stage job stating where the second-stage job has been routed, whether plotting tapes will be produced, and the required procedure for collecting them.
The full version of G-EXEC has now been implemented on the 370/158 at ERCC, by Steve Henley. The PDP 11/45 version development has been halted pending the implementation of a new operating system (RSX-11D) which will allow the use of 32K of core instead of 20K, and allow a wider range of application programs to be used.
External distribution of subsets of the system has continued; we were very pleased to hear at the beginning of March that the Dutch Soil Survey have decided to use G-EXEC as their data management system. Most of the new users of the system are contributing programs; the Dutch survey have offered us their free-text input scheme which appears highly successful, and the Swedish team hope to add a plotting section. The Geology Department of Leicester University has shown interest in adding resource evaluation programs dealing with ore bodies, and Dr O'Leary of that department has offered his services as secretary of a User Group to coordinate new programs and new users outside IGS. It is hoped to have the first meeting of this group in November after a COGEODATA meeting which most of the users are likely to attend.
Work on the G-EXEC interface to FILEMATCH Communication Format has continued, modifications being made in the light of the current definition of the Format. Further testing is to be carried out during Mel Cooper's visit in May.
In February, Dr Neil Burk, National Coordinator of the Canadian Centre for Geoscience Data (Department of Energy, Mines and Resources) visited both the Lab and the IGS offices. Dr Burk was very interested in the "communications format" concept, and in fact it was he who coined the term "FILEMATCH". It was during the time that Dr Burk was at Atlas that the proposals with respect to the development and testing of the FILEMATCH Format were finalised and presented. The proposal and a definition of FILEMATCH has been prepared and will shortly be issued as Note 2 of "Management of Data Bases".
In late March, PGS travelled to Prague where the FILEMATCH concept was presented to Earth Scientists of the Czechoslovakian Office of Geology.The visit was arranged by Dr Jiri Hruska of Geoindustria (Prague), an organisation which is involved in the computer processing, collection and storage of earth science data. Dr Hruska is also a member of the Committee on Storage, Automatic Processing and Retrieval of Geological data (COGEODATA), a committee of the International Union of Geological Sciences (IUGS). PCS also visited Dr Andre Hubaux of Euratom, Ispra, Italy. Dr Hubaux is a member and former Chairman of COGEODATA, and it was to solicit his opinion on the validity and general applicability of the FILEMATCH concept that the trip to Italy was made.
| The S68 Project: Information Note for the Astronomy II Committee of the ASR Board, 29 January | P F Smith |
| S68 Data Committee: List of publications received to date, 29 January | P F Smith |
| Thermodynamic Properties and Self Diffusion of Molten Sodium Chloride, J Chem Soc, Faraday Trans II, 1975, 71, 41 (with K Singer) | J W E Lewis |
| Thermodynamic Properties of Liquid Ionic Salts obtained by Monte Carlo Computation, J Chem Soc, Faraday Trans II, 1975, 71, 308 (with K Singer and L V Woodcock). | J W E Lewis |
| Tree-searching and tree-pruning techniques. Conference on Computer Chess, Balliol College, Oxford, 25 March. | P Kent |
| The Calculation of Valence Shell Ionization Potentials by SCF Method, Molecular Physics, 29, 873 ( 1975). | M F Guest, V R Saunders |
| Computer-Assisted Text Editing at ACL - Future Policy | V R Saunders, M F Chiu |
| G-EXEC - Input/Output Efficiency | V R Saunders |
| XRAY-74 on the ICL 1906A and IBM 370/195 | P A Machin, M Elder |
| Report on visit to ILL Grenoble | J W E Lewis |
| Crystallographic plotting Packages. Use of ORTEP on the IBM 370/195 | M Elder |
| Laser-Plasma Interactions, Meeting held at Colchester on 3 January | J E Crow |
| Crack Propagation. Conference held at Cambridge on 6-8 January | J E Crow |
| Seminar on Laser-Plasma Computations at Culham | J E Crow |
| Seminar on Engineering Graphics held at RL on 29 January | J E Crow, B Stokoe, J W E Lewis |
| Seminar on Direct Methods for sparse Least Squares Problems held at AERE Education and Training Centre | J E Crow |
| Seminar on Boundary Value Problems held on 27 February at AERE Education and Training Centre | J E Crow |
| Seminar on Invariant imbedding for boundary-value problems held at the Oxford Computing Laboratory on 13 March | J E Crow |
| Seminar on Black Holes held at Culham Laboratory on 19 March 1975 | J E Crow |
| Course on NASTRAN and ASKA held at SIA London on 21 March | J E Crow |
| Computational Physics Seminar held at Culham Laboratory on 24 March | J E Crow |
| Visit to Leicester University to see Interactive Graphics System for CAD research held on 26 March 1975 | J E Crow |
| Meeting with Professor Mason and Dr Blundell to discuss microdensitometer use. Held at Sussex university on 7 January 1975 | M Elder |
| Symposium of the Oscillation Method, Groningen, Holland held on 17-21 February 1975 | M Elder, P A Machin |
| Meeting with Mr A Tamburrini to discuss the Italian microdensitometer service held at ACL on 24 February 1975 | P A Machin, M Elder |
| Chemical Society (Faraday Division) meeting on Potential Energy Surfaces, held on 5 March | M F Guest, M F Chiu, W R Rodwell |
| BOON progress Meeting, held at Cranfield on 8 January | B Stokoe, P Kent |
| ALPS Progress Meeting held at ACL on 16 January | B Stokoe, P Kent, A J H Walter |
| Meeting to discuss the Proposed Seminar on Two-dimensional Signal Processing | B Stokoe, P Kent, M B Sherwen |
| ALPS progress meeting held at ACL on 10 March | B Stokoe, P Kent, A J H Walter |
| Meeting to discuss the conference on Two-dimensional Signal Processing held at State House on 14 March | B Stokoe, P Kent, M B Sherwen |
| Conference on Computer Chess, Balliol College, Oxford 24-25 March | P Kent |
| Conference on Solid State Physics held in Manchester on 6-8 January | J W E Lewis |
| Discussions at ILL Grenoble with Drs Lovesey and Copley from 5-12 March | J W E Lewis |
| Liquid Crystal Forum held at the Cavendish Laboratory, Cambridge | J W E Lewis |
| Seminar on Satellite Data Processing, Culham Laboratory, 25 February 1975 | P F Smith, B Stokoe |
| S2/68 Joint Executive Committee Meeting held at ACL on 4/5 March | P F Smith, B Stokoe |
| Geological Society of London Information Group Committee meeting held on 28 January | E M Gill |
| Computer Working Group meeting held at IGS on 31 January | E M Gill |
| Discussions with Dr Burke, 10-14 February | P G Sutterlin, E M Gill |
| Discussions with Dr H Berner (Swedish survey) held on 18-19 March | E M Gill, P G Sutterlin |
| Visit to IGS Edinburgh on 12 March | E M Gill |
| Visit to Czechoslovakia, Italy on 29 March | P G Sutterlin |
