Contact us Heritage collections Image license terms
HOME ACL Associates Technology Literature Applications Society Software revisited
Further reading □ OverviewQ1 1972Q2 1972Q3 1972Q4 1972Q1 1973Q2 1973Q3 1973Q4 1973Q1 1974Q2 1974Q3 1974Q4 1974Q1 1975Q2 1975Q3 1975Q4 1975
ACD C&A INF CCD CISD Archives Contact us Heritage archives Image license terms

Search

   
ACLLiteratureProgress ReportsBasic Software
ACLLiteratureProgress ReportsBasic Software
ACL ACD C&A INF CCD CISD Archives
Further reading

Overview
Q1 1972
Q2 1972
Q3 1972
Q4 1972
Q1 1973
Q2 1973
Q3 1973
Q4 1973
Q1 1974
Q2 1974
Q3 1974
Q4 1974
Q1 1975
Q2 1975
Q3 1975
Q4 1975

Quarterly Progress Report 1 January - 31 March 1974

F R A Hopgood

5 April 1974

1. INTRODUCTION

The Group increased its membership by two during the quarter. Steve Perkins, who has been working for both Basic Software and Resource Management, formally joined the Group. Ray Waters also transferred from Resource Management so that his work could be more closely integrated with that of Cliff Pavelin. Chris Hims: a sandwich student, joined the Group in February for six months.

A number of interview boards were held near the end of March and more are scheduled for April. It is hoped that the Group will be able to reach full complement before the end of September.

To some extent Group structure has changed during the quarter. With the increasing involvement in communications work, a separate section has been formed under Paul Bryant. He still remains in charge of the 1906A operating systems work - but some of the day-to-day running has being taken over by Cliff Pavelin. In particular, co-ordination of the work on Accounting, Budgeting, Scheduling and Performance Measurement is his responsibility.

Present Group structure is:

2. COMMUNICATIONS

2.1 GEC 2050 (MDF)

The following improvements have been made to the 7020 emulator during this quarter:

  1. Teletype output: the introduction of double buffering has increased the speed of this device and it is now possible to overlap the output for the teletype with I/O for other peripherals.
  2. Paper Saving: files output from the 7903 consist of two distinct parts. The first is the file itself ending on an end-of-file terminator. The second consists of the line of asterisks followed by a page throw. For some unknown reason, the latter item is received by itself at frequent intervals when the lineprinter is idle. To avoid wasting paper, this is now trapped by software and the page throw replaced by a line feed.
  3. Diagnostics: the diagnostic dump facility has been enhanced to produce more timing information. This has allowed a more critical assessment of the efficiency of the emulator. The main conclusion is that the 7903 does not attempt to drive the 2050 at full speed and, therefore, further efficiency enhancements would have no effect.
  4. Statistics: the emulator is currently being modified so that the user can request details about the number of lines printed, cards read and transmission errors on the communication line.
2.2 Front End Processor (PEB, FRAH, MDF, DCT, JDT)

The GEC 4080 was finally ordered and should be delivered at the beginning of September. The machine has a minimal configuration consisting of 64 Kbytes of storage, 4 synchronous interfaces for attaching workstations and connecting to the 360/195 and a lineprinter interface. The paper tape equipment on the 2050 will be used as the slow peripherals on the 4080. The lineprinter interface will allow the 2050 lineprinter to be switched between the two machines.

Connection to the 1906A will be by a locally produced interface. This has now been designed and all the critical components are to hand. In particular, the cable drivers and receivers have been obtained from ICL. The original design incorporating packed mode has been abandoned and only the unpacked mode will be implemented. This will mean that loading the system from the 1906A may not be possible. This has been forced on us because the Rutherford Laboratory had insufficient staff to construct the complete interface (the cost of producing it outside would have been more than £1,000). The simpler interface is about a third of the complexity of the complete interface and should be within Rutherford's capabilities.

GEC provides two operating systems, COS and DOS, for the 4080. The core-based operating system, COS, will run in the 4080 configuration ordered. It looks as though GEC will be putting more effort into the disc-based operating system, DOS, and, therefore, the aim is to try to connect the Data Products disc used on ATLAS. This disc is much larger than is needed initially. The aim is to run the disc up and check that it is still functioning correctly. One problem is that ICL refuse to maintain the disc. Assuming the disc still exhibits its earlier reliability, the aim is to design and build an interface to the 4080. It is possible that problems may arise due to the high speed of the disc. Work in this and other hardware areas tends to be held up by lack of manpower.

The main aim of the project is to allow the existing 2050 workstations to access both the 1906A and 360/195. This will allow the batch user at the workstation to enter jobs to both machines. Teletypes connected to the workstation should be able to connect to either MOP on the 1906A or ELECTRIC/MUGWUMP on the 360/195. Subsidiary aims of the project will be to reduce the number of ports required on the 360/195 and to provide access to any networks that may come into existence. The intention is to use the Rutherford Laboratory's workstation software without modification if that is possible. This means that the line protocol used will be HASP interleaving with binary synchronous line protocol.

Some work has been done on the design of the 4080 software. The GEC operating system provides good hardware protection between separate processes and this will influence the form of the system. The work falls into three main areas:

  1. Communication to the workstation and 360/195 using HASP interleaving. The main problem here is the association of input and output with a particular workstation. The 4080 appears to the 360/195 like a large workstation with many lineprinters and card readers. Unfortunately, the standard HASP system does not allow a particular lineprinter and card reader to be grouped together. Output for a job input at a particular card reader may be directed to any lineprinter. Edinburgh University do have modifications to HASP to cure this problem. However, these changes conflict with Rutherford alterations to allow teletypes on the workstation. An alternative solution is to assign specific properties to each lineprinter and alter the job control cards as jobs are input to insist that output goes to a lineprinter with the particular property. Directing control information to a particular console may also be difficult.
  2. Communication to the 1906A: the 4080 must appear like a 7903 to the 1906A except that input and output needs to be interleaved. A number of experiments have been made to check whether the non-interleaving on the 7020 is controlled by the 7903 or GEORGE. If GEORGE stops interleaving then considerable modifications to the 1906A operating system will be required. A detailed study of the code in the 7903 is currently under way. This is hampered by out-of-date and incomplete documentation. ICL's policy of not making the assembler for the 7903 available means that it is impossible to add diagnostic prints to the 7903 program - a frustrating and absurd situation.

    A large amount of experimental work has been performed on driving the 7903. Conceptual multiplexors in GEORGE 4 have thrown up a number of problems. Various facilities were found not to work (for example, a configuration request obtained a garbage reply; an identifier free segment was not necessarily generated when another teletype was attached to an active conceptual - the program would, in some cases, never know about the device). Attempts to drive both the lineprinter and card reader on the 2050 crashed GEORGE on several occasions. ICL now claim to have corrected these faults. However, tests to verify this are still to be performed.

    A further experiment to drive a conceptual multiplexor with a multi-member program showed that this ran exceedingly slowly and always broke GEORGE after about 5 minutes. This has now been abandoned.

    Driving the 7903 online has turned out to be a hit or miss affair - if every thing is just right it works quite well. The difficult part is finding out what is the correct thing to do (the documentation is notable only for what it omits to say - for example, the meaning of the error replies from the DCP in the 7903).

    It is now possible to drive teletypes successfully (without GEORGE breaking!) through both the 7903 online and a conceptual multiplexor. The main result, as far as the 4080 project is concerned, is that it appears as though interleaving is being stopped in the 7903 rather than GEORGE. Further tests should verify this.

  3. Control of peripherals: the central part of the system will be controlling the attachment of a particular workstation peripheral to one or other computer. It must keep a record of assignments and be able to reassign peripherals whenever an output document is completed.

It looks as though a considerable amount of work will be involved in setting up the operating system and making local modifications. The line drivers for communication to the 360/195, 1906A and workstations will need to be written. The aim will be to have a standard buffer length (probably 80 bytes) for all I/O. A longer input record will use several buffers. The standard 4080 peripheral transfer is controlled by a pair of words containing the address of the buffer and a count of the number of characters expected. It looks as though it will be possible to alter the control words on the fly so that input is switched to the next buffer. If this is possible, it will not be necessary to have buffers as large as the maximum input record. Tests have been carried out on the 2050. Generating a new input request as soon as the first buffer was full did not work. All subsequent data was lost as the hardware required synchronisation characters at the start of each transfer. By overwriting the control words (Way Control Block) before the last character in a buffer was received, it was possible to switch to a new buffer. It seems likely that this will also work in the 4080.

The aim in the next quarter will be to connect the 2050 to the 360/195 so that familiarity with the workstation software can be obtained. This will be essential for testing purposes if HASP modifications are made in the 360/195. A 4080 simulator is available on the 360/195. JDT has spent a considerable amount of this quarter getting familiar with this. Some diagnostic routines have been written and tested. Access to the simulator will continue via the 2050.

It is hoped that most of the design problems will have been solved in the next quarter and some coding started.

3. 1906A SYSTEM

3.1 Introduction

Mark 7.7 has been in use for all of the quarter and no major modifications have been made. Mends still arrive from ICL in a steady stream. These tend to be for faults that rarely occur on ACL's 1906A. There are one or two outstanding faults that have occurred more than five times and ICL are unable to solve these. The break rate is still higher than expected and it is mainly made up of one-off oddities. However. exploring previously unused GEORGE facilities (for example. conceptual multiplexors) usually throws up a crop of incidents until mends are received.

A bug has recently appeared in ACL's split residence mends. It occurs in a severe core jam situation of a sort which is avoided at ACL (the bug was discovered at Bracknell). A mend has been produced and seems to work although, as it did not occur here anyway, it is difficult to test.

The 1906A hardware has been reasonably stable over the quarter with not too many field modifications. It is therefore unlikely that the hardware reliability will improve significantly in the future. This is rather disappointing as it is usual to have hardware faults that break the machine most weeks. It is possible that GEORGE enhancements will allow the machine to fail-soft more often but that is all.

The next quarter should see the introduction of Mark 8. A copy of the symbolic code has been obtained and copies of this on microfilm are being produced. The move to Mark 8 is likely to take quite a long time. The number of ACL mends to add to the system is quite extensive. However. it should be a relatively simple matter to move backwards and forwards-between Mark 7 and Mark 8 so that the introduction of Mark 8 is likely to be less traumatic than the introduction of Mark 7.

3.2 Security System (GWR)

Work has continued this quarter in reducing the number of users who need to know :MANAGER's password. A system now exists for running specific jobs under :MANAGER without the user knowing the password. Budget controlling and other similar facilities are executed by this process. Further extension of the system to other high security users has been postponed due to awkward problems that can arise in certain crises situations. Alternative means of maintaining security on these users is being investigated.

3.3 Libraries (SRP)

The program libraries :LIB and :SUBLIB have been rationalised with :LIB containing standard ICL software and :SUBLIB containing software from other sources. The directory, :SUBLIB, has been moved in the filestore structure so that it is now directly under :SYSTEM. This should provide quicker access to-the files.

The macros for updating the libraries are now working and have been documented. Any future updating should now be a routine task. Restoration of DISC 26 after a cartridge failure is now working satisfactorily and it is hoped to restore the EDS DCP files automatically in due course.

3.4 SYSTEM JOURNAL (CJP, SRP, CH)

Both the ALGOL 68 and PLASYD interfaces for accessing the SYSTEM JOURNAL archives are complete. All the analysis programs produced so far use the ALGOL 68 interface. Some of the programs in existence or under development are:

  1. Retrievals and unjams: work is in progress to produce statistics of rates of retrievals and unjams at the same time as the accounts run.
  2. RJE jobs: a program exists to list the jobs run from a particular RJE station over various time intervals. This will be of use when the job finished message is removed from the remote consoles.
  3. Paper consumption: a program exists which produces a weekly list of large users of central printer paper. The results are being supplied to Resource Management Branch to assist in any action needed during the current paper shortage.
  4. Job analysis: the SYSTEM JOURNAL job analysis program has been made available for general use.
  5. Individual job statistics: work is proceeding on a FORTRAN program which will generate a compacted journal with each entry containing statistics about a complete job. This is similar to the statistics produced on ATLAS and should provide suitable data for JEH's existing statistical programs.
3.5 Filestore Analysis (CJP)

A filestore analysis program has been produced. This gives for each user his total size and number of:

  1. All files
  2. On-line files
  3. Directories
  4. Files unaccessed for a year

It also gives various totals. It is planned to obtain further information about distribution of times of access to files. Apart from general interest, this program will provide data for the new filestore accounting system and the files tore summary program which will list key details about the filestore each week (Technical Notice 64). Work on the latter has begun.

3.6 GEORGE Performance (CJP, CH)

An analysis of a week's EDS performance was made (Technical Notice 63) and a program written to analyse all backing store activity over a given week (Technical Notice 76). The macro to drive this program has embedded in it the distribution of units over the controllers thus allowing the run to be adjusted for any configuration changes. The results imply that backing store is not near to being overloaded at present.

Efficiency figures have averaged about 40% on object program and 43% on EXECUTIVE and idling in recent months. This is worse than hoped for - although the number of jobs run (particularly MOP jobs) continues to increase. The lack of any obvious bottleneck in the system implies that it is the nature of the work run which limits processor efficiency. The EXECUTIVE time (as distinct from genuine idle time) is unknown at present. However, experiments with a looping real-time program suggest it could be as much as 30% in prime shift! There is an ICL EXECUTIVE mend to measure this and it is hoped that this will be delivered in the next quarter.

3.7 Accounting and Budgeting (RJW)

Upon transfer of accounting and budgeting functions to the Group during the quarter, work has started on restructuring the current suite of programs starting with the filestore control process. New rules for the. system were defined and published in a User Note. The process has been made more modular by removing it from the previous general accounting program and writing a new program which reads a filestore statistics file produced from a previous directory-reading program. The statistics are interfaced with budgeting information from DICTIONARY and a subset of statistics are written to the user-information data base while the control process is implemented. The new system is substantially complete and final testing in parallel with the old programs is currently in progress.

A new macro for running administrative accounting and budgeting processes (that have to run under :MANAGER) was provided during the quarter. The macro is used from a low-level username. It analyses parameters before using the STARTER macro to finally run a terminal job under :MANAGER to execute the required commands.

Work is in progress to move the directory :ACCOUNTS, which is currently a pseudo-user of :MANAGER down to a lower level in conjunction with other directory cleaning-up and reorganisation.

The accounts end-of-financial-year process is being organised.

4. 1906A SUBSYSTEMS

4.1 GERONIM0 (JDT)

The ACL version of GERONIMO has been released to ICL for distribution from Bracknell. The version in use at ACL has been enhanced by a package for dynamic monitoring of 1906A performance. Changes have been made to avoid generating the same screenful of information twice for two terminals.

4.2 XPCH and XPCK Consolidators (JDT)

Progress on understanding how XPCH and XPCK work has been quite slow due to the complexity of the code. Some minor alterations have been made to XPCH to improve its efficiency. A testing version is now in use. This work and that on XFEH has a need for a DISC EDITOR similar to the ICL program XMED but with a number of additional facilities. Work is in progress on this. As both the FORTRAN and PLASYD compilers will accept source from direct access files, such an editor could be of more general interest. It is intended to release it to internal users when it is available.

4.3 Leeds Multi-Access Benchmark/MOP Simulator (GWR)

Due to increased interest in the system, a number of errors in the system under Mark 7 have been investigated and cured. Leeds University are planning to run the benchmark on their new configuration in order to assess the gains made due to Mark 7 and their extra hardware.

4.4 TASK System (GWR)

Despite being available since December, the TASK Manual has only . just been issued to external users and this has resulted in a jump of 50% in the systems usage in the last two weeks of the quarter when TASK ran over 50% of the background jobs in the 1906A.

Enhancements, which are documented in :NEWS.TASK, have been of a fairly minor nature but have proved numerous and time-consuming. The most major change has been the introduction of the ACL versions of the XFIH and the optimising XFEH FORTRAN compilers along with the XPCH consolidator. These are currently accessed by using the parameter TEST but after a period of testing, they will be made the standard compilers, with any newer versions being made" available via the TEST facility.

TASK will now supply the correct default Program Description segment for an EBCDIC FORTRAN program so that card decks from the 360/195 should run on the 1906A with the minimum of changes. Users should note that the default CR input channel (5) and LP output channel (6) are the same on both systems.

New parameters CCORE and CORE were introduced to cater for very large overlay programs which required large amounts of core for consolidation and for binary programs which wished to specify with what size core they wished to be loaded. Modifications to cater for the consolidation of very large binaries were also needed and a new version to overcome errors in this area has been introduced.

The most involved changes were to reduce the amount of monitor file output from TASK. Due to limitations of GEORGE in this area, only a small reduction was achieved but the effort expended was considerable.

The JANE (ex MVSL) system has also been interfaced onto TASK and now awaits a compiler. GROATS has also been added. The inclusion of ALGOL 68 awaits inspiration and time as does analysis of the TASK statistics and evaluation of exofiles.

Effort is currently being directed towards getting the NUTS system off the ground.

4.5 NUTS (GWR)

Discussion about the New UTilities System revealed that the proposed parameter formats were inadequate for the system envisaged. Various other ideas were investigated and current schemes are based on the concept of the GEORGE COPY command where the first parameter is the source and the second is the sink (or destination). The proposed formats for the new system are about to be presented in another paper.

Implementation of the NUTS kernel has started, but is proving a long and tedious job due to the extreme generality of the system and the large number of features that the kernel must contain to be of any use to a utility writer. It is unlikely that anything will be available to users for some considerable time since the system is now so far removed from the ideas of TASK that all its code has had to be written from scratch.

4.6 FORTRAN Compiler Development (DCT)

New versions of the optimising and standard FORTRAN compilers (#XFEH and #XFIH respectively) have been produced and are available to users of TASK. These new compilers provide a number of enhancements, the main ones of which are:

  1. *INCLUDE facility.
  2. input from disc (exo)files, rather than GEORGE filestore (card reader) files : the compilers also provide a line editing facility in conjunction with this.
  3. symbol table listing : this is at present a little crude in that it is printed in an apparently arbitrary order (the order of the variables in the symbol table), but it is intended in a subsequent version to print this in alphabetical order, and also to provide a listing of each COMMON area.

Documentation on these enhancements is available.

4.7 NULLIB (RET)

The only modification this quarter to the library update program has been the changing of the largest part of NULLIB to sparse (as suggested by Jed Brown). This allows the program to expand as required and allows it to fit in with normal 1906A programs. This system is now ready for general issue by Resource Management Branch.

5. GRAPHICS SYSTEMS

5.1 SPROGS VERSION II (RET, ACD, AHF)

Version II of the SPROGS system was finally released in February, rather later than had been hoped. An internally produced version of the new manual was issued to current users. A bound version is at the printers and should be back after Easter. This version of the manual will be made generally available.

The additional features specified in the last Quarterly Progress Report were completed in time for release. In particular the IFS routine was added which allows a FORTRAN-like IF facility from within SPROGS files. The STSQLS routines have been added to allow the sequence list to be initiated without adding a file to the list. Previously, the only method of starting a sequence list was by calling DRAW which automatically adds a file to the sequence list. Dotted lines are now available. A routine for retrieving single characters rather than complete fonts from the system library has also been included.

With the introduction of the new spooling system and the reversion to user owned tapes for microfilm output, some problems arose in finding the name of user-owned magnetic tapes. It was thought at first that it would be necessary to rewind the tape and read the title. However, it was found that an open-mode PERI can be performed after an ONLINE or GETONLINE command provided the GIVE qualifier was not present.

With DCT's improved version of XFEH becoming available during the quarter, the SPROGS generation system has been altered to make use of *INCLUDE for COMMON blocks. This has shortened compilation time by 15-20%.

5.2 SPROGS Experimental System (RET, AHF, ACD)

Following usual practice, the issued version has been frozen and the source saved on magnetic tape. Meanwhile, new facilities are added to a current experimental system (version III in embryo). The extra features included so far are:

  1. Tektronix : during the quarter PEB managed to get GEORGE to produce upper and lower case output from GEORGE files. It is therefore possible to access a Tektronix storage tube as an output device from the 1906A. Unfortunately it is only possible to display one frame at a time at the moment (the second frame in a file currently just overwrites the first). A macro (:SPROGS.TKFILES) has been provided which will split a file containing a number of frames into several files each containing one frame. These may then be listed and viewed on the Tektronix one at a time.

    Code has been written to include the Tektronix as one of the SPROGS output devices. An ALLCHAR file is written to TPO which can subsequently be listed at the terminal. The file starts with 4 dummy records to allow the screen to settle, and each frame ends with a command to reposition the display back to the top left-hand corner. Using the macro described above, it is possible to split the file up so that the individual frames can be viewed. As a test the 1906A performance program written by CJP has been modified to produce graphs on the Tektronix.

  2. Dummy Devices: by replacing a number of dummy routines, it is now possible for a user to insert a new output device without recompiling the system (the facility is akin to the provision of dummy display routines.
  3. Irregular Regions: it is now possible to specify the shape of a region by providing a NULL file containing the coordinates of the boundary. The region must be a simple closed figure - but can be concave. Lines intersecting such a region are correctly segmented into inside and outside parts.
  4. Graph Plotting Package: ACD has produced a graph plotting package which can loosely be divided into two parts. The first is a set of routines which output a graph with no user control of variables. The second is a set which output a graph with many user-controlled options (for example; size, markings etc).
5.3 3-D SPROGS (RET, ACD)

It is intended to provide a simple 3-D capability by specifying a 3-D region. Such a region would take a 3-D object and project it onto a 2-D region from a specified view point. A new set of 3-D display routines will be provided to allow a Z-coordinate to be specified. At the moment, the conversion routines have been written and tested but the mechanism has not been included in the SPROGS region system. It is possible to produce 3-D pictures by including the conversion package at the top level of the chain. Allowance has been made for the 2-D plane to be behind the view point, and for the view point to enter the object being projected.

5.4 SPROGS Film (AJP)

Work has continued this quarter on the SPROGS film. The section dealing with the use of the SPROGS filestore and sequence list is now complete. A number of modifications in speed were required after seeing the initial version. The fairing algorithms have been improved and all movement is now smooth. Some movements have had to be slowed as strobing occurred on the vertical lines.

The machine time taken to generate the film has been reduced since the introduction of the SAVE feature for SD4020 instructions. Consecutive frames of identical output are now produced by storing the relevant SD4020 instructions in a file and outputting from here rather than recalculating each frame.

5.5 LOGO Sequence (AJP)

Some effort has gone into seeing if the new Atlas Laboratory logo can be used as the basis of a title or end sequence for laboratory-generated films. A number of ideas have been tried and a final sequence will probably emerge from these.

5.6 POLYMER Film (RET, ACD)

Following a request from Professor Edwards, SPROGS has been used to produce a film showing polymer chains deforming in three dimensions. A draft version of the film lasting about two minutes was produced in three days and took about one hour of 1906A time. Further work on the film is continuing. An introductory sequence is being produced which shows mid-point and end-point movement of the polymer chain before giving a close-up of a short 20-point chain depicting both types of movement. The film itself will probably be extended to show the steady state position. It may be that sequences will be added which start from different initial positions. For example, starting with the chain in a knot is apparently of interest.

6. PDP15

6.1 Hardware Modifications (JRG)

The only hardware modification this quarter has been the addition of a Tektronix 4010 storage tube to the PDP15 via the LT19 multi-teletype controller. Neil Parker has added switches so that:

  1. the Tektronix and a teletype can share a line from the LT19. Each line costs sufficient to make this a worthwhile saving.
  2. the Tektronix can be driven from either the PDP15 or a 600 baud line from the 1906A.

Some problems do arise. Switching the Tektronix to the 1906A without connecting the teletype to the PDP15 can cause interrupts in the PDP15. The standard multi-teletype handler in DOS can produce text for the Tektronix. However, some basic routines must be written to draw lines.

6.2 Hardware Faults (JRG, WDS)

Numerous non-complex faults occurred this quarter (tape deck errors, faulty switches, noisy fans, failed fuses and bulbs). Of the faults brought forward from the last quarter, the DMAC interface appears to be fixed. Also the randomly occurring system errors have disappeared without further intervention.

The main outstanding problem is the spark pen/keyboard problem. DEC had hoped to obtain from the UK distributors of the spark-pen, a finger switch to inhibit sparks when the switch was released. The aim was to prolong the extremely short pen-life and to prevent interference with the keyboard when the pen was laid down. The finger switch that arrived had an entirely different effect which was of no use. DEC are still negotiating for another switch! The pen being used at the moment has lasted considerably longer than its predecessor. This could be due to WDS rubbing graphite on the tip.

6.3 Operator Training (JRG, WDS)

As part of the hand-over of some PDP15 activities to Operations Branch, one operator from each shift is gaining experience of the PDP15 - each for a 6 week period. It is intended that some software can be run by the operators for users and that the operators should be capable of sorting out simple user problems. So far, two operators (Alan Hancock and Gordon Dawson) have finished their 6 weeks. By the beginning of May, Trevor Amos should also have finished the training and each shift will have a PDP15 expert.

Each operator has a programming project. AJH has been writing a program to exercise the whole machine and its peripherals. GVD is producing a system to read points from the DMAC and display them as well as adding the points to a disc file. It is hoped that DMAC users will find this helpful. The disc files produced can be sent to the 1906A via the link. Some simple editing will be included.

6.4 RSX PLUS III Operating System (JRG)

A prelease, version of RSX PLUS III was installed on 16 January. This version includes the facilities of RSX PLUS and DOS as well as some additional facilities. Some serious bugs were found in the DOS part of the system. This means that DOS users are still using the old version. One of the problems with the system is that the FORTRAN library assumes that the hardware floating point is available. As the old library will not work under RSX, it makes the system virtually unusable at present.

Some work has gone into evaluating RSX. A number of design problems have emerged and are described in SPROGS Paper No 43. One major problem is that the list of installed tasks is kept in core .and is, therefore, insecure. However, the task for saving the current core image can only be run under extremely restrictive conditions. It is therefore difficult to secure the system. Another major problem is that tasks are not attached, in any sense, to the console that started them and no console has a complete set of commands. DEC is aware of the problems and some action may be taken. In order to get better understanding of the system, a course on RSX is being run at the laboratory by DEC in May.

6.5 PDP15 - 1906A Link (JRG)

Most of the hardware problems now seem to be cured. The PDP15 to 1906A character transfer system has been tested with several DOS system programs and has caused a number of PDP15 system errors. Transfers are successful using PIP or a user-written FORTRAN program. The current speed is about 80-130 lines per 1906A job-time second and 30-40 lines per real time second. Further software will be buffered differently and should show an improvement.

6.6 PIGS (WDS)

The assembler for the PIGS Menu Definition Language (MENDEL) is nearing completion. A usable version of the first system should be available next quarter. Work has begun on the first PIGS manual.

6.7 Synthesizer (AHF)

The interface unit for the synthesizer has been installed in the PDP15. Initially, the interface unit did not pass all the signals correctly along the I/O bus. After checking all the wiring and adding additional earth wires, the fault was finally found in the DEC-supplied I/O bus cable itself. The cable was modified and signals then passed along the bus correctly.

After one or two other minor wiring changes, the first computer-controlled notes were produced on the last day of the quarter. There should be music while you work in the computer area next quarter!

7. MISCELLANEOUS

7.1 COLAB (RET, DCT)

COLAB was used successfully in an SRC training course in February. As the course has moved from Cosener's House, it is no longer possible for the participants to have a dummy run on the machine the previous evening. Instead a hand simulation was tried and this proved quite successful with the aid of a FACIT. In spite of requests not to do so, a number of players managed to BREAK IN and disrupt the system (this problem will be cured if a conceptual multiplexor version is introduced). The main alterations still required are in the order and content of the final performance analysis. The time allowed for the game proved adequate and similar arrangements have been made for the next Course in April.

7.2 Audio Equipment (AHF)

The audio equipment for the new Colloquium Room has been purchased this quarter so that it could come out of 1973-74 budget. The equipment consisting of a power amplifier, pre-amplifier, record deck and two loudspeakers will be installed as soon as the building is ready. Modifications are being made to the new 16 mm projector so that soundtracks can be played through this system. The equipment has been tested and leads made for connecting to the laboratory's other audio equipment. Some experimentation is still needed before a decision can be made on the exact placing of the speakers.

8. VISITS AND COURSES

24 January
BCS Symposium on Job Control Languages London, P E Bryant
4-8 February
SRC Management Course III, High Wycombe, C J Pavelin
January-March
Course in Electronic Music Composition 6 sessions, 1 evening a week, Bracknell, A H Francis
3-15 March
Advanced Course on Compiler Construction, Munich Technical University, Germany, D C Toll
11-15 March
SRC Management. Course II, HTS Management Centre, Farnham Common, A H Francis
11-15 March
GEC 4080 Systems Appreciation Course GEC Training Centre, Borehamwood, P E Bryant, M D Fowler, J D Thewlis
27-29 March
Special Course on Minicomputer Interfacing Polytechnic of Central London, M D Fowler
8 February
Visit by SRC Management Course III for COLAB exercise (R E Thomas and D C Toll)
12 February
Visit to Daresbury to investigate facilities available at the site, F R A Hopgood, D G House
25 March
Visit to Southampton University, Dept of Sound and Vibration Research on behalf of SRC/AME Panel, R E Thomas, J R Gallop
26 March
Visit by members of Queen Mary College, Department of Computer Science to discuss 1900 compilers with D C Toll and J D Thewlis

9. MEETINGS ATTENDED

3 January
ICLCUA GEORGE 3 User Group General Meeting, F R A Hopgood
21 January
SRC CAD Working Party, London, P E Bryant
1 and 13 February
ACTP Project Meetings, Stevenage, P E Bryant
18 February
SRC Networks Panel's visit to Daresbury, P E Bryant
1 March
SRC/AME Panel meeting, Leicester, R E Thomas
7 March
Communications Co-ordination Group ERCC Edinburgh, P E Bryant
8 March
NPL Computer Performance Meeting. London, P E Bryant

10. GROUP TALKS

2 January
ATLAS Supervisor, P E Bryant
16 January
Advanced Course on Computer Systems Architecture, G W Robinson
30 January
History of Computing, film and general discussion on possible Daresbury move
13 February
Structured Programming - or Life in 1984, F R A Hopgood
27 February
Virtual Memory on IBM Machines, A J H Walter
27 March
The Architecture of the GEC 4080, J D Thewlis

11. PAPERS

1906A INTERNAL USER NOTICES
70 Specification of UTILITY macro, S R Perkins
71 Economies in use of lineprinter paper, P E Bryant
72 GEORGE Bug, P E Bryant
73 :LIB and :SUBLIB Directories, S R Perkins
74 Banner Output Macro ZLETTERS, P E Bryant
75 New FORTRAN compiler #XFEH MkIE Version I, D C Toll
76 GEORGE 4 Mk 7.7 Version F,P E Bryant
77 Information Retrieval, GLM Berrey-Roghe and A M Walter
78 Alterations to the FORTRAN run-time I/O Package, A J H Walter
79 New FORTRAN Compiler #XFIH'Mk 1E Version 2, D C Toll
80 Program Source in Direct Access Files, J D Thewlis
81 Changes on introduction of Mk 7.7 Version J, P E Bryant
1906A TECHNICAL NOTICES
61 Filestore Characteristics (3), F R A Hopgood
62 GEORGE 3 User Group Meeting 3.1.74, F R A Hopgood
63 Use of EDS 30 and EDS 60, C J Pavelin
64 Filestore Monitoring, C J Pavelin
65 Accounting and the filestore, C J Pavelin
66 User/Job Statistics, C J Pavelin
67 Report of a meeting on certain sites special requirements for GEORGE Mk 8, P E Bryant
68 Report on GEORGE 3 User Group Operations Management Activity Group meeting held on 21 January 1974, R J Waters
69 Changes to Filestore Control System, C J Pavelin
70 Statistical Analyses, C J Pavelin
71 Dumping. and Archiving - An interesting idea, P E Bryant
72 Use of the MAGIN Macro, C J Pavelin
73 Library Maintenance, S R Perkins
74 Tape Library Growth, P E Bryant
75 Security, P E Bryant
76 EDS and drum analysis program, C Hims
77 SYSTEM JOURNAL Interface, C J Pavelin
78 GRAPHIC, NORMAL AND ALLCHAR files and device independence, P E Bryant
79 Operator Macro JOBLIST, S R Perkins
FRONT END PROCESSOR PAPERS
8 Notes on a meeting to discuss GEC 4080 Project, F R A Hopgood
9 Connection to the 360/195, P E Bryant
10 Record of a meeting with GEC Computers Ltd held at ACL on 16 January 1974. P E Bryant
11 Notes of a meeting with members of Rutherford Laboratory on 5 February 1974, F R A Hopgood
12 Discussion paper, P E Bryant
13 A simplified view of the ELECTRIC multiaccess system and the GEC 4080 simulator, J D Thewlis
14 Front End Processor project at ACL, P E Bryant
15 GEC 4080 Project - Progress Report, P E Bryant
16 Cyclic Redundancy Checking in the 2050 HASP Emulator, F R A Hopgood
17 Report of a meeting on communications at ERCC, Edinburgh, on 7 March 1974, P E Bryant
18 ACL's software scheme for the GEC 4080, P E Bryant
19 Attachment of the Data Products Disc to the GEC 4080 via the existing interface to Sigma 2 modified, P E Bryant
SPROGS NOTE
18 Notes of a meeting held with DEC to discuss problems with RSX, R E Thomas
SPROGS PAPERS
41 A proposed graph plotting suite for SPROGS, A C Dennis
42 Possible DMAC Monitoring System, J R Gallop
43 Changes and additions to RSX PLUS III, J R Gallop
44 New Output Devices in SPROGS, R E Thomas
SPROGS TECHNICAL PAPERS
16 (a) Currently available documentation on RSX PLUS III (b) Titles and page numbers of RSX PLUS III Specifications, J R Gallop
17 SPROGS Generation System Version II (update), R E Thomas
18 Advance Repeat, R E Thomas
19 PDPI5 Graphics Software FOG and VTPRIM, J R Gallop
20 Irregular Regions, R E Thomas
OPERATING THE PDP15
(Notes for Operators), J R Gallop

12. LECTURES

11-week course on Translator Writing Systems, 4th year undergraduates, Brunel University, F R A Hopgood

10-week course on Compiling Techniques, M Tech students, Brunel University, F R A Hopgood

⇑ Top of page
© Chilton Computing and UKRI Science and Technology Facilities Council webmaster@chilton-computing.org.uk
Our thanks to UKRI Science and Technology Facilities Council for hosting this site