The size of the group has increased considerably during this last quarter. It now consists of:-
In addition, David Sedar (DAS), a sandwich student, has been with the group this quarter.
The graphics hardware has still not been accepted. At the end of the last quarter most of the graphics control hardware had been replaced by DEC. This should have cured our display problems but in fact produced a number of minor faults which took a considerable part of October before they were cured. The new hardware improved the accuracy on the display but not to a level where it could be accepted. There is still a considerable amount of overshoot on long vectors which are nearly vertical or horizontal. In addition the positioning of short vectors is sufficiently inaccurate that any small object is just about unrecognisable. DEC think that the trouble is due to lack of accuracy in the deflection yoke of the Kratos tube. This was replaced near the end of December. The display has not yet been set up correctly since then so that it is impossible to say at the moment whether or not this will cure the faults. Our impression is, however, that although some improvement is exhibited it will not be within the manufacturer's specifications.
The basic PDP15 hardware has performed reasonably well this quarter although a number of minor faults have occurred. The most serious fault was a core store one which took a number of days to sort out completely.
The service on the SD4020 has been rather disrupted this quarter. At the beginning of the quarter considerable time was lost replacing charactron tubes until finally a new tube with sufficient intensity was encountered. Once installed this had quite a dramatic effect on the output. Instead of requiring an aperture setting of F3.4 to generate reasonable output, the cameras had to be closed down to F16. For the first time for about two years we now have more than enough light intensity on the tube.
Once the new tube was sorted out, the new combined 16mm/35mm Vought camera was installed. Installation of this did not go quite as smoothly as it might have done. The image size on the 16mm film was not large enough to get a full cine image. In addition it was found to be impossible to abut 35mm frames. Consequently new lenses are required for both cameras and these should be arriving in the next quarter. The registration on the new camera is much better than that on the old ones and there should now be no problem with camera jitter.
The SD4020 has again had another bad quarter with a stream of minor errors making it almost impossible to get any production work done at all. Some of the faults were:-
Installation of two minor modifications has taken place during this quarter. The first is the variable intensity option which allows characters to be output at a range of intensity levels. The second is a modification which allows a carriage return without line feed in lineprinter simulation mode.
The major effort this quarter has been getting CAMP working on the PDP15. The major bug which was causing a lot of trouble last quarter has been discovered and removed. The system is now working on the PDP15 apart from a few minor bugs. The system has been modified in a number of ways to make it more amenable to running on the PDP15. The SAVE facility which allows static pictorial information to be saved and output subsequently requires large amounts of storage. On Atlas this is available as core storage whereas on the PDP15 it has been necessary to move this to the disc. When SAVE is called a file is opened on the disc and the data corresponding to the saved instructions is written to this file. When the REPEAT instruction is called the data is read back down from this disc file. A subsequent call to SAVE overwrites the existing file.
On the ATLAS version, CAMP instructions are provided on a card input stream. The simple alternative on the PDP15 is the teletype. However, there is a need to store the current program for subsequent modification. In addition it would be very tedious to re-input the fixed CAMP commands defining the character set each time from the teletype. Input is now allowed to the CAMP system from a disc or DEC tape as well as the teletype. The user has the ability to switch from teletype input to a file named by a CAMP command. Although some problems arose in getting the file name accepted, this has now been overcome. The standard mode of operation will now be with input of the constant commands from one file and the variable data from another.
The output from the system is a magnetic tape which can either be processed on the SD4020 or viewed on the display using JRG's program. At the moment some problems exist on the output side. SD4020 tapes, although correctly simulated on the PDP15, fail on the SD4020.
Most of the GROATS package has been transferred to the 1906A, and, although it has not as yet been exhaustively tested, it works. A certain amount of re-programming was necessary; in particular the ABL routines have had to be re-written in PLAN. The Algol, however, ran on the 1906A almost without alteration.
At present the Algol routines have to be recompiled every time they are used, but work is at present in progress in converting the entire package into a set of independent procedures suitable for creating a library which can be searched by the consolidator. This involves removing the global arrays (but not the global simple variables). This will be tested on the 1906A as soon as it comes back into service.
All of the basic GROATS package, and the character and font definition procedures from the library, have so far been implemented, with the notable exception of the facilities for saving picture or plotting orders. On Atlas, there are two saving systems available, one using an integer array to save the plotting orders, and the other using a real array. On the 1906A only one of these, the real array version, will be implemented.
The GROATS KDF9 and GROATS ELLIOTT versions on Atlas will not be available on the 1906A as the use of these systems has been small. Also, the procedures PRINTTX, PRINT INTEGER will be omitted on the 1906A as the standard 1900 Algol routines WRITE TEXT and PRINT are available.
To the user, the package on the 1906A will appear very much as on Atlas: the global simple variables will be present without declaration, as a: special version of the 1900 Algol compiler will be used (at present XALT, but this may later be changed to XALV); however, direct use of the global arrays will be impossible as they will not exist in their present form.
It is hoped that the system will be available for users by sometime in March: detailed documentation on the changes will be issued at the appropriate time.
In November we received a copy of the new PDP15 disc operating system called DOS. The previous system was mainly an update of the DEC tape monitor. Unfortunately the new DOS is much larger than the old operating system. Without making any changes it needed all our available space on the disc apart from 15 blocks. The new operating system has a greatly improved FORTRAN compiler and an editing system designed to work with the VT15 display. It is also more versatile in the way it can dump core images and handle user files.
The system was pruned to allow sufficient space for user files by deleting rarely used system programs: (TKB, 89TRAN etc) deleting the Bank mode user; deleting programs in the Peripheral Userlib (PER) and transferring source to .LIBR BIN. This allowed us 660 (octal) free blocks. A number of minor inconsistencies were found. For example the name assigned to blank COMMON had changed from .XX to XXMON. Also overlaid programs had to be recompiled due to differences in the calls to library routines. This does mean that we have virtually to recompile everything before using DOS. We are currently continuing to use the old monitor system and will make a concerted effort to move to the new DOS when the extra disc surface arrives.
The display tablet (VW01) has been attached to the PDP15. The handler VWA. and a number of useful routines have been written for it. These are defined in Sprogs Paper No 10. It is hoped to write a routine TRACK equivalent to the light pen tracking routine so that tracking can be done using either the tablet or the light pen. This should improve the speed of tracking considerably.
A number of errors in the basic DEC software have been located and corrected. The ICLOSE command in the editor now works correctly. In FORTRAN, expressions containing division now always produce the correct sign!
An attempt was made to link the PDP15 and 1906A using the BSI interface just before the 1906A was withdrawn from service for the paging modifications. This was unsuccessful due to programming errors. A fresh attempt will be made as soon as the 1906A starts up again.
The listing program on Atlas which prints out PDP15 files is in the process of being rewritten to run on the 1906A.
This has again been modified and we have further changes in mind. The VT15 contains certain useful display orders which are not used at all by the software; for instance, if each coordinate (DX and DY) of a beam movement is less than 32, the line can be specified in one word only compared with the usual two. Making the software produce these shortened orders, has produced a worthwhile saving in picture storage space, when the picture consists of many short straight lines - for instance curves or DMAC input.
We intend to modify the way the software adds display items (lines, text strings, calls to other sub-pictures) to an existing display file. At present the display is switched off by program, the item is added and the display is switched on again. The result is a distracting flickering effect when a picture is being continually updated. Hopefully we can make the picture steady while most kinds of updating are taking place.
A film showing ways of constructing a cardioid is being produced. If this is successful it is hoped that a commercial outlet for a number of films on similar topics may be obtained.
The production of a sequence for use in a film describing Aerial Synthesis and as a film loop for an exhibition next year was successfully completed on Atlas. Subsequent runs on the SD4020 proved less satisfactory and good copies are still awaited.
It is hoped to modify this sequence in order to obtain a multi colour print for the exhibition film loop over the next few weeks.
The FORTRAN package on Atlas is being converted by Nick Modi to run on the 1906A. This entails rewriting most of the machine code routines. However it has been decided to leave out the Extended Benson-Lehner Model J System. This package was originally written to help users of the B-L J plotter to convert over to the SD4020 without them having to change their programs. It should now be possible to phase this out.
VT15. display programs have been written to aid in the reconstruction of digitised photographs. Inconsistent results on the SD4020, however, have delayed the production of high quality prints. The arrival of the polaroid camera will be of value in assisting in the evaluation of the range of significant intensity levels from the display.
We are currently looking into the possibility of purchasing a new microfilm recorder. Some tapes have been prepared for testing the performance of a range of machines. Results from these tests are expected in the next quarter. Already some results are to hand which indicate line quality and picture linearity of a number of machines.
A revised draft of the new system has been written and will be discussed and expanded during the next quarter. It is hoped that programming should also start in this quarter assuming that a reasonable service is being provided on the 1906A to make it worthwhile. There seems little point in starting until the 1906A is available during the working day from MOP consoles.
It is hoped in the next quarter that AHF will be able to complete the PDP15 versions of CAMP and CAMPER. In addition the implementation on the 1906A should be started. GROATS should be implemented on the 1906A as well. We should then be in a good position to estimate the relative efficiencies of the systems on Atlas and the 1906A.
The link to the 1906A from the PDP15 should be made this quarter. Also work on SPROGS implementation should start.
F R A Hopgood attended the UAIDE conference in Los Angeles
