The major happening in the past quarter was the arrival (at last) of the VT15 display. This has, of course, meant that the group could start getting to grips with the graphics side of the work on the PDP15. Consequently a number of pieces of work have been put aside while we concentrate on getting familiar with the VT15 display and how we should use it. The people currently involved in the work are:
A student from Brunel University, David Ralphs, should join the group at the start of the next quarter and will be here until the end of August.
D-MAC Package (GAE) The manual for the D-MAC package is now written and being typed. It should be at the printers some time in April. As well as giving information about the software provided, it does give a reasonably comprehensive definition of how to use the D-MAC and will be available to users in general even if their application does not require our software.
The use of the D-MAC has been increasing and, already, the Operations Group have been able to provide a useful digitising service for users requiring reasonably large volumes of output.
CAMP and CAMPER (FRAH) Work on these two simple packages for production of 2 and 3-dimensional films has been suspended for the time being. It is hoped that David Ralphs will be able to finish off this work when he arrives. Both packages are currently working on Atlas and the 360/75. There is, however, a number of differences in the two packages due to incompatibilities in the Fortran dialects and also differences in the character sets.
A version for the PDP15 has been compiled but is impossible to run due to insufficient store being available. The PDP15 version will have to be split into a number of overlays which requires a certain amount of restructuring of the main routine. It will also be necessary to redefine a number of parameters so that storage for picture files can be changed without having to make major changes throughout the package.
The aim here will be to get the PDP15 version working and then make the other versions compatible with this. Otherwise we shall have a major updating problem. We nope to generate all the diagrams for the manual using the package itself. This is particularly useful in the case of CAMPER, the 3-dimensional version.
Films (FRAH, JRG) After five months of trying, we finally managed to generate a copy of the Syntactic Dominoes film using the SC4020. The tapes defining the film had to be restructured to use the minimum number of facilities of the SC4020. Work is progressing on the sequel to this film (JRG). The two films between them describe the basic methods of syntactic analysis used in a compiler and also describe the top down fast back method in detail. FRAH has used the first film in his Compiling Techniques course where it was appreciated by the students.
FRAH has been investigating the possibilities of producing a set of three films on various Hash Table techniques. A number of different methods of presenting the techniques have been tried without any being completely successful. Draft copies of these have been shown to a number of computer scientists who have expressed interest in the work.
Another three months has passed with virtually no progress on the SC4020 faults that have plagued us since August.
The Advance repeat (AR) option is still not working and little has been done by Datagraphix to get it working. There is still the basic problem of deciding whether the fault is in the tape deck or the SC4020. Our evidence seems to suggest that the SC4020 is incapable of recovering from an attempt to reread a record from tape. We have been promised an American engineer by Datagraphix to work on this problem but he appears to be more mythical than real.
The current situation is not improved by us having an SC4020 engineer who has responsibilities at other establishments. It makes it difficult for him to concentrate on any fault as he is continuously being called away. We are currently attempting to negotiate a new maintenance agreement with Datagraphix.
The current state of the SC4020 is not good. Due to lack of effort in ensuring that output quality is maintained most of the faults experienced in the past are reappearing. We still have random forms flash. The camera has got significant jitter on it once more. The SC4020 is doing advance repeat instructions (re-read ?) when none exist.
The intensity of vectors on the tube has gradually been dropping to an almost invisible level. Some recent adjustments have, however, cured this certainly for the moment. Unfortunately we have been modifying software in attempts to compensate for lack of illumination. We now have to reprogram to adjust once more to current conditions. This whole problem of quality control is becoming very tedious. Until we can get some confidence that the output on successive weeks will be comparable, it is almost impossible to make any progress.
Basic Systems work (GAE) The interface between the D-MAC and PDP 15 has been completed and appears to work reasonably well. The output speed of the D-MAC is sufficiently fast that care has to be taken that input is not lost if the user is attempting a reasonable amount of computation between reading the D-MAC input.
A new monitor system has had to be generated to include the handlers for both the D-MAC and the VT15. We should have received a handler for the VT15 keyboard from DEC but this has still not arrived. Consequently we have been simulating the keyboard by using the D-MAC keyboard. GAE is currently generating a handler for the keyboard on the assumption that the one from DEC will not appear. A few minor errors in the VT15 handler have been found but that is all.
VTPRIM The Fortran display package produced by DEC has now been in use for about one month and already we have found a number of bugs and omissions which make us seriously consider dropping the use of DEC graphic software completely. However with the small amount of effort currently available, this would considerably disrupt our current program of work and we are persevering with it. Some of the bugs are so major that the relevant routine has obviously never been tested at all. The omissions tend to be more serious. For example the display package allows you to receive interrupts from the pushbuttons but does not allow you to turn them off once the interrupt has been handled. When we contacted Maynard, their suggestion was that we inserted a PAUSE in the Fortran program and turned the pushbutton off while the program was stopped!
The display processor works in Page Mode and therefore all display jump orders refer to the current page. Unfortunately DEC software makes no attempt to stop COMMON areas crossing Page Boundaries. Consequently a program can get loaded so that the display file is incorrect.
SD 4020 Fortran Package (PMN) A mini-version of this package is now available for the PDP15 which can readily be extended as the need arises. Plotting points is twice as fast as Atlas, but lines take three times as long.
DOOM (DMAC Output onto Microfilm) (PMN) A program which takes input from the D-MAC pen follower and outputs this to magnetic tape for the SD 4020 is working. It is possible to save complex backgrounds and have these automatically re-output after a frame advance command to the SD 4020. This means that only moving objects need be re-digitized for each new frame. Certain gaps are appearing in the SD 4020 output which have yet to be explained.
Aiding this is the ability to output the digitized pictures on the VT15 display, although problems which are believed to be due to hardware have hindered this.
An extension to be included in the system is the facility to store specifications of several objects, together with a digitisation of the path they are to follow over many frames.
It is hoped that it will thereby be possible for someone to be able to produce a film sequence purely from movements of the pen on the D-MAC table with its position defining the various functions. The division of the D-MAC Digitising Area into various functions is:
|
SAVE |
ADVANCE FILM |
STORE |
|---|---|---|
|
THICKEN |
PICTURE AREA |
NEWLINE |
|
SPATH |
FINISH (RESTART) |
FPATH |
5. Font Definer (FRAH) We still have a number of Atlas users defining SC4020 character sets. This program is designed to generate a new character font using the light pen and VT15. Basically a mesh of points is displayed and the user can indicate which lines should make up the required character. Most of the basic package is written but still requires debugging.
It is hoped to extend the system so that, once the font has been defined, it is possible to generate slides or film titles using the VT15 and output the results to magnetic tape for production on the SC4020.
PMN has been making a number of enquiries concerning possible use of video tape for generating colour films.
A TV system has been demonstrated for producing false colour images ie colours which differ from the original. The original could, in fact, be a purely grey-scale image.
A vidicon camera scans a picture and a black box converts the light received into three intensity levels. The resultant levels are displayed in three colours on a colour TV screen. The intensity levels and colours can be arbitrarily chosen by manual controls. It should be possible to extend the present system to distinguish up to 32 distinct intensity levels.
The advantage lies in the possibility of recording in colour direct onto video tape from a film sequence produced on the monochrome VT15 display.
A photographic pseudo-colour technique developed by the Rand Corporation is also currently being investigated for stills.
We have continued to examine the possibility or producing colour films using the standard technicolor process. Apart from jitter problems which will not get cured until the new camera arrives, it is almost certain that any attempt to produce colour films will require a considerable amount of time spent in film editing and processing. This requires the expertise of a photographic technician. Similarly our efforts at generating film sound tracks indicate that we require the services of somebody with a good understanding of audio equipment . Neither of these people are available in the laboratory and we have neither the time nor the expertise to do the work ourselves. There is a need for somebody with audio-visual expertise and we would like to see such a person on the laboratory staff.
We hope to continue with our current work on the VT15 involving self contained programs until the end of April at least. We should then be in a position to decide on the best way to use the VT15 in the PDP15-1906A system. Design of the 1906A package should start in May. It will be designed as a film generation package and should therefore complement the current SC4020 Fortran and Algol systems which are being implemented on the 1906A by E B Fossey's section. The aim will be to produce a package that can be used either in a batch-mode or alternatively interactively with data being input from the display, D-MAC or keyboard.
We include here a list of papers and talks given by members of the Graphics Section in the past. We hope to keep this updated in future Progress Reports.
In addition to these Graphics talks given externally, there were a number of lectures also given on compiling techniques. These include four complete courses on Compiler Writing Techniques given at the Universities of Brunel and Swansea.
GROATS: a Graphic Output System for Atlas using the SC4020 by F R A Hopgood, Proceedings of the 1969 UAIDE Conference
Chance and Thermal Equilibrium by F R A Hopgood, P J Black and J Ogborn, Proceedings of the 1969 UAIDE Conference
Computer Generated Educational Films by F R A Hopgood and P M Nelson, CG 70 Proceedings
Change and Chance by F R A Hopgood, P J Black and J Ogborn, Proceedings of Computers in Undergraduate Science Education Conference, Chicago 1970
Animation in the 70's by F R A Hopgood, Proceedings ACM National Conference 1970
Increasing the flexibility of an SC4020 installation by P M Nelson, Proceedings of the 1970 UAIDE Conference
Putting soundtracks on your 16 mm films by P M Nelson, Proceedings of the 1970 UAIDE Conference
Computer Generated Films by P M Nelson, QUEST
Computer Generated Films by P M Nelson, Amateur Photography
Fortran routines for animation by P M Nelson, Guide to Computer Animation (ed by Hallas)(to be published)
Compiling Techniques (book) by F R A Hopgood, has been translated into French, German and Japanese
A short sequence of film on Continental Drift was shown on the BBC-TV's Horizon programme.
FOCUS was shown on Austrian television.
Film sequences generated by F R A Hopgood, G A England, R Asbury and J Lickness have been shown in the Open University's TV programmes for the 'Foundation in Mathematics' course.
FOCUS was presented at the Brighton Conference of BISFA last year.
