AHF Alan Francis FRAH Bob Hopgood GAE Graham England RET Eric Thomas WDS Wade Shaw
1 . The SPROGS project will be divided as follows:
BSI Link and Associated Software: GAE
PDP15 Software and Syntax for SPROGS P Processor: JRG WDS
SPROGS Fortran Package on 1906A: AHF and RET
DCT, who is currently finishing the GROATS package on the 1906A, will no longer be working in the SPROGS Group.
2. The SPROGS Group will issue three series of papers. These will be entitled:
SPROGS Papers: complete ideas which have not yet been implemented.
SPROGS Notes: comments on these ideas and meetings held.
SPROGS Technical Papers: details of finished work and other relevant items of a technical nature.
RET will be responsible for updating the SPROGS Manual at present held on :GSIN00.SPROGS MNL.
3. The £olloving possible enhancements to the PDP15 are being considered.
(a) A replacement display, 21" screen. It was thought that we could obtain this new display as a trade-in for the old at an extra cost of £2000. JRG has prepared some test programs, which produced bad results on our current display to send to the States to be tried on the new display.
(b) More core.
(c) Hardware sound box. The device and interface will cost about £2000. AHF could design the interface.
(d) Another disc. It is apparently necessary to get a new disc drive as the current cabinet is full.
(e) Drum.
(f) Line printer. Data Dynamics are offering a printer which they are prepared to interface to the PDP15 at a total cost of £4000.
None of the Group feels that a card reader is necessary. GAE will obtain more information about these options.
4. The BSI link to the PDP15 is still giving trouble. GAE has managed to receive some signals at the PDP15 end but is still having trouble with the 1906A side of the interface. The BSI on the PDP15 is not wired according to the logic diagrams. Further information is being sought on this point. An extra delay has been caused through the engineering responsible· for the BSI being away on a course.
5. The rest of this paper deals with discussions on the current SPROGS manual. As the manual is updated, some of these comments will be incorporated.
5.1 One of the aims of the SPROGS system is to make it readily available to users at other places. Hence the version of Fortran used to implement the system should be as machine-independent as possible. Thus, for example, no identifier should be longer than 6 characters, and no recursive techniques should be used. As there is some disagreement among Fortrans on the passing of text arguments it was agreed that such arguments should be confined to functions such as NAME and FLNMB. All variables in the system would follow the implicit Fortran conventions as regards type. No REAL or INTEGER declarations would be used.
5.2 The current SPROGS description in the manual does not deal either with error messages or with organisation of backing store. These points will be considered at a later date.
5.3 The routines EXPAN and ROTN, section 3.7, will in fact act immediately on a file with no continued application through the region levels. Their arguments can therefore be taken as absolute.
5.4 Routine THICK will have to be implemented in the following way. For NEGATIVE arguments a preliminary reduction through the regions of a pair of current coordinates will provide the conversion factor necessary to allow the extra lines to be inserted such that, after reducing all the lines, the final effect will be a collection of lines raster unit apart. Scissoring effects prevent leaving the application of the required thickening until the last region. Initially it will not be possible to take advantage of hardware output devices which possess a hardware thickening. For positive arguments the required number of lines is inserted immediately but as the lines are reduced through the regions it will be necessary to add more lines to prevent the thickened line moving apart.
5.5 When using the routine DOTTED it will be necessary to calculate the required line lengths in the current region, store the information with the line segments and convert these line lengths as the line is reduced through the regions. In this way the correct density of dotting will be maintained and there will be no problem of crossing a region boundary. As in GROATS a negative argument will imply that the line is stripped into segments all of which will remain visible. This is to assist with wipes.
5.6 Instead of separate routines FIDF and FINRDF to end a picture definition, a single routine FIDF will be used with one argument. In this way, more reset options will be available and it will be possible by making this argument an index variable to store single copies of library pictures.
5.7 By defining subfiles and possibly editing them later it will be possible to introduce recursive file definitions. No check on such recursion will be provided. If a closed-loop definition is encountered the picture file will continue to eat up store until no more is available.
5.8 When a name or a picture file is deleted it will be necessary to leave that name in the name table in case there are other files which refer to that name. In this way there will be no problem of pointers in a file pointing to old blank entries.
5.9 It is intended that the sequence list be executed from the top downwards; hence the DRAW routine will add the referenced picture file to the end of the sequence list and then begin execution at the top of this list. Thus, if the routine ADSQLS has been used previously the file referenced by that routine will be executed before the file referenced by the DRAW routine.
5.10 When using the routine SPSOLS it will be necessary to check that there is still something in the sequence list that is capable of being executed. In this way a possible deadly embrace situation will be avoided.
5.11 It is envisaged that the global priority variable PCPRYV would have the value Zero when the file contains machine code orders for one or other output device. Thus it will have a value -1 only when the file is being directly executed. There is thus no need for the extra variable INTP to specify whether a file is being directly executed or not.
5.12 Rather than using the routine PRMT to indicate whether an argument is an actual value, an index variable or a trio of index variables, functions will be used acting on the actual arguments themselves in the call of the desired routine. The idea of a trio of index variables specifying file name, file number and position within the file, allows one file to get at arguments stored in another file, thus using that file as data. This can be a general feature, although to be used generally, the user must know something about the way items are stored within a file. An extra routine might be used to set up a true data file which could safely be accessed linearly.
