(a) The replacement 21" screen display. We await a reply to a letter asking for a quote from DEC.
(b) More core. DEC supply core at the following prices
8K £9,500
16K £16,600.
SARL can supply core at the following prices
8K £1,250
16K £8,030
32K £12,720.
All these prices assume we claim full Duty remittance. SARL are prepared to provide a maintenance contract. We have written a letter to DEC to see if they can quote more competitive prices for their core and to discover what conditions they stipulate for the addition of another manufacturer's core.
(c) Hardware soundbox. AHF has submitted plans to DEC for an interface for the proposed box. Apparently the interface requires to use -3 volt and 0 volt levels instead of 0 volt and +3 volt levels and so extra level converters are required. The total cost of parts for the interface is £1,500. The soundbox will cost £350. DEC have approved the design. AHF will provide a full specification of the work required for the interface, so that a quote can be obtained from the Rutherford Laboratory for this work.
(d) Another disc. An extra disc surface containing 260K words will cost about £4,500. Although a new cabinet is required this will be provided free of charge.
(e) There is in fact no drum marketed for the PDP15.
(f) Line Printer. DEC market line-printers as follows
132 columns, 215 lines per minute, 96 characters at £10,580.
A version with 64 characters costs £9,820.
Datadynamics can provide a faster printer, 64 characters at £2,500. but an interface is required.
GAE will write to Datadynamics to ask for a quotation for such an interface.
We have about £20,000 in the forward-look for this year earmarked for PDP15 enhancements of which £4,500 has already been used.
2. Two points were raised on items described in SPROGS NOTE 1. The method of operation of the routine dotted gives rise to the following problem.
Suppose a dotted line sweeps across a region, such that the line crosses the region boundaries in two places. If the dotting of the line is left to the lowest level, and if each line segment begins with a complete dot, the dotting interval will vary as the line moves, giving the impression that the part of the line within the region is bouncing. This problem could be solved by storing with the line segment the original XY coordinates of the line of which this segment is a part. However, the extra complication does not seem worthwhile. The only point of delaying the dotting is to make use of possible hardware assistance and in this case the bouncing is still likely to occur. Thus lines will now be split into their dotted segments at the highest level and each segment of the line will be reduced through the regions.
3. On further consideration it was decided that the variable INTP does in fact perform a different job from the variable PCPRYV. The latter is positive or zero during a file definition and is -1 during execution, whether the orders being executed come from a file or not. The variable INTP distinguishes from where executed orders have come. Although the variable INTP will be left in the SPROGS Manual for the moment, it will probably be possible to get this information from other variables and so INTP may not be needed.
4. More comments on the SPROGS Manual:
4.1 It was decided to limit all text names to eight characters maximum, so that only one double word need be allocated for their storage.
1.2 Currently the second argument of the routine WHILE is a two character code describing the required conditions. For example, GE, LT, EQ etc. If this code were replaced by a numerical code it would be possible to provide a variable condition making use of an index variable. However, the actual condition required would not now be so obvious. Whether we should go for the generality of a number or the readability of two characters is not clear.
4.3 It was decided that it was worth including the inverse function of WHILE, that is UNTIL
4.4 There was some discussion on the modes of storage proposed for picture files. It was decided that initially routines written in FORTRAN only would access the file store and hence all arguments, numbers etc will be stored as reals. Subsequently, as much packing as possible would be incorporated in FORTRAN routines and finally the full power of the various modes proposed would be incorporated by means of PLAN routines. In this way a general FORTRAN system would be provided and extensions for our own use would also be possible.
4.5 It was suggested that the marker variable PRMINF should contain information as to the number of arguments following. However, as the name table for a given routine contains a count of the number of arguments and as a routine will know how many arguments it itself expects, this suggestion will not be implemented.
4.6 It was decided that Region 0 would be provided in the system and would allow output both to the SD4020 and to the PDP15. The basic SD4020 routines would invert the co-ordinate system as is required for that device. Thus the ability to specify basic regions for regions above level 0 will only be used for any other devices which are added on.
4.7 FRAH will provide more information on his ideas of the use of the PDP15 light buttons with respect to the SPROGS system.
4.8 There are two possible ways in which a user may add his own display routines to the system. Either he must re-compile part of the routine INTERP including calls to his own routines, or the routine INTERP should be written in PLAN so that the addresses of the new routines can be passed across and the routines entered when required. We shall use the former approach.
4.9 The specification of data mode will be left for JRG and WDS as this mode will be used mainly from the PDP15.
4.10 A user will be limited in his choice of region numbers so that it will not be necessary to store the region number in the region table. This will greatly ease the problem of access and pointers.
