3- ICT/ICL 1900 Range Software

 

During the period of development of the1900 Range, the Software activity evolved from a Sales Support Aid into a major computer industry driver, which, towards the end of the period, was absorbing development resources in excess of the hardware.

 

This Section will outline its evolution in the context of the ICT/ICL 1900 Range

 

 

 

3.1 – The FP 6000 Software (Before the 1900 Range)

 

The following is the relevant extract from the “FP6000- Report on Visit to Ferranti-Packard” in April 1963, detailing present or planned software while recommending the adoption of the FP6000 as the Ferranti 1900 (To be noted that Executive was considered an item of software in Ferranti-Packard, while it became part of each “processor” development in the ICT 1900 Range and, together with Test Programs and Diagnostics, was no longer listed by ICT in the general software lists):

 

 

“7.         Programming

 

7.1 General Software

The software which either exists or to which Ferranti-Packard are committed is as follows;-

 

(1) Executive

This is the F-P name for what we would call the supervisor. It is written in modular form and provision will be made for controlling up to 20 peripheral equipments. In addition it deals with extracodes and time­sharing activities. For any particular system only those parts of the executive which are required will be supplied. The version for the paper tape system is virtually ready. The dead-line is 25th April when it is needed for the N.R.E. machine (the P.R.B machine is a special case and does not require executive). This version will occupy 960 words.

The typewriter is used for communication between executive and the operator and a set of stereotyped messages exists (e.g. SUSPEND, GO, ALTER, REVISE PRIORITY).

 

(2) Debug

This is a program that may be regarded as an extension to executive. It assits in the development of programs and can be 'steered' to permit, for example,

'Single-shot’ working with print out,

Monitoring at prescribed points in a program, or

Monitoring at successful jump instructions

This program has been flow-charted and partially written and is required by the end of April by the N.R.E.

 

(3) Assembler

This is a simple input scheme for machine language programs. It was referred to as 'a poor man's Pegasus, Initial Orders with some extra features'. It is a 'load and go' assembler with a 1-1 transformation from the written form of instructions to the machine representation. Names(of up to k characters) are allowed for symbolic addresses and provided that all symbols are pre-defined, simple arithmetic can be done on them.

The assembler enables library routines to be incorporated into a program by means of the directive "Library" followed by the names of the routines required. Other directives specify the amount of storage required by the program, the peripherals required and can allocate the priorities between the main program and up to 2 sub-programs which can time- share with the main program.

This assembler appears to provide an adequate input scheme for machine language programs. It has been checked out and is used at present as the standard way of taking in programs that are being developed.

 

(4) Fortran Compiler

A Fortran compiler, modelled on the 1620 Fortran but being made compatible with Atlas Fortran, is being written for the N.R.E. It is scheduled for completion by October, 1963. This date may be a little optimistic but the commitment to the N.R.E. should ensure that it is completed.

 

(5) FP6000 Autocoder

This is a more sophisticated input than the assembler. It handles mnemonic forms for the functions in machine language and has no restrictions on symbolic addresses. It also enables macro-instructions to be assembled. Further features which are planned include the ability to translate simple logical and arithmetic statements.

No precise specification of this language exists and its status will be affected by any decisions regarding the implementation of COBOL. It would appear that F-P have in mind a language which should be relatively easy to implement and which would make a useful alternative to a much more expensive COBOL or NEBULA compiler.

No firm dates were mentioned for completion of this project.

 

(6) Library

At present the library consists of routines concerned with the input and output of numbers, fixed to floating conversions, floating point operations, the usual mathematical functions and certain double-length routines. In all some 40 routines have been written and are being tested. The library list is appended.

 

(7) Diagnostics

This is the F-P word for engineers test programs. These have been written for the paper tape system and cover

Core store tests

Individual function tests

Paper tape reader tests

Paper tape punch tests

Typewriter input and output tests

Further tests will be written to cover other parts of the system as later orders make them necessary.

No attempts have been made to write diagnostic tests in the sense that they are used at West Gorton.

 

(8) Matrix Interpretive Scheme

It is intended to write a matrix interpretive scheme based on the Deuce scheme rather than the Pegasus one. The main difference is that storage can be freed when no longer required by the scheme.

Little work has been done on this scheme to date.

 

(9) ALGOL compiler

It is believed that an ALGOL compiler will be written in collaboration with the Saskatchewan Power Corporation. It is expected that the Corporation will supply a programmer who has previously written an ALGOL compiler for another machine to work in Toronto with F-P for one year.”