Even by 1964, the number of proficient computer programmers was low. The number of research scientists with a wide understanding of computational techniques and also proficient programmers was even lower. A partial solution was for those with skills to make them available to those without.
At Harwell, the choice of Fortran as the computer language to be used brought with it the possibility of having a library of subroutines defined either in machine code or Fortran and made available via a set of relocatable binary cards to anybody who wanted to use them. Slotting a few binary cards into your card deck might save you 6 months of hard programming.
A library in use at Harwell early on was the IBM 704 Fortran Program Library distributed by SHARE. It was a mixture of machine code and Fortran subroutines of general and specialised use with a classifications scheme.
A second widely used Library was the Quantum Chemistry Program Exchange (QCPE). An example was MIDIAT (QCPE No 29), used and modified by Harwell on the Risley 7090. It was later ported to both the Atlas and the the IBM 7030.
In 1962, informal sharing of routines was replaced by the creation of the Harwell Subroutine Library (HSL) which was added to through out the period 1965-1973 and made available on Atlas.
It was clear that similar specialist Libraries on Atlas could be of value in specific areas. Some examples of Libraries mounted on Atlas were:
No complete record of libraries were kept and it was quite possible for a group of a specific discipline to instigate a library between themselves without needing Atlas staff support.
Most of the libraries written at Chilton were moved to either the ICL 1906A or the IBM 360/195 when Atlas was turned off. As talk to an Atlas replacement started in 1967 and the ICL 1906A was ordered in 1968, this was a long drawn out process as the ICL 1906A did not arrive until 1971. Effectively, full running of the Chilton Atlas only lasted for a year before the rundown was started resulting in its closure in 1974.
