John Honniball wrote: I did a year's sandwich at UKAEA Culham Labs, Oxon. from 1987-88 and used a bunch of Primes there. I thought they were fascinating machines. They had recently upgraded to Primos Rev.19. I was told at the time that even numbers and odd numbered revisions were for different purposes -- anyone know what the difference was? I know that a big part of the change to Rev.19 was dynamic linking of libraries, so I was involved with recompiling loads of utilities to go from using SEG (static linker) to BIND (dynamic). I liked CPL, Prime's shell language, but most users of the computing services used UKAEA's own shell, GPL(?), which had been implemented on Primes, VAXen and an IBM 3084 at Harwell. This gave a common ground to physicists who had to use many different machines. One common scenario was develop a program on the Primes, transfer it through the IBM to the Cray 2 at Harwell, run it, and transfer the results back to the Prime for printing out. At the time I was there, many physicists were moving away from this complicated system and getting more power on their desktops. There were a lot of Whitechapel MG-1 workstations around the place, and Suns were starting to appear. Whitechapel came along to demo their latest and greatest colour workstations, running a windowing system which I think was called Oriel.