George, the Disc A Hung wasn''t a hardware fault - it was a race condition in the hard disc controller code. I''ll never forget that bug, it took me 18 months on-and-off to fix the b****y thing. No matter how hard we worked the system it wouldn''t fail. It turned out you had to work the CPU hard, and the disc hardly at all to make it happen.
And the fix was to swap to adjacent sections of code, one 3 lines and one four.

I worked for the IT department of the UK Home Office. We had quite a few systems running on these machines. The ones I was involved with were for the centres that maintained and repaired radio equipment for the Police and Fire Brigades.
I wrote a system using the Dataflex 4GL system to track repairs through the workshops - up to 4 terminals, some in the stores, others in the workshops.
I do remember having to write a small assembler program to allow the users to enter the system date in UK format on startup (the clock was not battery backed), as these were really rebadged Rair machines, which expected data entry in US format.

I worked for a computer dealer which sold these to run our proprietary accounting software. There was a problem whereby, seemingly at random, the machine would freeze with a "Disk A hung" message (Disk A being the hard disk). ICL eventually acknowledged this as a hardware problem although, by then, our small number of ICL customers had been moved to other systems due to their unreliability.

Our software was written in MS Basic and I spend two weeks converting it it to CBasic - only for an MS-DOS emulator to appear which ran the MS-DOS versions of our programs substantially faster than the native CBasic ones.

The OS, Concurrent CP/M, didn''t have tree-structured directories, it had numbered areas on the disk, 0 - 15.