The subject came up at work today of software that takes in feeds of radar rainfall data and lets you display it, do calculations on it, etc. I mentioned the fact that the first company I worked at had done software on PCs to do this about 28 years ago. I racked my brain and remembered it was called RAU for Rainfall Analysis Unit. A search for "software sciences" "rainfall analysis unit" comes up with one hit (and "software sciences" RAU comes up with two irrelevant ones).
The hit is to this PDF http://nora.nerc.ac.uk/15330/1/N015330CR.pdf which has the foetid spore of an EU project, done in the late 80s / early 90s I guess. The address is the Institute of Hydrology (now part of the Centre for Ecology and Hydrology), which is down the road from where I work. Popular legend has it that when the research councils were set up in the 1964 (? CBATG ?) reorganisation, hydrology fell under NERC and hydraulics fell under SERC so they built a new building for the NERC stuff next to what then became the SERC one. (All together now 'back in the DSIR, you don't know how lucky you are boys').
Anyway, the paper says "The data were restructured on the University's mainframe computer (VAX 1170, subsequently Sequent Symmetry) to be campatible with the visual, graphical display package "Rainfall Analysis Unit' from Software Sciences Ltd.""
It also says "The simulation software used is the WASSP-SIM section of the Wallingford Procedure (National Water Council, 1983), a software package which allows an urbandrainage network to be mathematically modelled and simulated using either design
storms of stated frequency or by actual event data". WASSP is a distant ancestor of the software I now work on - about 6 generations of software before the current one if my recollection of tribal lore is correct. Thus a link between my first job and the current one. Then again there have been strange congruences between my jobs so I shouldn't be surprised, I suppose.
Interesting, of course, to note that, IIRC, the RAU software ran on an 8086 DOS PC with twin floppy drives (and an expensive display card and some sort of interface to a leased line, I think). I imagine the software being discussed today would require rather more resources.