Experimental Physics and
Industrial Control System

No one was directed to use EPICS from the DoE. People that joined, did so to reduce their cost and risk. Marty Knott from the APS published a paper on this savings at an ICALEPCS conference. The number that sticks in my mind - was $1M on software development costs. Subsequent laboratories had a greater benefit as after this first phase, the Channel Access Protocol, the Process Database, a good suite of hardware drivers,  the state notation language, and a set of client tools that were adequate. Many project were able to use this version with moderate development for integrating new hardware devices. A limitation was that the process database only ran on vxWorks. 

When those of us that developed the predecessor, GTA Control System, we presented it at the ICALEPCS in 1989 at Vancouver. Marty Knott, the control group leader for the Advanced Photon Source at Argonne National Laboratory, and his engineers were ready to develop their own control system. Marty Kraimer was tasked with evaluating GTACS. Argonne decided to work with us to further develop GTACS and the result of that was EPICS. There are some other early papers that were written about collaborating. They are ICALEPCS papers as I recall.