Perhaps about 1958 Dan Slotnik visited Livermore from Westinghouse where he was proposing to build a highly parallel computer. He called the proposed design ‘Solomon’. It would be bit serial and memory reference to core would bring many bits, each of the same weight but for different values. The design was inherently fixed point. Its precision was very flexible.
At the end of the day and presentation he invited those in ear shot to San Francisco for dinner. 12 accepted. He asked what was a good place to eat and someone suggested Ernie’s. I sat near Dan. I recall hearing something of the sort “bring us your best”. The meal was indeed superb. At the end of the meal he looked at the check for a moment and said: “Now I know why you want floating point.”.
Slotnik indeed went on the build the Illiac IV, with floating point, which broke new ground in parallel computing.