Although the first version of Microsoft Windows had been released in 1985, Scott compiled his code to run in DOS, as it was faster for the kind of stuff he was doing…but it still wasn’t real fast. Part of the problem was that DOS was single-tasking and single-threaded (this is from my memory, which isn’t quite as good as it was 30 years ago…I will try to confirm the terminology).
I remember going over to Scott’s new office to see how things were going. He asked me to boot up a computer (into MS-DOS) on the workbench and run a test program that was sampling data. Once that was running he asked me to type in a box on the screen. As I typed there was a lot of lag between when I pressed a key and when the letter appeared on the screen…sometimes the screen was half a dozen letters behind my typing.
He then rebooted the computer and asked me to repeat the test. This time the screen refreshed as fast as I typed. And the incoming data was processed at the same speed. He explained that he had re-written DOS to make it multi-tasking.
I was kind of blown away…this was obviously a superior version of DOS…if Scott could do this why hadn’t Microsoft done it a few years ago?