I recall that Peter Norton was also considered responsible for why many DOS applications required pretty much error-for-error IBM compatibles - because he advoc

I recall that Peter Norton was also considered responsible for why many DOS applications required pretty much error-for-error IBM compatibles - because he advocated for directly dealing with hardware and disregarding DOS APIs (which, admittedly, were bad)