Making Better Software
I’m reading Kent McPhee’s paper ‘Design Theory and Software Design’. From the introduction:
The question that was asked in 1968 [at a NATO scientific committee conference] is just as valid now: ‘How can we predictably create better software systems?’
This is a recurring question. But isn’t it backwards? Isn’t the problem rather how to discover which software system would be the best in a particular problem domain, and within a particular budget and timeframe?