One of the author's frustrations, seeing months of work abandoned - is the byproduct of training programmers as scientists and engineers rather than designers...or rather it is the byproduct of a pedagogy which attempts to extend mathematical logic and proofs to messy social problems.

Landing pages and NoSQL backends are not deductively better or worse: at best they are statistically or inductively so. Typically, however, choice is simply intuitive.

Making good intuitive choices is the essence of design and the way to develop good design intuition is not to sit down and bash on one's first solution until it plausibly works. Education should attempt to teach students that falling in love with one's first idea is a tarpit.

Teach students the practice of coming up with three independent design schemes and developing each to the point that the merits and flaws of its underlying design concept is evident. Maybe one is worth developing.

Throwing away months of output is not wasted effort - just as lines of code are no more a relevant measure of productivity than pencil lines on a blueprint. Artists have stacks of sketches and studies, architects stacks of abandoned schemes.

Most ideas are not worth loving, let alone executing and most solutions cannot be demonstrated correct via mathematical logic. The education of programmers should align with this reality.