To be clear I'm not trying to say "lets use waterfall to preplan the software"
I'm more saying things like "Does design, modularization, organization, or things like SOLID principles" matter at all? When i've had projects that I've had near total control over I've been able to get to a place where many product asks were trivial. Or the obvious extensions were. Things like "Can you add a new permission to the sytsem, or cover a new resource type with permission constraints?" I built a system where the other proposal would have taking a large constant amount of time to add each new resource permission (each permission was a bool column on the user table) . Whereas my design I simply added a couple string consts and a wrapped the resource's ORM portions in a decorator. Basically done in the time it took to have the meeting where the product person asked.
But these kinds of future value software seems to be denigrated because the opinion seems to be that its going to take far too long to implement, but in my experience the "good" solution and the crap ones (accounting for marginally excess bugs) are at most 10% longer to implement. Sometimes the crap implementation appears 50% faster to do, but then an observant person would note that it has pernicious bugs for months or years that destroys the team's velocity.
I'm more saying things like "Does design, modularization, organization, or things like SOLID principles" matter at all? When i've had projects that I've had near total control over I've been able to get to a place where many product asks were trivial. Or the obvious extensions were. Things like "Can you add a new permission to the sytsem, or cover a new resource type with permission constraints?" I built a system where the other proposal would have taking a large constant amount of time to add each new resource permission (each permission was a bool column on the user table) . Whereas my design I simply added a couple string consts and a wrapped the resource's ORM portions in a decorator. Basically done in the time it took to have the meeting where the product person asked.
But these kinds of future value software seems to be denigrated because the opinion seems to be that its going to take far too long to implement, but in my experience the "good" solution and the crap ones (accounting for marginally excess bugs) are at most 10% longer to implement. Sometimes the crap implementation appears 50% faster to do, but then an observant person would note that it has pernicious bugs for months or years that destroys the team's velocity.
Also just an anecdote and hypothesis.
But here's an appeal to an expert who I agree with https://martinfowler.com/articles/is-quality-worth-cost.html