In my experience, a lot of the bloat in software comes from the decision to specifically pick user interface designers that know nothing of software engineering. The basic idea is that the user should get the perfect interface for them instead of whatever monstrosity is more convenient for the developer.
I agree with the sentiment to a large extent, but the limit should be where the change slows down the software significantly. That doesn’t mean the user interface shouldn’t have features that are very expensive at run time, but it does mean that there should be a dialog about cost and benefit.





Pseudocode: