[__alias]

I buy into the same belief as you here, but I guess you could easily argue that you could create a suitable fitting abstraction earlier on with the understanding that you can "detach" them once the point that they're fundamentally different comes

[consilient]

The point of abstraction is to reduce the number of concepts in play. If you're still tracking which old concept is "really" being used every time, you haven't actually abstracted over anything, you're just naming things badly.

[capableweb]

> The point of abstraction is to reduce the number of concepts in play.

I'm not sure I agree with this. For me, the point of abstraction is divide the number of concepts between the layers you introduce, effectively to hide concepts from the layers where you don't want to have to care about them. Often times, abstractions adds the total number of concepts at play, but hides them beneath/above the layers.

[fluoridation]

The problem is that there's an impetus to continue working on top of established facilities, because it's usually incrementally less work than reworking a piece of code into something else. Plus it's difficult to recognize ahead of time when something is about to become a problem, rather than fix something that's already a problem.