Great comparison.
As a sidenote, basic authentication is part of the HTTP standard, so suggesting a truly open, standards based way of identifying a user for the purpose of accessing a service/system is not such a crazy idea.
It's not a dying concept, so there's scope still there.
For how long was CSS developed before there was large scale adoption? It seemed to be seemlingly forever people were doing table based layouts even after it was available everywhere, and it probably wasn't available everywhere until many years after its inception.
> relatively few folks were willing to actually back up that belief with action
Which in this case turned out to be the correct decision. Nobody wants to implement a dying technology, no matter how good it is technically.
Most generally wait until it is best practice among the bigger players until they implement something, uness there is a clear user need.