In fact, the response to KDE 4 was quite similar.

The problem with KDE 4.0 was that it was simply nowhere near release quality; huge chunks of it were either massively buggy or simply not implemented, and it took them until 4.3 to get it to a state that I was prepared to use it as my desktop. At that point it worked pretty much identically to KDE 3.x, even though the underlying code had been completely rewritten, and I could just ignore the new stuff I had no use for such as activities without it affecting my desktop experience at all.

Unity, while it's less ugly than GNOME 3 and not quite as awful, still shares many of the same fundamental design flaws. No amount of bug fixing and tweaks can make a desktop usable when the basic ideas behind it are so wrong.

What concerns me in particular, is that at least with the KDE 4.0 debacle, there were quite a few "This is BAD... But I can see it being good in a few versions" type comments.... with Gnome 3, can't think I've seen a single positive comment, at best "changes are sorta manageable".

Yup. The problem with KDE 4.0 was that it simply wasn't even close to release quality; it was early alpha at best, with large chunks of functionality either buggy or unimplemented. 4.2.2 was the first version that I could imagine anyone using as their default desktop; I switched somewhere in the 4.3.x sequence.

The problem with GNOME 3 is that the fundamental design decisions are wrong. No amount of bug fixing etc. is going to make it usable.