It's a nice case study of too much engineering and too much technology, with far too many unnecessary features to start with, with too much of a roll-out (the basic feature demo took over 45 minutes) to too few people without nearly enough thought into actual use cases of a rather complex piece of software that simply didn't perform that well in real world scenarios.
"What's a Wave?" - should be answerable in a couple of sentences, I don't think it ever achieved that level of refinement before being put out into the world.
Even though the basic idea was great, performance of the software in-browser has never been that great, server performance was problematic, especially when waves became too long, no export option of collaborate waves to documents, no real integration with email. Plus it was hard to organize waves. After it went pseudo-public nothing really new happened to it and old weird problems like the funky scrollbar were never addressed.
That being said, I've used it on a number of projects, a couple quite big six-month ordeals, and found it an amazing collaborative environment if you disciplined your team a bit and organized things well at the start of the project.
The concept, combining several, different, modes of communication in one environment is pretty sound actually, I just think that the modes were combined a little too singularly. Certain things like real-time chat and document collaboration are fundamentally different things that should have been relegated to different parts of the interface rather than jammed into one place.
"But despite these wins, and numerous loyal fans, Wave has not seen the user adoption we would have liked." Duh, the rollout was rather botched and given what I think I've detailed it's no wonder.
"What is Wave" should have been "a document collaboration tool with real-time side-channel chat and group email" would have been better for the use-cases my group fell into.
What I was really excited about, and what never seemed to materialize in terms of a release from Google, was that ability to run my own Wave server and federate with other Wave servers. That's what I wanted, and what never happened.