It's good to get a list of these milestones as a general food-for-thought exercise, but I also think it's a bit dangerous to make generalizations from any of them individually, rather than thinking of them as part of a continually evolving and extremely complex relationship between the people who make software and the people who use it.
The fact is that none of these events occurred in isolation; each new major site redesign comes with the baggage of the last. As the demographics of Internet users shift (with technology becoming cheaper/more accessible and techno-jargon spreading with it) and as the very concept of a "redesign" becomes more mainstream, people approach change with a very different set of expectations year to year. The first time Facebook introduced a major new redesign, there were mobs armed with torches, out for blood. The 23435th time Facebook made some tweak to their interface since then has been met with shrugs of apathy – in fact, if someone were to start a petition today to abandon Facebook because of some new change to their interface, I think most Facebook users would giggle and find that adorably quaint. "Aw, grandpa hates the new like button..."
But that was a very particular – and in many ways irreproducible – set of circumstances. We have no way of knowing what would have happened if the first redesign had been rolled back, or if it had been a set of incremental improvements rather than one big one. And no website can ever do a redesign again without risking comparison to that moment, rightly or wrongly. That's why it's dangerous to try to use history as hard data – not enough data points, huge bias, and no way to A/B test ;)
So, I don't think we should be looking back at any one notable redesign event or set of events for literal arguments pro/contra one huge redesign versus many small incremental changes, how we should or shouldn't be communicating changes to users, whether we ride the wave of criticism or roll back, etc.
Or, at the very least, we can't *just* be doing that. What we should be doing more of, imho, is looking forward, thinking about who our users are today and who we want to attract to our projects tomorrow. The very fact that "well of course the average person hates change" is becoming a platitude among the digerati is a strong indication that the trend toward more change isn't likely to go away anytime soon. It's extremely unlikely that in the next 5-10 years, fewer people will own multiple Internet-enabled devices, visit fewer sites and apps, and know less about technology... what does that mean for the interfaces we build today? How can we ensure that we're not spending all our efforts solving for problems/users that won't exist in 5 years' time? Those, I think, are things that are just as if not more important for us to start thinking and talking about.