> This worst thing that the GNOME devs did
> was make it so GNOME 2 and 3 could not be
> installed side-by-side.
(btw, just compile one of them with a --prefix and I'm sure it could be made to work. The only problem is FHS fundamentalism.)
> But it was clear that the features were indeed
> coming back and things would get better.
It wasn't! A lot of features ("features"?) never came back.
> Also I think it's foolish to assume that history is repeating.
FWIW I agree that it's foolish to _assume_ this.
However. I also think we can say that people thinking/saying the same things about GNOME 2 were _wrong_. So I think it's also foolish to assume that similar comments are now on-target.
Now, I agree. GNOME 3 _may_ turn out badly.
But it won't be for the _reasons_ most people are talking about here.
We know from experience that the "methodology" Linux discussion forums have for trying to understand desktop UI changes and their effect is flawed.
It's because commenters have a lot of wrong "folk models" about what makes a good UI and what "most people" are like and so forth, which are simply not accurate. Commenters also aren't able to see into various tradeoffs (both design and resource based) that become a huge factor in real life outcomes. And nobody can predict what "batting average" developers will have in making the right judgments; or what random external factors will get involved.
I think a lesson from GNOME 2 is that flames and "widespread outcry" can be wildly wrong, because we have GNOME 2 as an example of success despite that.
GNOME 2 doesn't give us evidence that flames are _always_ wrong, just that they _can be_ wrong. So for GNOME 3 it remains to be seen.
However: here's what it means for the GNOME 3 developers. They should not "listen to" the flames. They may want to _extract information_ from them - there's some content there, about certain users. But they should not "listen" in the sense of doing exactly what those flames are advising.
The flames are one data point among many, just as they always are with every piece of software. Developers and designers have to go chase down the broader, more thorough data they need, and apply copious amounts of their own judgment.
They may get it wrong, or not. But that doesn't mean using judgment was a mistake, just that when using judgment, one can be wrong.