If they had created a new product instead of bumping Gnome's version number all these problems would have gone away because the user's expectation is to have something different instead of an evolution of an existing product.
They're trying to build a better product, and they're only a bunch of part-time volunteers.
At some point they had a bunch of features, but to keep up with the other desktops (Aero, Quartz) they had to focus on rearchitecting and, being a small team, there haven't enough manpower to keep all the features.
I also don't think language (api's) are not at all comparable to desktop environments. There's just so much going on in a desktop environment.
I thought quite a few of them worked for Redhat?
>> At some point they had a bunch of features, but to keep up with the other desktops (Aero, Quartz) they had to focus on rearchitecting and, being a small team, there haven't enough manpower to keep all the features.
That seems a bit of a rosy view of what happened. A cynic might say that they decided to scrap everything and start again, users be damned, and went out of their way to make 3/Shell incompatible with 2 to force people away from it, announcing that it was now abandoned and of no further interest. Then, as now, they explained how everyone that didn't like that move was wrong and stupid, and they set about making the desktop experience that you will absolutely love because we say so and if you don't do it our way you're a curmudgeon and intellectually subnormal.
Ironically, the Redhat-employed ones seem to be working on Classic mode (so they can release RHEL 7 someday), which minimizes the damage the non-Redhat-employed ones have done: http://news.softpedia.com/news/Red-Hat-Linux-7-Will-Have-a-T...
and not only this cynic would be wrong, but he would also be full of crap, because it's such a trivial thing to check in the public repositories, public mailing list archives, and public bug tracking system.
but, obviously, crapping on a bunch of volunteers trying to work on the project they like is easier.
Are you saying they didn't force incompatibility with Gnome 2? Or that they didn't just decide to change everything to exactly how they wanted the desktop, regardless of users?
Maybe they didn't scrap some of the underlying libraries, but the whole look, layout, interaction pattern/user flow and pretty much everything else was a total change.
How much time that is spent on packaging, documenting existing features, building communication systems, and doing cool new features is indeed hard resource questions for a project manager. Then there is policy questions which effect resources.
Python demands PEPs to be written before changes is considered. That demand for PEPs is less a resource question, and more a policy question. It forces developer to write down in words what they try to accomplish with a change, and also write down what effect it has on users.
In my opinion, the community reaction to gnome3 is not very surprising. Its a result of "less restrictive" developer process and few efforts into mitigation of community outrage. I can only hope they will find some insight from other projects, and start to handle feature changes with more user awareness.
Next step is of course to find understanding that users' reactions are based on numerous reasons, rather than originating from one source. Losing access to features previously found is one cause, but so is the lack of rational and balanced discussions explaining why the features were removed, and missing instructions on how to get similar functionality in Gnome 3, or the lack of tools to migrate old Gnome 2 applications to Gnome 3. To take a comparing example of Gnome2->Gnome3, lets look at Python 2->Python 3 and see what the Python Software Foundation did right.
#1: Features removed were announced years before they were removed, with clear instructions on how to migrate to better and improved features.
#2: 2to3 was invented, an automatic translator for source code from Python 2 to Python 3, to make it easy for developers to migrate over to new version.
#3: Any feature that was removed had clear documentation and discussion that explained why it was removed, and what users could do to move forward.
#4: New features were back-ported from Python 3 and made available in Python 2. This allowed users to test out things from the security and familiarity of Python 2, without getting tangled in all the changes at once in Python 3.
#5: Python 2 and Python 3 was/is developed in parallel, giving people plenty of glimpses (and warnings) of the future, and insight over what would be future safe activities.
Gnome 3 did none of these things. One could argue that there were practical reasons why, but not doing any of those things will hurt a project. At any rate, I have not seen Gnome 3 give any practical reasons why they did such an abrupt move from Gnome 2 to Gnome 3, beyond that they "had to".
(Written by a Gnome 3 user, on Gnome 3)