First of all, there is of course GNOME Shell, which, yes, is somewhat exciting. But apart from the integrated notifications and a few other hings i cannot really see the ‘revolution’ that couldn’t have been achieved by extending compiz.
Today i found this great post on planet gnome. Ever since these new switch buttons were introduced to GTK i was asking myself ‘what for?’. Checkboxes serve the purpose completely well and we have spent years finetuning implementation recommendations around them (like having the label follow the box or phrasing it to express a positive state). Now we’ve got these toggle buttons emerging, that
- take up more screen real estate
- are somewhat ambivilent in expressing their state
- obviously don’t provide the direct ‘this statement is active’ metaphor that checkboxes come with
- are not available on websites and hence introduce a metaphor gap between desktop and web
And all that just because they provide a somewhat more spacial interaction on a certain proprietary touchscreen phone that too many people are hooked up with, anyways?
GNOME 2.x has been criticized a lot, particularly for it’s slow movement in regards to adapting new features. But it had a very well founded premise and that was to strive for simplicity and to evaluate what really makes sense. Somehow the ‘let’s be fancier’ movement that drove 3.0 has managed to bypass these principles, at least in some areas.
I guess my biggest criticism is, though…3.0 has been the major development focus for about 2 years now, during which time 2.x has seen only minor feature additions. I know a lot of work has gone into refactoring and that is great. But where could 2.0 be now if all that GNOME Shell and certain other fancification effort had gone into adding the same functionality in a more modular fashion as 2.x used to be. We already had compiz and we already had dozens of fancy docks.
Now what’s done is done, but it looks like a GNOME 3.1 will be required quite quickly after the big release to provide remedies for a whole bunch of silly decisions.