This thread started here. When you're done reading please come back and weigh in.

I've been spending significant time using my iPod touch, and as some alluded to in the previous thread, the foremost reason for Apple's iPhone OS animation is to optimize the human interface. By doing this the claim is that productivity, experience and usability are improved. I agree: things are happening with my cadence, and thus I have an easier time interacting, my eyes spend less time waiting and more time focused, and the smooth transitions might keep me from wasting expensive starting and stopping energy. And when there's a multi-core Cortex A9 is in these things, I'm guessing Apple will try and reinvent multi-tasking.

I've always looked at transitions as way(s) to cover that you were doing something in the backend that would invoke a "now loading" or slight delay in displaying what the user has requested - mind you, I'm an Adobe Flash/Flex dev first so I'd love to hide "Now Loading" as much as possible.

The whole conversation about static versus animated UI's will ultimately boil down to the type of user. Some people don't mind the animated "fluff"... most others think it's wasting precious CPU cycles. Everyday users versus "purists"...

If done well, I don't care. If it's obvious... well, it's not done well. Vista is a good example of good intention, overdone and badly executed since it slowed down the OS. The iPhone... is a gesture based UI that uses movement, transitions to show progress, participation and in some cases show more information. Sometimes... it's just all flash and glitz... it just is done because it could be.

I'm just not a fan of how most UI conventions are now.

No, mine was just an analogy: in the same way that you learn a keyboard shortcut on a desktop, you learn what is going to appear on the screen, so you don't need the animation anymore.

I don't think that's a problem in the next generation of tablets, as it isn't a problem on an old desktop with semi decent hardware acceleration (at least under linux, under windows I don't know, I'm still using windows xp with the classic skin, teletubbies stopped to amuse me a long time ago), the problem is that they will slow you down.
Not an issue if you can turn them off (as my kde desktop allows me to do).

Really quickly thinking, I can think of three major categories for visual transitions:
1. To give the impression of quicker performance. For instance, in the iPhone, when you click on an application icon, a transition instantly comes that the application is launching, even though it is only launching a blank background image of the application. This creates perceived performance, and as long as devices cannot launch everything instantly, is generally seen as a useful thing, in any case. I.e. they're not there to slow down the device, they're there to make it seem faster. Omitting graphical hints does not usually make the application load any faster, or make the device load and compile the necessary data from the storage faster.
2. To communicate navigation. For instance, in the iPhone, transition movements from left to right, right to left, dialogs appearing from the bottom etc. all communicate what is happening within the application. If view A would switch to view B in the blink of an eye in all cases, without these transitional cues, users would have to decipher where he has gone within the application. Done right, these are also very useful elements to add. For instance, zooming out from an application vs. going into another view within the application.
3. For superfluous stuff. That's the 3d whizbang effects that look nice in TV ads but become quickly very much annoying.

I think it was Luca that had arguments that even though these transitions are useful in the start, once the users learn the device, they stop being useful and start being harmful. I would say differently. You need constant reinforcement: the transitions serve as these small success signals. And then again, you install new applications, use new services etc. - as long as the language of transitions is holistic, so to say, seeing the same transition in a new application instantly communicates what the action was, since it was the same transition as seen previously.

Yes, there is vast potential for misuse with transitions, but in general I at least feel - and I hope that most agree - that they definitely can serve a very useful purpose within a mobile UI. At least with 1. and 2.

*randomly jumping in*
I think a well done UI adds value to a platform as a whole. If the platform usable and it has advantages, developers will be naturally attracted to a platform. "Everyday" users, however, probably will not. They *need* the whiz-bang to validate their purchase price and get the "experience".
Since a lack of these "whiz-bang" effects/UI enhancements of other sorts leads to a lack of users/smaller user base, fewer developers dedicate time to the platform. I bought my n810 because it was purported to have an excellent internet experience, including flash, etc. Day by day, those technologies are moving forward, and we have only some efforts to stay with the times. If we had a dedicated developer base serving millions of users, we would have a much better chance at having device that keeps up with technological standards. I think the smart play for Nokia is to add whiz-bang effects to the OS and add a monetization scheme (read: application catalog) to attract more "closed-source-monetizing" developers (I know, I know - this is pretty vilified in these forums) and (more important) more open-source developers to move the platform forward.

Summary:
No whiz bang/UI optimizations = fewer users = fewer developers = stagnant platform
whiz bang/UI optimizations = more users = more developers = developing platform

Ask yourself:
Why is it that every other site these days has an iPhone/iPod touch (free) application / website modification to suit the device?
Answer: Everyday users like the experience = bigger user base; Websites/companies recognize the need to reach out to those users and are willing to take the time to develop applications in-house/outsource development to get a solution in place for those mobile users.

Just my .02 cents.

The drum that is banged most often for new mobile UI is 3D graphics - gotta have 3D graphics. Can't most agree that 2D graphics will suffice and 3D graphics are not necessary to accomplish 1 and 2?

Well, kind of yes. But the 3d accelerators can be used for 2d stuff just as well. (As probably some remember from PC's ten or so years ago, at the age of Matrox 2d accelerated vs. 3DFX display cards.)

You can use 3d for instance to do overlay and zooming in and and out and things like that. Kind of like locking the viewpoint on the Z axis and looking downwards towards the application surface.

It's a very limited use of what is possible in theory, but using it only in a limited fashion make it any worse. Just that hardware is advertised as being 3d accelerated naturally shouldn't rule out good taste in using the capabilities that it offers.

And more analogy. Emacs has "keyboard shortcuts" only, and it is
mostly old generation hackers using it. I prefer emacs myself over
many fancy IDE's but i don't try to sell it any more only truth.

When we like attract mass users, non hacker community we need
to have good looking and easy to use UI .

I remember from my university usability course that the shotrcut
way works if some user uses all of the day same application
but then entry threshold will be really high.

If some fast use expert mode is needed, we may always
implement it.

I personally think the major chunk iphone's success is its capacitative touch screen. The whole user experience changes with that. That is the single most important thing in the iphone being a big hit. The user interface is nothing great.. it is just a normal finger optimizzed ui.... but everything makes it so easy and smooth because of the capacitative touch screen.