For the title "Glad there is no universal portrait support", the word universal here is the key.

As people discuss here, there are certain features that would be very useful for portrait. My personal #1 vote would go for the Media player, the Now playing view. I've been using now the N900 as my mp3 player (32 gigabytes of space and all that), and changing songs would definitely benefit from the portrait mode.

Then again, there are a lot of applications that ... Well, I couldn't care less if they wouldn't support portrait.

However, when we talk about universal portrait mode support, and somebody would actually do it, it would mean spending weeks and weeks (months, really) trying to make all views work in portrait. Even those that are hard to fit in portrait and that are really not that needed...

... I guess the question is that what are you trying to do portrait mode for. Is it either portrait mode for one hand use tasks or portrait mode for everything. For Maemo 5 I would personally say that trying to fix the "smaller issue" of portrait mode for certain critical one hand use tasks. If one would try to get the best "bang for buck", one would do a list of most critical missing one hand use cases in applications and try to fix those, one at a time.

What if the use case is simple data entry? Right now, its 2 hands. What then?

Data entry with one hand in portrait is definitely tricky. Just making an UI portrait doesn't automatically enable one hand entry. It's a rather limiting area on screen that is easy to press with the same hand that is holding the device. In theory you can try to reach all the corners with the same thumb, but in practice it becomes very cumbersome very quickly.

Data entry, if talking about text input, is also problematic. (And even more problematic with one hand.) You could do a T9 design, but then again, T9 is just far slower than the HW qwerty. The effort of switching to landscape and starting to use the qwerty becomes very small in comparison to the effort of trying to type something with T9. Then again, Apple has a qwerty in portrait. In order for that to work nicely, you really need a predictive engine (as Apple has) on top of it, trying to smooth out and predict your finger presses. (Yes, you can try to live without it, but I'm pretty pessimistic about how long users would be willing to use a portrait qwerty without it.)

HUH? Its the oldest trick in the book, and biillions of portrait devices shipped with qwerty and T9. You lost me there.

Not true again. Check the real tests. The fastest texter in the world uses T9, and couldn't get 40% of the speed in qwerty. Where do you get the idea qwerty is faster?
check these links:

http://www.symbian-freak.com/news/00..._or_qwerty.htm
http://jace.livejournal.com/446536.html

I'd love to do another speed test shootout to prove a point...

Your speed with T9 slows down when you type a lot of non English words especially when using slangs

I'd say it comes down to what you are used to. After using T9 ever since it started to appear in cellphones I'm quite sure I'm way faster with it than the 3 row qwerty on the N900, but after some time I think I'm unlikely to reflect over the difference in daily usage.

doksng: After you used the slang word once it *should* be in the T9 dictionary and shouldn't slow you down the next time

yes it should but, it seems to have a buffer and replaces older ones with new ones as you add more and more

This is a big limitation, and I've begged Nokia, who helped develop the T9 system with Samsung, to update the spec to include multiple dictionary support for using two or more languages at once, plus increase the base size of the T9 dictionary, which doesn't store enough words. These are big issues that turn people off to T9. Were they fixed, it would proliferate even more.

<passing.opinion>
If you're too busy to free up both of your hands, you are too busy to
type on your device.
</passing.opinion>

Creating two interfaces for any app is mandatory if system-wide-ASR is implemented.

Also, the user would have to learn twice as much (or more if no similar logic is used in every app in rotating - this would have to be clearly defined in guidelines) to use the device. This adds to learning curve and slows usability in those situations where the system tries to choose between orientations. Perhaps a way of lessening the confusion in users when the screen is rotated could be a way to show the user how the interface has changed - maybe by animating the change completely, showing each button or image moving to the other location. Otherwise the interface would have to stay pretty similar and that creates issues due the aspect ratio and resolution. This naturally depends on the app.

Delay will be introduced and this creates frustration, especially when the user will have to experience and recognize the screen rotation situations to know how to avoid them if speed of use is needed. This adds to the burden of what-not-to-do if robust functionality is expected. This surely is not wanted.

If screen rotation has to be implemented, it must not confuse the user in any way. Some devices that have ASR have a very steep learning curve, ymmv. This lessens the overall appeal of the device.
ASR would have to be clearly defined and explained.

IF ASR will be implemented in the OS, I propose a control panel app where the user can allow/disallow/let-the-system-choose screen rotation per application basis. A complete system wide ASR disable is stupid, the phone functionality is based on the screen rotation at the desktop (the 'complete' button should perhaps be relabeled as 'the way this device was designed to be used').

As the current situtation is that any 3rd party developer can choose whether to have a portrait interface... I'd reallly leave it at that. Atleast, if ASR would be implemented Nokia shold have it as a installable feature afterwards, and not as a forced update later. I imagine Qt and maemo 6 will have more flexibility to have ASR functionality.