Both of your examples are wrong. It is well-known in aerospace industry that cockpit interfaces need vast improvements, and that is actually an active area of research. Pointing to the complexity of an airplane cockpit and saying "look, not all interfaces need to be pretty" betrays a fundamental misunderstanding and/or lack of knowledge about the subject.
Second of all, the reason desktop environments are still preferred for "serious business work" is because of their features, not their UI. I work for a business-to-business software company that has been around for over 25 years, and our product is on version 9. We recently released an iOS app. It is being adopted slowly, but not because of its interface. The interface on it is actually very, very clean and effective. The problem is that we have not yet had the time to port all the features on the desktop version to the iOS version. Therefore it is not used for "serious" work... yet.
I didn't say the cockpit couldn't have improvement, I said could you imagine running one with a smartphone interface. It's really not appropriate for the situation.
Your other example seems to willfully ignore that features are stripped out of mobile OSs for the cause of simplicity, not because they're 'catching up'. Yes, for things that are solely within programs themselves you have a point, but we're talking about the operating system here. Quickly switch between windows? No. Have more than one thing visible at a time? No. Default input mode allows for quick entry of data? No. I mean sure, your program might be the ants pants, but what happens when the user needs to enter data while viewing another document?
>>Yes, for things that are solely within programs themselves you have a point, but we're talking about the operating system here.
I don't think the distinction really matters. In this context, you can think of an operating system as a "master app" that has escalated privileges for things such as root-level access to system functions and the hardware.
>> I mean sure, your program might be the ants pants, but what happens when the user needs to enter data while viewing another document?
This soon won't even be an issue, because the data will not have to be manually transferred between documents by a human user.
Can you really not think of any situation where a person might want to enter data while being able to see another document? Not even a coder looking at a reference guide?
Second of all, the reason desktop environments are still preferred for "serious business work" is because of their features, not their UI. I work for a business-to-business software company that has been around for over 25 years, and our product is on version 9. We recently released an iOS app. It is being adopted slowly, but not because of its interface. The interface on it is actually very, very clean and effective. The problem is that we have not yet had the time to port all the features on the desktop version to the iOS version. Therefore it is not used for "serious" work... yet.