That's arguable. Companies which are involved in PlasmaActive see no problems in making a fully open UI. Closing stuff is not the only way to be successful.

Canonical and Microsoft don't do that, but Apple and Google do.

There are plenty of pitfalls here (pricing, distribution, patents???), but I'd argue the basic strategy, vague as it is thus far, seems sound.

Four words: one click purchase patent.

Really? Even if you want OSX?

From a commercial perspective, where a new or novel idea provides a level of differentiation that defines potential for success, I can understand, and even support (The FOSS angel sitting on my other shoulder is cursing at me atm) the decision to have some closed components in a "package" that includes hardware, drivers, os and ui.
The level of support through bug reporting and the speed at which fixes are pushed, will determine how much of a problem they present to an end user (not talking devs here).
For many here, the biggest gripe about the closed components with Fremantle, related more to the delay in bugfixing (let alone all the WONTFIXes) than the challenges these components presented when devloping via the sdk.
Jolla, to all extents will aim to release a device that is as open as possible, while still protecting any IP relating to aspects of 3rd party hardware and their own IP that they see as being marketable/profitable from a commercial perspective.
This model is not really any different from that used by Nokia on the N900 or N9 and it certainly didn't stop me from lining up and buying either of these when they were released, not for what they couldn't do, or didn't have, but for what they represented and it won't stop me from buying a device that Jolla releases if it is available in my market.
It certainly represents a whole lot more than what exists on the market right now.
I certainly like the idea of a visual alert notifying of cpu/memory hogging processes. With the N9 as a reference,the app in question could glow or have a red tint in the running apps screen. Having a "top/htop" balloon over the individual open apps in the running apps screen would be good, adding a "nice" button to click on when long pressing each app would be good too.
This is the sort of thing that could be enabled in a "developer mode", leaving a minimalist view for the average user and yet still being available for power users.
Given how the swipe UI is so intuitive, I look forward to what the Jolla team can deliver, whether it's a completely different UI or just building on the concept.

Patents are such a mess - you never know what kind of ridiculously simple thing can be patented. So when you create something new - just ignore all this nonsense. Don't let your creativity be limited by this idiocy. Debian patents FAQ even recommends not to research any existing patents.

Personally I don't care that much if something is open or closed as long as it works as expected. Jolla wants to be a comercially successful company so they can't rely on thousand different opinions, whishes and desires on how things should look and work. That just wouldn't work.
So if the tipical lifespan of a device is 2 years then the company should properly support it for those 2 years and not put "won't fix" next to the multiple times reported bug. Or if some feature is missing then the company should implement that in its own (open or closed) already existing program. I don't see the point in that some community members spend weeks or months in recreating music player just to add equilizer if the company can do that much faster.
There should be an open and honest communication channel between community and company, a place where we can ask a question and get an answer on why something can or can't happen.

But I do agree that a device should belong to the user as much as possible.

There is a Ted Talk on "how I beat a patent troll."
There has to be some judge somewhere that understands that the rate of acceleration of technology leaps causes old lame patents nowadays to seem relevant even though they are not.

I don't get why anyone would want that personally, but sure, see osx86project.org.

I haven't read that FAQ, but I can bet that the reasoning behind that is the same as the reason that I've heard around many times before: if you do patent research, and create something which infringes on a patent somehow, you're now wilfully infringing on it, which means you're now liable for a whole lot more damages than before.