Firstly, unlike backend/system/OS/desktop components where lots of people can write little bits and bring them together to form a whole, you have to consider the whole UI from the start and create a consistent interaction model and look and feel, probably with some visionary type figure wearing a nice scarf who is leading a small team of trendy designers and UX people. These kinds of people tend to be relatively rare in OSS, as undertaking initial design and then supervising a whole project through to "completion" is a sizeable undertaking and not really "scratching an itch" as many OSS projects start off as.

Secondly, the UI is where all the components come together. If you're making something that uses components A, B & C and when you use them all together there's some funny race condition or you don't quite have all the right APIs to make it all work together in the way that was envisaged, you're the one doing the sometimes thankless task of delving into all the bits and seeing what's broken where and how you can fix it.

Also you're the one looking at performance and final testing of the whole system - UI, middleware and firmware together. Two apps open at the same time cause too many wakeups and kill the battery? You need to instrument and fix it. Using the browser while crossing an international border and starting roaming kills the data connection until reboot? Sucks to be you - fix it.

Yes it's probably easy to lash together some kind of UI which allows you to "use" various features you'd expect from a mobile. It's very hard to make a "product" which isn't a complete inconsistent mess full of weird edge-cases that might be somewhere approaching usable as a daily driver or sell-able as a product.

It's taken Jolla, what, 3 years to get where we are now? And most of their staff are ex-Nokia so will have done and launched a similarly complex UI before on the N900 and N9. A good-quality UI for a consumer product takes a surprising amount of effort.

And as we've been discussing here, until you have something "nice" that people enjoy using, you're very unlikely to reach that critical mass where some of your users are developers too and they see value in jumping on board and adding stuff (in fitting with the OS design language too)

Nicely said!

thanks for explaination..it seems reasonable. and what about buying sailfish closed source components from investors? how much it could cost?

however for glacier it seems there are already nice concept and vision.of the ui,so the leading small team maybe is already there

Uh-huh. You want a "trendy" UI. You want a "consistent" UI. Those things are better provided, in my opinion, from closed-source entities (where you will find the visionary types wearing the nice scarves). And those things are already quite well provided by Apple and Google today.

I don't personally see the need for a gorgeous, massive UI aimed towards lowest-common-denominator users. You might be able to make money that way; but I don't think you'll be perfectly serving any specific users that way.

Target the UI specifically towards the needs of the folks who are building the UI! At least that way, it'll have a chance to become self-sustaining...

Again, who cares about making this thing usable for the average cellphone user? If there's no reward in fixing a particular problem in an open-source program, then congratulations, you've just discovered that your problem isn't sufficiently important. Sorry that you've got a corner case that nobody cares about, but unless you can fix it yourself or pay someone to do it, it's beyond the scope of the project.

If a mobile OS would be sell-able as a product just because it is consistent and usable as a daily driver, Jolla would be raking in the dollars right now. Sorry to say this, but iOS and Android are already "good enough" in that department.

So yeah, if you want an alternative to iOS and Android, then you probably don't see consistency and user experience as the most important features in an OS (or, at least, you're much more interested in some other aspect). As such, I say it would be fine to have a UI that (at least to begin with) is incomplete, inconsistent, and glitchy. A minimalist UI that gets you off the ground and lets you use mobile hardware would be a great start...

I dunno. Everyone on this board seems to have conflicting views about what is "nice" to use. Personally, I'm not a big fan of graphical UIs to begin with; the more minimalist the interface, the better I like it.

I don't really see why a UI has to have one single way of doing things, either. Apple, Google, and Jolla all try to limit their mobile UIs to do things in one particular way, and thus constrain both users and app writers into following their particular system. Perhaps this helps the ordinary user avoid having to learn lots of different mechanisms for accessing data; but it also makes the device very inflexible.

Instead, better to focus on providing a minimal set of functionality, and then allow the user to choose how they want to access that functionality, be it by swiping, pressing buttons, pulling down menus, or whatever else they would prefer. Don't bother creating a massive, full-featured UI; just give me the parts I need, and let me put it together myself.

copernicus i agree for what regards ui and masses, but not for consistency.
if you have a ui without any direction or any recalls to anything, you will get a lot of fragmentation. if you dont choose anything, cause you want it all,you will have a bunch of flies in your hand and nothing more. of course its better have a way to personalize, but not to destroy all the directiob that characterize that os. for example, if i think a system with pulley menu,swipe, but also button everywhere like android ...i wouldnt like it. for me is like wearing an elegant jacket with sportive short pants, and maybe with white socket with sandals (please let then to guys from Germany/just joking)

You already have the answer:
On one side, you say you need millions to do it, on the other side you want a bunch of TMOers to do it by themselves on their free time.
Every big OSS project that succeed was backed by a company (Red Hat, SUN, Trolltech/Nokia/Digia, ...).
If that backing company is not Sailfish, you have to find another one.

Or maybe the missing part is less work than I think and 10 people can do it in their free time ?

And about the consistency : it is hugely needed as it is directly correlated to usability.
We have seen a lot of inconsistent interfaces, and they are a pain to use (Windows 8, Android, ...), compared to the consistent ones (Gnome, Meego, Sailfish), even if at the end you can do your tasks on both.

AFAIK Blender is not backed by a company, but I agree one single exception doesn't make a rule.

Blender *was* backed by a company, and is not really an exception in this regard

Yes you're right. And this explains why it's probably a better idea, in case of Jolla bankruptcy, to continue with SFOS (of course, it should be fully open in such a case), which is pretty mature, than starting a brand new GUI.

mee want swipey swipey

is that really such a hard requirement in a mobile UI.