Mer-based OS on a Tablet: what would you like to see?
 

In this thread, I would like to collect some ideas about what would comprise a good, "usable" open operating system running on a tablet.

Here's my situation: I'd originally planned on diving into writing apps for the Jolla tablet, which I'd hoped to have received last summer. But, of course, it's now winter, and given current circumstances, I'm not sure I'll ever receive the tablet. So, I'm trying a new tack: I'm now planning on picking up a cheap Nexus 7 tablet, and installing something other than Android on it. If this requires me to do some UI coding myself, so be it; I might as well spend my energies on that, if Sailfish is going nowhere fast...

So, here are some questions to consider:

1) Use Mer or not? I think there are a number of advantages with Mer, but given that the initial target here is a tablet, maybe a full-blown desktop distribution of Linux would be doable...

2) Assuming Mer is chosen, what UI to use:

a) Sailfish? I'm kinda hesitant about even touching Sailfish at this point, as it may soon end up entangled in Jolla's death-throes, and is in any case way too big for just a handful of open-source advocates to maintain. But it really is a beautiful UI...

b) Nemo? Support for this UI seems to have dried up, but it is cleanly open today, and seems more lightweight than Sailfish. I want to at least give it a try.

c) Hildon/Cordia/etc.? It'd be really nice to see a Maemo-ish interface come back again. (And it includes stuff that I really like: extremely light-weight UI elements, lots of support for alternate input methods -- such as stylus and keyboard, support for widgets, etc.)

3) Given a particular OS and UI, what features are essential? I myself demand a decent terminal and shell. ;) But my guess would be that tools like a web browser and e-mail client would be significant for most users. Media players as well. Again, I'm not a tablet user, so I don't have a good feel for what people do with these things...

4)The fun question: what could we do differently? For example: one thing that absolutely every phone OS designer has chosen to do is to kill tasks whenever a memory overrun is threatened. This is, of course, to ensure that you can always use the device to send and receive phone calls. Without the need to protect the phone functionality, a tablet UI could reasonably return to standard memory paging techniques to bring true multitasking back to mobile devices...

Anyway, let me throw the floor open for comments here. If you had the chance to put together your own tablet with your own OS, what would you want to see? :)

EDIT: I don't think I put enough of an emphasis on the idea that this is a question about what you, yourself would like to see in a mobile OS, not what you think would sell as an OS for the general market. :) So, I've modified the thread title a bit...

Re: Mer-based OS on a Tablet: what is needed?
 

I fear that engaging this lot will end up in the same mess that plagued Jolla. The too many cooks syndrome. But what the heck, here's my tuppence.

For starters, I think a tablet is much closer to the desktop than to a phone. I think the major mistake all tablet OS producers make is to use the same UI on the tablet as they do on the phone. So perhaps the best starting point would be to look around the existing desktop environments and seeing if any of them could be suitable on a keyboardless touchscreen device with minimal changes.

The second thing I would suggest naturally comes from the first anyway but I will spell it out just in case. Focus on function, not the form. No facy animations that slow down the work flow. No unnecessary transparencies. Really count the number of taps (swipes, key presses...) it takes to do each task.

Case study: the N900 keyboard. I know, I said keyboardless. This is just for illustration. I do not know about other language variants but the English keyboard has the dot in the main layout. No shifts, no symbols. So far so good. BUT - consider a text entry box that accepts only numbers. The numbers on the N900 keyboard are on a symbol shift, which is helpfully pre-selected for you when such an entry box receives a focus. The trouble is, the dot (in this case, the decimal point) now requires pressing the sym shift key. Annoying as hell. Good old Palm did that just right: the dot was a dot without a shift in both the letter and the number mode. Why? Because they cared about workflow down to the minute details like that.

Re: Mer-based OS on a Tablet: what is needed?
 

Aha, now that brings up something I recall from long ago; back when screen space was at a real premium, I tried playing around with some tiling window managers. The main advantage of these is that you can put pretty much every single pixel of the display to use; no wasting of space on drop-down shadows, minimal window border decorations, etc. Plus, you'd get Dave's favorite split-screen functionality for free, only you'd be able to split the screen into as many pieces as you'd like...

http://i3wm.org/screenshots/i3-9.png

Ah, well, to do that you really need to understand the workflow down to the minute details. I would know that for some things, not for others. I would think the only way to really accomplish this goal would be to make it easy both for users to explain their workflow needs in precise detail, and to make it easy to adapt the UI to those needs. Making the UI open will help a lot for the second task; not sure how to manage the first task, though...

Re: Mer-based OS on a Tablet: what is needed?
 

I'll just note that there is the Fedlet project that aims to bring Fedora to Intel Bay Trail based tablets.

Re: Mer-based OS on a Tablet: what is needed?
 

Very cool! Any chance of this getting anything like official support from RedHat? (And, I guess, given that the Fedlet page says that it is now "Semi-Dormant", is there a future at all for Fedlet?)

Re: Mer-based OS on a Tablet: what is needed?
 

I'm all for mer based os running...
but for gawds sake make it a rr. that way it doesn't become a nightmare in upkeep.
I love arm and all...
but if you have the serious hardware...use it.
Base it off Debian...
make it an rr ...
throw, mer into the mix,
and then you have a bloody unstoppable tank of a distro ..
clean, spartan and not top heavy?
antix is a good debian rr example.
absolutely fab.
runs on a bloody toaster..
old hardware or new hardware...
looks beautiful, runs beautiful...
I don't suggest making any distro that can't be applied to as many devices as possible...nor would I suggest a distro which has a lifespan of 6 months or 2 to 4 years tops...with nothing but headache along the way...
I suggest another way after seeing (pretty much) a decade of the same problems attempted to be solved using the same solutions over and over...with the same result.
Inevitably I believe that making mer or fremantle or hildon or diablo ...essentially a maem-esque distro properly laptop/tablet compatible is the way to transcend the limits of the initial devices they were built for...
which is what solution we should (and seem to ) be running toward.

I think that making things more complicated than necessary esp concerning onscreen keyboards...welllll...

I do think utilizing existing (very plentiful) brand model devices which are as open and ...um hm...receptive...to linux is important..
the nexus line is a good example.

in particular..some device brands appear to work well until looking under the hood and there is something ugly like broadcom wifi chip or some other component that doesn't play well.

As far as tablets go...
I would seriously suggest a convertible hybrid which has a keyboard.
It isn't that I'm biased to my fav. and I'm extolling it's virtues due to a brand loyalty.
If any other product did what my hybrid does and can do...I'd happily be saying the same of it...
but my fav became my fav because of everything it contained, how well it's internal bits got along with one another and how well it got along with linux distros compared to sooo many laptops, hybrids and tablets I have used in the last 2 decades...
I've used uncomfortable desktops, netbooks, clunky and sleek laptops, antiquated (now) umpc's, the early tablets, recent tablets too, with and without attempted physik keyboards ...and hybrids...hybrids with docks that have the keyboard, hybrids with docks that have the optical bay...etc...
best model I've used in my time is the toshiba portege m700 m750 and m780 (i7 is my latest and last model they made)...everything in the brute.
the stylus is nice...it's great to use it as a proper laptop..it's great to flip the screen around and use it as a pure tablet....it gets the job done..it works with just about every flavour of linux I have thrown at it..
I rarely use the optical drive ....if ever...
8 gb of ram,
a tb hard drive internally ...purely for easy casual storage...
and I have a second hard drive caddy for the optical bay which I pop in and out as the desire suits me to play with any of 4 or 5 hard drives with various distros ( I don't know how many I'm currently playing with...I've lost count..) I shove in and spark it up for...
So in all...
I think It would also be a contender device line I think...
they are plentiful and cheap enough...

But no matter what decisions are come to (if any are actually to materialize...) concerning the next steps with distros and such...
I do hope sanity is a primary ingredient when considering the end result and what legacy it will have...if any...

Re: Mer-based OS on a Tablet: what is needed?
 

Note that Fedlet is Adam Wiliamsons own hobby project. But its is true that the Gnome developers do work on making Gnome shell more touch friendly, even though this is probably targeted mostly on touchscreen laptops rather than on "true" mobile devices.

As for Fedora, it already works on various arm boards, but given the condition that the target device must be able to run mainline kernel then I'm afraid we would not see it officially on many mobile devices any time soon due to the overall kernel forking madness.

Also I don think anyone has used libhybris with Fedora so far, but even when he did I'm not sure this would be supportable as an official Fedora project deliverable. But I could imagine a libhybris using Fedora spin/remix - if enough people from the Fedora community would be interested working on it.

Never underestimate AdamW, he is an one man IT army. ;-)

Re: Mer-based OS on a Tablet: what is needed?
 

is it just me or does gnome 3 not remind us all of Hildon?

Re: Mer-based OS on a Tablet: what is needed?
 

I still think libhybris is a game changer for supporting other devices. I'm seriously thinking about a pi2 with touchscreen at the moment for testing my *little* side project, but if I can get it to a reasonable point I will be looking more at libhybris for other devices. if a decent, open intel x3/x5 based tablet with intel supported gpu driver comes along...

A tiling WM could work quite well if allowed to do full and horiz/vert split. Just needs a little UI polish compared to some existing ones.

As some of you know I've been experimenting with hildon/gtk3. At the moment it is still running on a matchbox2 based hildon-desktop. One idea I had as a possible future direction is to take inspiration from Harmattan and Budgie and use mutter.

I have a very small test app that literaaly just gets mutter running but has no ui elements (taskbar, switcher etc) using Budgie as a reference. Obviously there are a few gnome dependencies, as there are with Fremantle, but mutter could make a good start for you.

Alternatively, my other option should GTK3 prove too much of a hassle was to dump everything gtk and follow the Maemo>Meego path and use qt+qml. Down this route I was thinking what about expanding the qmantle2 demo (if you can find the source code, might have to contact rburchell directly). It uses lipstick iirc so you get a qt5 wayland based ui. I did have the source code but have misplaced it or deleted it by accident. You could mix in either stock qtcomponents or create a custom set based on marxians hildon versions.

anyway, thought I'd share some of my ideas wih you. also just reminded me i still haven't pushed mb2/gtk3 to github. one day i'll remember..

Re: Mer-based OS on a Tablet: what is needed?
 

If I remember right, the guy, who made campaign of tablets with QML based DE long time ago on TMO and as far as i remember(I might be wrong) was involved in mer or it was based on mer, now have few interesting projects :)
Qt(QML)+Wayland and all this have pretty nice UI already. It's all opensource and on github (btw maybe worth to open new thread for that one). anyway here are projects:
http://www.maui-project.org/
http://hawaiios.org/