The new QWERTY device project
 

Now that we are getting closer to finish the Moto Keyboard Mod, I have started to work on the next excitement - reviving the N950/Lauta concept with a modern SoC and hardware.

The first choice has been made: SoC will be Qualcomm 64-bit. There are a few choices but Qualcomm's is the best in overall performance and support.

Qualcomm (Chosen): Solution extremely expensive. Components are hard to get. But support is great (in terms of drivers). Performance at top tier, power management is great.
Intel: They exit mobile chipset business and atom range is not as power efficiency as Qualcomm.
MTK: Cheaper than Qualcomm, performance is OK. But drivers often being the issue. Components often out of stock...
Allwinner: Lower end cheap SoC with unknown support other than Android
Rockchip: Strength on Media but not a mainstream chipset.

Then here comes the question: Which OS we can use? I have a few in mind but the community's opinion will play a great rule.

AOSP: Can be shipped as a default choice, for warranty purpose, at least every basic function is working. But as Wicket mentioned in here, it seems to be causing barriers porting proper GNU releases?

The advantage for AOSP is mainly cost-free and lawsuit free, I can't sell something under other OS that might have broken functions...

Sailfish OS: A proper OS that I want to ship with default OS. But to obtain official adaptation costs a lot of time and money. On the other hand, we can leave it to a community port only. The question will be is it worth the cost adding Android layer.

Nemo: Any opinion on advantage/disadvantage on that? On official wiki page looks like the development ceased 4 years ago... If this can be achieved in terms of open UI and community/developer friendly in a latest release/kernel that will be a great choice.

Ubuntu Mobile: Is it cancelled? Is it worth developing?

GMS Android: To make this project successful and have funds improving HW/SW in the future, selling it to general public in a GMS version seem to be a good choice. Although I hated Android...

Lineage OS: As an alternative to GMS Andoid, but is it possible shipping as the default OS?

Any suggestions/ideas/loves/hates please post here. ;)

==================================

Update:

Chipsets and specs has been decided.

At the moment we will do our best releasing the device with Sailfish OS, while keep an Android version (Either Lineage OS or Google GMS) for general users. But SFOS will remain 1st priority.

=================================

Update 2:

The project is moving forward but since I am no longer seeking for crowdfunding, I will share progress only when major milestone has been made. This will be a great one. ;)

Re: The choice of OS of the new QWERTY device project
 

I vote any OS except sailfish until Jolla step up and take responsibility for the failed campaign. That said if they do that with a full refund I think sailfish can be an option. What about android? I don't think we shall exclude any option this early in the game.

If we got out second refunds we could donate it to this project and maybe you could afford a basic prototyping.

Happy to see this thread! It's very welcome here. GL chen.

Re: The choice of OS of the new QWERTY device project
 

well in my view your target market has two clear groups within it
1, those who just want a hardware keyboard for their phone
2, those who want real linux on a phone with a hardware keyboard

group 1 ( in my view the larger group ) would prefer something normal like AOSP ( or Lineage OS? ).

group 2 ( the group i am in ) can cope with messing with system stuff to get whatever they need on it, as long as it is setup to be easily done (what defines this needs its own thread *sigh* ).

so i my view AOSP (or Lineage OS) is a good choice to appeal to the largest group, giving us the largest chance of this device going into production at a reasonable price, while providing what support you can to allow users to install themselves what they like on their own device.

Re: The choice of OS of the new QWERTY device project
 

Aaaaand... Here we go again. Dave999 barging in and spoiling Yet Another Thread™ with his drivel about Jolla's refund.

Chen, I would like to ask, why not a proper Debian or any other mainstream Linux that has an ARM support? If you are in control of the drivers and your device has a real keyboard then I can see no reason why not.

Re: The choice of OS of the new QWERTY device project
 

Nemo is very far from being appropriate for daily use. There are simply too few developers that have time to work on it (I think there's three people working on it off and on), so over the past years there hasn't been a lot of progress.

Out of moral concerns I definitely wouldn't go for GMS Android, and practically it wouldn't make much of a difference either. Someone who wants a hardware keyboard is more likely to be a 'power user', and they would either not want Google, or have no problem installing those packages by themselves to get their desired functionality.

I agree with chainsawbike and think Lineage OS would be best. AOSP is very out of date, the music player hasn't seen any update since Android 2.3, when Google started exclusively developing their proprietary version, and it's like this for nearly all of the default applications. They're even worse than Jolla's applications that are never updated.

Re: The choice of OS of the new QWERTY device project
 

How about dualboot between Android/Sailfish? For me Sailfish is enough for daily use. Perhaps sometimes Android app is needed.

How about good old Nokia Communicator look (9500/9300..)? The main screen is in safe when closed. Small screen to see clock/number etc.

Be sure, that there is enough memory and a good camera! Dual-Sim!

Re: The choice of OS of the new QWERTY device project
 

How about a voting system? I would prefer a N950 like design with Sailfish on it

Re: The choice of OS of the new QWERTY device project
 

while that would be very convenient for those who want a real linux on their phone i do no believe that will provide a good enough out-of-the-box user experience (or just plain too different) to be mass-marketable and in my view we need the device to be mass-marketable to make it viable

as long it is made with the intent of being able to install real Linux on it, I think releasing it out of the box with an android variant would be a smart move to attract a larger number of potential buyers

Re: The choice of OS of the new QWERTY device project
 

You want this to be commercially successful or at least not a flop so android OTB is the only thing that make sense.

Re: The choice of OS of the new QWERTY device project
 

Well, to get this really difficult project to a reality I think we need all kinds of options and supporters. Not only yours or mine. In addition to that we probably need a lot more backers than tablet project to even get this to the starting line.

If you can't accept others opions I think it's better you leave the union :)

This thread will be awesome!

Re: The choice of OS of the new QWERTY device project
 

You are taking a (calculated) risk with this project already and it would make sense to ship with the most risk-free solution: GMS Android. That way you could differentiate yourself from the rest via hardware and appeal to the maximal number of customers. I would expect that many are missing hardware keyboard, better get this out for as many as possible.

Would be great to make this device as developer friendly as possible. I don't know how compatible it is with the GMS, but making bootloader unlockable as simply as on Nexus would be great. Its another problematic area right now and making it appeal to developer community would surely help you.

I wonder whether its possible to ship with TWRP out of the box. If it is, why do others keep their crippled recoveries ...

I wonder if you could consider keeping the warranty after device was re-flashed, in terms of hardware failure. Again, its an option that would appeal to many developers.

As for Android base, Lineage would be great. Porting SFOS to it would be simpler if it has Lineage base, at least it used to be in CM days. However, I don't know whether Lineage folks make it possible to preload it on the device.

I presume that you want to be ready ASAP since all components get older every day. So, if Lineage takes too long time, ship a working product as soon as you can - being AOSP + GMS.

With the product available for developers, there will be people porting Lineage to it anyway. Maybe it would be great to get in touch with their community and ask if there are any specific requirements that can make their work easier.

I presume many in this community would be very interested in getting such device with SFOS. I do wonder if we can get SFOS images for extra fee, similar to Sony Xperia X program, and how large would the fee be. With Lineage ported/shipped it should be possible to make community port as well. When the choice of chipset is done, maybe its great to check whether it would be problematic with Jolla and/or porters.

Few additional pointers: please include SD Card slot. If possible, please use AMOLED screen.

Good luck!

PS: Ubuntu Mobile is discontinued and there is only a fan-based UBports

Re: The choice of OS of the new QWERTY device project
 

To this users groups distinctions, there are also two ways to use the device:
* keyboard closed : a touch friendly OS is needed
* keyboard open : a landscape friendly OS is needed, with good support for keyboard shortcuts

Sailfish OS is one of the best touch friendly OS in my mind, as it has been design with this in mind, whereas android is usable but can be a pain (small buttons, things in top and low bar, and all), and debian would need a UI for touch (KDE Plasma ? Hildon ?) and phone (with calls, sms, contacts and apps like that) that I am not sure exists.

Sailfish OS is not the most landscape friendly OS, for example the home screen or events views don't exist in landscape. If this phone is a success it may help to push them to add it. But for people who used an other-half hardware keyboard, it looks like a lot can be done, especially on the shortcuts side.
Android is a bit more landscape friendly, but I don't think the keyboard is a first class citizen. On my "old" Motorola PhotonQ, it is usable, but there are no shortcuts that I know to switch between apps, most buttons can't be reached by the keyboard, so that you constantly switch between the touchscreen and the keyboard.
Debian is obviously landscape friendly, but the chosen UI would have an impact to the usability too.


And finally, the OS will have an impact on the apps we can run:
* Sailfish : can run Sailfish apps, and can also run some standard linux app with some work. For android apps, Jolla's support is needed (or an open source emulator like AnBox or fDroid is finished), and Google Service is not available legally.
* Android : can run Android apps, and some linux apps with some work (chroot ?). Google Services are needed for some mainstream apps. Sailfish apps cannot run without a lot of work and maybe legal issues (Silica ?)
* Debian : can run linux apps, maybe Android apps through emulators (AnBox, fDroid). Google Service is also not available legally, and Sailfish apps would probably not run without a lot of work.


So I would prefer Sailfish, but for the success of the phone I understand that an Android flavour would help getting it to the masses, and that a community port of Sailfish would be enough (especially if you can help with as much info and sources on the drivers used).

Re: The choice of OS of the new QWERTY device project
 

Unfortunately, I agree.
For example "sticky keys" function (when I don't need to push Shift + other key nut once but I can click Shift and then other key, of course using both keys at once also do work) is very useable for me on Motorola Droid 4. Sailfish lacks this function.

If you remap keyboard files located in /system/usr/key*/ directories you can use ALT+TAB to switch between apps and CTRL+C/X/V/A on text editing.
I have remapped Caps Lock key to CTRL and ALT so now I can easily type using it Polish characters (ą,ć,ę, ...), use ALT+TAB (using one finger as the keys are located next to each other) and mentioned CTRL+C/X/V/A.

Re: The choice of OS of the new QWERTY device project
 

What's the cost for shipping a device with sailfish/android? The license cost taken by Jolla and google?

Re: The choice of OS of the new QWERTY device project
 

First define a 'commercial success'. How many units need to be sold for what price? I think it is already a success if the product is launched and Chen and company will make some profit.

Don't forget that the Android crowd is a lot less tolerant about hardware or software issues. You need to launch with an almost perfect product and can't ask for a premium price if it is not.
People using debian are used to solve their own problems, looking for workarounds and will actively assist improving the OS.

Re: The choice of OS of the new QWERTY device project
 

Jolla should be bending over backwards to get involved with this project.

Re: The choice of OS of the new QWERTY device project
 

Hi Chen,

First of all, congratulations on your successful tablet campaign! I hope my post that you linked to didn't come across too negative. I'm glad you've created this thread to seek input from the whole community.

I'm less concerned about what OS is bundled with the device and more concerned about what can be installed afterwards, and for that, the choice of SoC is critical. For me the most important thing would be that you choose a SoC which already has decent mainline Linux support. The main point being that we are not forced to use an Android kernel because nothing else is available. Android kernels generally go unsupported after the SoC vendor has stopped manufacturing each chip, thus the device would be forced into obsolescence after maybe only a couple of years.

You say you've already chosen Qualcomm (which isn't necessarily a bad thing depending on your target audience) but I've a few comments on your summary of SoCs which you may want to take into consideration.

Intel have indeed exited the mobile chipset business however China-based mobile chip vendor Spreadtrum have partnered with Intel and have been making new x86-based SoCs for smartphones.

Allwinner is quite the opposite of what you describe. They are actually one of the best vendors for non-Android support and use Mali GPUs which are compatible with the Lima free/open drivers.

You didn't mention OMAP, perhaps because TI are no longer developing new chips, but OMAP has great mainline Linux support. The main disadvantage with OMAP is the PowerVR GPU for which there are no free/open drivers.

Qualcomm has traditionally been a bad choice for privacy reasons. Their SoCs are known to have bad hardware design where the modem controls GPS, audio (including mic), RAM, NAND. If you choose a Qualcomm SoC without a built-in modem, it might be okay. Mainline Linux support for certain Qualcomm SoCs has improved in recent years and Freedreno (along with Lima) is one of the more advanced projects for free/open GPU drivers.

Have a look at my thread on mainline Linux devices which provides some info on which SoCs are currently best supported by mainline Linux. There are currently two 64-bit Qualcomm phones with some support: the LG Nexus 5X (MSM8992) and the Huawei Nexus 6P (MSM8994). Other 64-bit Qualcomm chips with mainline support can be found here.

Re: The choice of OS of the new QWERTY device project
 

Whatever the OS is, I think people here on TMO need to be absolutely certain that it will be possible to install something else and to make it (1) work and (2) useable for an everyday carry device. Although an Android flavor would certainly be best to get media coverage during the campaign and therefore more backers coming from many communities or no community at all (as opposed as all people here on niche-TMO), it would be a big disappointment if it turns out that the device that was first discussed here cannot comply with what people here would want. I myself would not want to use it if it came only with Android as a really working solution (I'd buy a Moto with your keyboard otherwise).

Sailfish is fine to me, it already has proven being capable of proper keyboard support with the daemon that came with the TOHKBD, although it would of course need some work to be adapted here. The OS itself is also very good for daily use and has active development with a company working on it, regardless of how much some people may complain about that company (and those people are still on TMO so they're not such haters either), it managed to make one of the only working alternatives to iOS and Android and it is a serious one, not something that disappeared after a few months (although there have been hard times). It is not perfect in every way, but none of the other OSes are and Sailfish is not a bad balance between privacy, openness and closed blob, although it does have some closed blob.

Sailfish can also come with Dalvik layer and I really hope you can set a deal with Jolla for an official Sailfish experience. Community ports are great, but there are far less users of community ports than official Sailfish. There are likely two reasons for that: (1) community ports require some hacking, (2) community ports can't run Android applications. I believe (2) is far more significant in the reduction of the number of potential users than is (1). Doesn't matter how schizophrenic it can be to want to run Android applications in an OS that is all but Android, we don't really get to chose if we want to use everyday life tools. Banking systems require Android applications and they do not all provide website alternatives, connected cameras require Android applications, town-related applications (parking, bus, etc.) all require Android, and I happen to live in a town where I am not allowed to park my car anywhere if I cannot use the application or buy a yearly subscription, believe it or not. We are already screwed. But running those applications on demand within an OS that better respects our privacy would still be better than running Android permanently and have to suffer the poor ergonomy (buttons on the top right and top left of the screen, really? Option enabled by default to send excerpts of what we type with the virtual keyboard to Google, really?).

I think official Sailfish would be the best for the trade-offs between everyday use, privacy, ergonomy, security (actively maintained). I don't really care if the phone ships with that OS out of the box (although it would send a strong message about what this device is for I guess) as long as I know for sure I can use it if I install it myself. The thing is, I'm a bit concerned there would be any guarantee to that if the phone was shipped with another android-based OS that was deemed acceptable, despite the Android ergonomy and paradigms. I would not mind paying extra for official Sailfish to help set the deal.

I must admit I would love being able to run a real Linux distribution on the phone too and I would have many uses for it, but this probably should not be the default OS provided with a smartphone. Also, although it does not satisfy the above rant about the need for Android compatibility layer, what about other OSes mentioned in this thread, like Plasma Mobile that seems to be actively worked on? Aside from the nice UI which is already an advantage, here is what they say about applications: "Plasma apps, Ubuntu Touch (.click) apps, Gnome apps (e.g.GnomeChess), X11 (e.g.xmame) and possibly other Qt-based apps like Sailfish OS or Nemo. Packages can be installed by “apt-get install packagename".

Good luck Chen and thank you for your continued efforts.

This may also be something important to keep in mind. The bad reviews that could come from these people could ruin a campaign and reputation of a device, although the device might be up to the standard of a community more resilient to fiddling.

Re: The choice of OS of the new QWERTY device project
 

I feel sailfish is close enough to a desktop experience that it covers the need for mobile phone users and desktop users.

Meanwhile, I think a voting system would help to see where the community is at. We may already agree on most aspects ...

Re: The choice of OS of the new QWERTY device project
 

In my opinion, the voting system (if it will be enabled) should be forum independent so that people from other communities might participate.

For example from XDA-Developers - I suppose also there are people interested in physical keyboard devices (according to Droid 4 and Photon Q device sections activity).

But it depends on the final target group for the device.

Re: The choice of OS of the new QWERTY device project
 

What ever OS is chosen, I hope that it is self contained and capable to the point I can write, compile and run an app on the device. That is, use vim or other standard linux editors and a "make" command to compile.

Re: The choice of OS of the new QWERTY device project
 

While that is great and agree with that. I want the device to come with the most user friendly OS available so I can recommend the device to more people here at TMO and in real life. And also get an easier time to find potential backers. While a geek device is cool and all, we are not enough to make it happen.

That means android or sailfish. While I don't like it currently, it's basically the only path currently. But sailfish has a long way to be able to handle a landscape qwerty slider.

And to be honest. The hardware is much more difficult issue. Where to find a working slider mechanism?

Re: The choice of OS of the new QWERTY device project
 

What is the actual support to this statement? Have you tried using Sailfish on the first Jolla smartphone coupled with the TOHKBD? It was actually as close as it gets to a N900 experience in my opinion, except from the general bulkiness, the concern of dropping the phone when it just holds with magnets, the fact that I still preferred the feeling of the keys on the N900, and the lack of stylus (but admittedly none of these limits were related to the OS).

As for the slider mechanism, Chen knows where he's headed according to the details he gave before (close to the mechanism already seen on an HTC device, with tilt adjustment and not just a slider). Not saying it will be easy, but at least he knows exactly what difficulties he will face and probably has a plan on how to circumvent them.

Re: The choice of OS of the new QWERTY device project
 

Personally i prefer Harmattan over Maemo/SailfishOS/Android, so maybe part of the campaign's funds should be spent on hiring some blackhat to steal sources from Nokia ;-)
]
It's good to hear about n950/lauta design. Keep this design in pair with quality(titanium body[not alu] or polycarbonate[better signal handling] and it will be success. Like others already mentioned add SDCard port, Amoled
screen, 3.5mm jack and maybe 2xUSB-C OTG ports and I will happily replace my n9 with it :-)

Re: The choice of OS of the new QWERTY device project
 

I dunno Dave999. Folks in "real life" already have enough options. This one should be just for us and, if it's good enough, more folks will come on board with time.

I do like the idea of Harmattan...

Re: The choice of OS of the new QWERTY device project
 

Basically from using jolla1 in landscape. Try that. Start with unlock your device with a code...

Nerver tried HTC slider but to me that sounds, tricky and expensive. But yeah. Good IF he has an Idea that can be relizied.


So you suggest we should not push it to real life but only as a ghost in this forum, like fiction of our collective imagination :D

If this one is for us only 300-400 supporters. Then we have to be ready to pay Approximately 2000 Euro each :)

Re: The choice of OS of the new QWERTY device project
 

you could go with Android (whichever version AOSP, Lineage, etc.) and then talk to Jolla about a system like the Xperia X, if they'll go for it.

Out of interest which SoC is it?, Snapdragon 430, 600 series?

Also what screen is it, AMOLED, LCD? In either case can you make sure it has Panel Self Refresh, for better power efficiency. I *think* the JDI (LCD) displays might have that by default these days, but not sure.

Re: The choice of OS of the new QWERTY device project
 

Gotta pay to play!

Re: The choice of OS of the new QWERTY device project
 

I thought so too. But then I met Jolla and apperently they don't. If they would have been a one man compony, fine. But the where a multimillion company with 100 plus stuff...:eek:

Re: The choice of OS of the new QWERTY device project
 

But how relevant is it? You should probably not base all your assumptions about the capability of Sailfish handling a landscape keyboard slider... On your experience using Sailfish in landscape without the keyboard.

The tohkbd really helped a lot in all ways, from just handling the device without an awkward uncomfortable grip (since you don't risk touching or obstructing the screen by mistake) to general use with the key input. Things like Alt+Tab or ability to type codes with the hardware keys really change the experience, and Dirkvl and Kimmoli managed to have them very well integrated in Sailfish. Heck, Sailfish even started to support hardware keyboards in the Input section of the settings in a later update.

From my experience with the Jolla+tohkbd, I can tell Sailfish is not bad at all for handling a landscape keyboard slider (I don't even see how Android would do better), and it does not cripple it either. You can still use it in portrait mode with the touchscreen without any sacrifice (which was not the case in Maemo 5, despite all the good things I think about Maemo 5, it did not shine by its portrait mode capabilities). I think this is what the Lauta is about: a landscape keyboard slider with a modern OS (not something limited to hackers, but compatible with them) and still an everyday "smartphone".

About Harmattan, I'm not sure it's an option. Is it fully open and maintained? I loved Harmattan (except the Aegis blob), but even if it could be ported, we don't want something that is not maintained anymore and will bring tons of security flaws or be incompatible with newer protocols and applications.

Re: The choice of OS of the new QWERTY device project
 

I thing you read way to much into this discussion.This what I wrote. Nothing more nothing less.

"But sailfish has a long way to be able to handle a landscape qwerty slider."

Not sure what you are on about ;)

Re: The choice of OS of the new QWERTY device project
 

I'll give that to you, I probably shouldn't have gone into such a lengthy message. :D Sorry about that, nothing personal anyway and I meant no offense.

Anyway, long story short, I personally disagree with the statement that Sailfish has a long way to be able to handle a landscape qwerty slider, as a user who has actually used Sailfish with such a keyboard for months. But since this opinion is totally personal, I wanted to know what supported yours. Since you seem not to have used a keyboard with Sailfish (but just landscape mode), I'm a little skeptical about this statement that Sailfish would presumably be less potent than another OS for this.

Re: The choice of OS of the new QWERTY device project
 

That's not true. Sailfish support even sticky shift + alt combination.

Re: The choice of OS of the new QWERTY device project
 

Kabouik and Wicket, thank you for both of your posts, I wish I could thank you twice! Lots of good points in there.

Chen, I think a lot of this depends on the price point.

If the price point is high, the device has to last a really long time and therefore relying on an Android kernel that will become outdated/unsupported quite quickly becomes a problem. In this situation, a chip with mainline Linux kernel support would be very important.

If the price is lower then people are more likely to buy this device and then replace it with your next product (!) in a few years, so a fixed kernel is less of a problem. If you ship it with Android (LineageOS preferably) and good libhybris support is available (polished community port of SailfishOS) then it would enable other mobile OSes to be ported too (many of them use libhybris) and the community could work on something like sfdroid for Android support.

Have you spoken to Jolla about this, and are you free (contractually) to discuss how much official SFOS support could cost? Presumably this depends on what is included - would be interesting to know what the cost of the following items is:

• providing "perfect" libhybris support
• licenses for proprietary libs xt9 and exchange
• aliendalvik

I expect aliendalvik makes up the biggest chunk of the cost, since it's a big chunk of development work.

Even though I got stung on the Neo900, I'm highly likely to back this project. Careful with polls - you'll probably pick up some noise from people who have no intention of putting their money where their mouth is!

Re: The choice of OS of the new QWERTY device project
 

I agree on that, polls about the direction of a project must be considered very carefully for many reasons, including this one. Same goes for crowdfunding stretch goals by the way. :D

Good point on the obsolescence of the device and the kernel. Since we're probably not going to see such projects very often (even though Chen spoke of the Lauta revival as only the beginning of something on the longer term), I would definitely prefer something long-lived. That, and community ports (either for Sailfish, although again I'd prefer having Aliendalvik included, or other OSes as well) take time, especially with limited ressources and contributors. So many good software/OS projects die before completion just because the target hardware ends up being considered too outdated for people to be interested (even though the hardware is still plenty enough for the said projects), leading to less contributors and slow death of development.

Re: The choice of OS of the new QWERTY device project
 

If I remember correctly, this is just a matter of which modifier is mapped to the key, at least I got sticky keys working last time I tried. Could/should be exposed in the interface, but if you're already remapping keys, a different Shift isn't much extra work.

Re: The choice of OS of the new QWERTY device project
 

By the way, it would be nice if people could stop directly quoting Dave999, because those direct quotes circumvent the ignore list, and I accidentally end up reading nonsense.

Re: The choice of OS of the new QWERTY device project
 

More than 17 of us here that are still hanging here and hoping for some kind of alternative.
Ideally, he needs HW that can run both.

Re: The choice of OS of the new QWERTY device project
 

What about Tizen?

https://source.tizen.org

I think we should atleast look at that as a option in this early state.

Re: The choice of OS of the new QWERTY device project
 

Can we buy harmattan license and bring that beast alive again?