Growing mobile Linux so other phone makers will adopt it

Starting up this topic again, in the What of the 180 day lead-time? topic, @amosbatto writes:

Because Phosh was designed as a thin overlay on top of the desktop GTK/GNOME ecosystem, it should be significantly cheaper to maintain and develop than the other mobile Linux interfaces, and unlike Plasma Mobile which has almost no corporate support, the GTK/GNOME ecosystem is supported by every single one of the large Linux companies (IBM/Red Hat, SUSE, Canonical and Google), so Phosh can take advantage of the dev work being done by those companies.

Building upon well-funded efforts is the way to go. However…

Maybe I am wrong about the financial situation of Purism, but what I am sure about is the fact that none of the big Linux companies care about mobile Linux and are willing to pay for its development. If we decide that we don’t need a company like Purism to pay for software development, then we are choosing to rely solely on volunteer labor to develop mobile Linux, which is going to be slow and unlikely to create a mobile OS that appeals to ordinary users who don’t have technical skills.

So, the big companies we rely on don’t care about the mobile market, and many of them are software and service vendors, anyway. It’s clear that relying on volunteers isn’t going to deliver the devices that some people want, otherwise we would have them already. Relying on one small vendor isn’t going to do it, either, because it’s too much to expect and everything depends on their success.

The question is: how can mobile Linux be sustainably funded? Perhaps, lurking beneath that, there is also the question: is mobile Linux actually what people want?

4 Likes

For some people there is an incipient realisation that there is something wrong with the business model of the Apple / Google duopoly, relying on or encouraging as it does the rampant plundering of your privacy and freedom. Open Source is a means to that end, and Linux is the obviously available Open Source answer (but not necessarily the only answer).

I contend most people don’t care what the OS is not whom “owns” it as long as they can do the things they want to do. So until mobile Linux can do the things people want to do it won’t be what they want.

3 Likes

The obvious answer is that people need to put their money where their mouth is. Software development is expensive no matter how “free” it is.

The problem as I see it:
Linux users like doing things different. Many of them feel special simply because they know how to use Linux. Others because they are more intimate with its inner workings, etc. But they all have different opinions about what is good and what is bad for Linux. All of them feel entitled in their opinions, and seem to take great pleasure in tearing down anything that doesn’t absolutely and completely conform to it. It is very dog eat dog.

I know of NO organization within the Linux world that is held lofty by all Linux users. None! That has been, continues to be, and looks like it will be the problem for Linux for as long as the eye can see.

2 Likes

The magic is going to be either:

  • Purism figures out a subscription model that allows them to make money off of an established clientelle in perpetuity, for which a smartphone is a perfect target.

  • Purism figures out the app store for their established clientelle and makes reasonable margin on it.

THE MOMENT anything along these lines happens is the moment blood hounds show up and start providing competing offerings.

I can see other roads to success:

  1. Support from foundations (similar to signal);
  2. Governments mandating the use of open source to protect the national interest. Sensitive personal data in the hands of foreign, private companies should be very troubling to politicians and government officials.

In the short term, I think it is going to be a small, relatively wealthy elite that will support mobile linux. I would hope that there is enough interest and wealth in this group that mobile-linux companies can make a reasonable return on effort to keep development moving forward.

2 Likes

I don’t see governments promoting open source. US government is lobbied by big tech, which cares about control, not about people, and US has influence in very large part of of the world. EU has the same lobbying problem as evidenced by recurring ACTA in various disguises. On another end there’s China and Russia - totalitarian states, and they care about control over citizens, not their liberties. So either way, there’s no motive for governments to promote open source in hands of citizens.

Edit: Open source is bad enough for governments, but free software is a total outrage for them.

7 Likes

Mobile Linux will be significantly funded the same way that any other Linux gets significantly funded.

When I started learning Linux, Redhat 4 was the latest version of what now has become Fedora Core 34, and is still evolving now. But back then, Windows 95A was popular. I had to partition my 5 GB hard drive in to three logical partitions because at the time, Windows 95 couldn’t read or write to a hard drive larger than 1.7 GB in size. People said a 5 GB hard drive was way too much. Who could ever use that much hard drive space? At the time Linux had no hard drive size limits that I knew of. Linux was primitive then. If you didn’t like a bash prompt, you were mocked as weak by most of the Linux community. No self-respecting coder would build ‘training-wheels’ kinds of apps and GUIs for those who were unwilling to learn a command line environment and how to write their own code. The internet wasn’t very helpful back then to help you with the command line information. You were on your own for the most part. That’s how the Linux community liked it back then. If they needed a new feature or program for themselves, they would re-compile the kernel if necessary, to make their own apps work for themselves. Everyone else had to use windows or learn to code or at least, live most of their lives at a bash prompt, completely held hostage by their own lack of programming skills.

The latest version of Fedora Core now is 34. You don’t need to know anything about Linux now if you can put a live CD or DVD in to your computer, and answer the questions that are presented to you in the Beautiful interface on your screen. Eventually, you click “ok” to reboot and you’ve just successfully installed Linux on to your PC. From there, most distros give you an operating system with GUIs that rival or beat the latest version of Windows. The command line may be helpful. But if you want to use ‘training wheels’ (GUIs) and never learn a command line for the rest of your life, that works too now.

Somewhere between those earliest linux versions (long before Redhat 4 and even long before the first version of Windows), someone had to do continuous improvements to Linux. Who paid who to make Linux what it is today? Is the development going to stop if we quit paying those people now? If you answer these questions correctly then congratulations ! Add another zero to your paycheck for your own most recent contributions to Linux. Because if you get paid zero to begin with, then my adding another zero to your paycheck is very affordable to me. In fact, while you’re at it, add two zeros. I am feeling especially generous today. Either way, Linux is going to continue to grow, whether or not you or I ever contribute even one line of code to it. Someone always will. That goes for Mobile Linux too. As soon as its value has been established in a phone, the development from that point forward is like free energy. No one will need to get paid to improve Mobile Linux after that. Will work for Kudos.

Richard Stallman used to say that open-source is “free as in freedom of speech, not free beer”. But the Linux operating system itself is definitely more like the free beer for most of us.

6 Likes

Nice link. How are you framing this to be relevant to the discussion here?

3 Likes

I think that Purism has already found the best model by trying to make a version of mobile Linux that needs the least independent development, and is closely tied to desktop Linux, where there are well-funded companies and big communities already doing the work. Canonical and Mozilla failed because they tried to create a mountain of siloed code without any outside help.

There should be a strong niche market for people who are willing to pay a hefty premium for privacy and security which is user controllable, just like people pay a premium for gaming phones and rugged phones. Purism can use those higher prices to pay developers, and we can see how quickly Phosh is getting adopted and outside contributors are already testing and adding to Phosh, so I think that the development of Phosh will be sustainable in the future.

The problem as I see it is that the Phosh ecosystem and the Librem 5 does yet have good enough software to attract many users who aren’t committed Linux geeks, and most Linux geeks aren’t willing to pay much for their hardware, especially when it is underpowered. There is a sizable market for people who want privacy or need transparent security, but Purism has to get Phosh, its apps and the i.MX 8M drivers to a good enough state to be able to tap into that larger market.

It will probably be another 2 years before the Librem 5 has good enough software to appeal to non-technical users without a background in Linux, so the question is whether Purism can keep financing development during that time period when it is getting few new orders. Purism has done a poor job of explaining why the development of Phosh was necessary for the future of mobile Linux. This is a delicate issue, because I don’t think that Purism can talk too much about the deficiencies in the other Linux interfaces without causing serious problems with the UBports and KDE communities which have collaborations with Purism.

In order for FOSS to work well, it needs both a core of paid developers and wider community of volunteers. The current involvement of Mobian, postmarketOS and the GNOME community in the development of Phosh is largely the result of the good sales of the PinePhone and the fact that Purism has worked hard to make Phosh compatible with the existing GTK/GNOME ecosystem and to contribute to it, so its existing applications will run in Phosh.

I see an overwhelming desire for mobile Linux, which is evidenced by the fact that there are still people who desperately want their Librem 5, and so many people have bought the PinePhone and are working on developing for it, despite its underpowered hardware.

There is a growing public awareness about the risks of surveillance Capitalism and the data shows that people are keeping their phones for longer which means that they are bumping up against the planned obsolescence in Android, and I think that both of those factors will fuel the desire for mobile Linux. Mobile Linux empowers ordinary people to fight for their right to privacy and avoid the potential for 1984-style government surveillance in the future. With China becoming an Orwellian surveillance state and the Snowden revelations about Google, Microsoft, Dropbox, Yahoo! and Apple sharing people’s personal data with the NSA, I think that people are starting to see the risks, but it is very important to have a functioning alternative to show policy makers and to prevent companies like Google from becoming too abusive because they fear losing users to mobile Linux.

Desktop Linux has a number of big companies funding its development, but that won’t be the case with mobile Linux as I explained in the parent thread. Only 16% of the commits to the Linux kernel between 2007 and 2019 were done by people who weren’t being paid (11.95% by people with no affiliation and 4.09% by people with unknown affiliation). If you check who develops Firefox, the vast majority of the commits either come from paid employees of Mozilla or pieces of code that were made by paid developers of other projects. 20.6% of the commits to LibreOffice come from unpaid volunteers and the majority of the commits come from just 21 developers who make more than 200 commits per year. If you check who contributes to Python, PHP and PostgreSQL, you will find that a lot of the developers work for companies that use these tools and they are being paid for their work.

Commits to LibreOffice between 2019-08 and 2020-09

Organization # commits % commits
Collabora 7518 46.0%
Red Hat 3525 21.6%
Volunteers 3367 20.6%
CIB 719 4.4%
NISZ 500 3.1%
Document Foundation 492 3.0%
Gov’t of Munich 114 0.7%
1and1 21 0.1%
SIL 14 0.1%
Other orgs 9 0.1%
Total 16341 100.0%

What company is going to pay for the development of mobile Linux? It is wishful thinking to believe that companies are going to appear to pay for mobile Linux’s development because there is no business model to justify it. What you are going to get is projects like Plasma Mobile, which has been in development for twice as long as Phosh, and is considered to be less usable than Phosh by most users.

I don’t mind people opting for slower, 100% volunteer development, just so long as they are very clear eyed about their choice and its significance. I can’t foresee Plasma Mobile ever reaching mainstream users due to its slower development and the focus of its developers, whereas Phosh has a shot in my opinion. The principal question is whether you want mobile Linux to ever be able to reach a mainstream audience or are you content to have a system that just appeals to Linux geeks.

5 Likes

I know this is a popular opinion in some circles, but this can’t be further from the truth. FirefoxOS failed commercially because of the lack of official support from a very popular messaging app starting by W. Given the distribution model (partnering with carriers), this was a chicken and egg problem that never got solved. Then Mozilla leadership made the mistake to not pursue a more independent model in 2016. They would be in a great position today if they had, but the project is not dead, see https://github.com/b2gOS/B2G/

@fabrice,
B2G only has 6244 lines of code, but the problem as I understand it was that Mozilla had to create and maintain the large number of the applications for Firefox OS. I recall reading that the phone makers using Firefox OS kept demanding that Mozilla Foundation needed to create more apps to match every app what Android was offering, and Mozilla couldn’t keep up with the demands, and Mozilla wasn’t getting hardly any help from the wider community.

When I run cloc on B2GOS’s tablet-ui directory, it says that there are 192,704 lines of code, which isn’t as much as I expected. Is that all the application code that Mozilla Foundation developed or did B2GOS strip out a lot of the application code?

Nonetheless, Phosh and Plasma Mobile have a much better strategy in my opinion of adapting the existing GTK and Qt desktop applications with libhandy/libadwaita and Kirigami classes, and taking advantage of the work of the existing GNOME and KDE communities, respectively.

I don’t disagree that Mozilla failed at mobile Linux partly because it couldn’t find hardware partners that were committed to Firefox OS, and Mozilla would have done much better to sell its own phones, rather than trying to convince unreliable phone makers who didn’t understand how to market phones to Linux enthusiasts. Canonical made the same mistake when it gave up on selling its own Edge phone.

However this is what I think was the principal problem:

Mozilla Foundation made similar mistakes in designing and marketing Firefox OS. It couldn’t run the GTK and Qt software used by desktop Linux, so it held little interest for the global community of Linux enthusiasts who wanted to be able to carry around Linux in their pockets. Instead, Mozilla’s plan was to market Firefox OS as cheaper alternative to Android in developing countries. Google responded with its Android One program to drive down the price of Android phones in developing countries, and most poorer consumers in developing countries would rather have Android with its millions of apps, and they have no appreciation for the fact that Firefox OS was built on free web standards. There was no compelling reason for tech enthusiasts to buy the low-spec Firefox OS phones and consumers in developing countries also couldn’t appreciate the phones either since their limited app selection make them appear as poor alternatives to low-end Android phones.

Ultimately, KaiOS made Firefox OS work by privatizing the code and focusing on phones with buttons that lacked touch screens, so it could reach an even cheaper market where Android simply doesn’t compete. People buying $25-$50 phones without touch screens don’t expect to have an app store with millions of apps, and the limitations of Firefox OS were less of a disadvantage.

Thoughts?

The code for FirefoxOS is spread out in many repos, not just the one I linked to (see https://github.com/mozilla-b2g/gaia for the smartphone UI for instance). Mozilla didn’t “fail at Linux”, this was not the goal at all - the goal was to leverage the Web as a viable mobile application platform, because this is where you can find the most developers out of Android/iOS.

I agree that it’s hard to sell “the Web” as a competitive advantage in itself if you can’t show additional user benefits. I think it would be easier now that in 2013-2015, due to a better general understanding of the abuses of the duopoly and how the Web can provide better user agency.

I would not be in this forum if I was not interested in Linux mobile, but I still think that fixating on GTK/QT will limit the reach to a very small niche. That can be absolutely fine if Purism or others like Pine64 manage to turn that into a viable ecosystem, but I don’t believe this will reach the scale needed to displace the duopoly.

The other elephant in the room is the coming of Fuchsia: if/when Google switches from a Linux kernel to Fuchsia for Android, this will make things even more difficult for Linux mobile, because the overall resources devoted to make Linux good on mobile will be starved. With Fuchsia not being under a GPL license, the blob situation will deteriorate even more. Note that Google started to test Fuchsia in production on one of the “home” products, and that the current Android hardware abstraction architecture already looks a lot like a micro-kernel with many services running in their own isolated process.

Mozilla is run by pure incompetence held afloat by magic money from Google because they have to prop up a corpse to pretend there is competition in the browser space in order to not get into legal problems. The entire org is a zombie on googles marionette strings. It’s hardly a mystery as to why the failed.

They really did have top notch engineering team though. We have to figure out how to fund the engineering vision of that engineering team and get the browser space moving again.

With that being said, what is a mystery is why Mark Shuttleworth failed. Maybe he too was running out of resources because he too (like Mozilla foundation) kept on hiring ideologues for positions that require pragmatic sales / bottom line oriented sharks. The Ubuntu phone has the whiff of a hail mary pass before he figured out that he needed to follow the Redhat model to make Canonical break even. But even then, had he set aside 10 million, he could have had an L5 / Pinephone on the market way ahead of anyone else (even at that time this was entirely possible) and could have actually owned this niche.

The sad part of the Mark Shuttleworth story is that the Ubuntu phone could have been THE SOLUTION to his need to generate subscriptions where his desktop offering couldn’t squeeze a penny out of a 100 dollar bill.

@amosbatto, dude, I think that either:

FIRST:

Purism needs to bring you on board as a consulltant. You seem to have the rare talent stack that spans understanding of the technology stacks, business, hardware / hardware manufacturing, specific use case of the several demos / niches that L5 is appealing to AND a general understanding of human nature and how that applies marketing and sales.

I am sure that I am not the only one that appreciates your well rounded and unbelievably well research posts.

I think Librem would do well to bring you on board and triage communication between them and the community.

While it looks like silence is killing them, I have enough experience to know that they could do FAR FAR FAR more damage with bad messaging. So the current strategy is a kind of passimum for minimizing risk.

They might do better though by hiring someone with your talent stack to massage the message the right way and engage the right communities in the appropriate way.

Although, to be fair, if they could just catch a break and get some supply of parts and start shipping the rest takes care of it self.

SECOND:

It might be interesting if you started a blog that was a kind of second opinion on how to understand the Linux smartphone space and how to think about addressing the needs of the various demographics that the L5 / Pinephone and others are appealing to.

You probably don’t have time for that, but, it could be the kind of thing that stokes the right kind of interest.

If anything, it would be nice to have someone that can provide a visionary bridge between the Pinephone and L5 communities to start. Despite the challenges all of these projects face the future for both projects and then further down the road for other projects that will come along to fill out the ecosystem is UNBELIEVABLY EXCITING.

1 Like

It is ridiculous what Mozilla pays Mitchell Baker, but I give Mozilla a lot of credit for its work on Rust and WebAssembly and being the first one to attempt to block tracking, which forced Apple and Google to follow suite. Mozilla has been trying to reduce its reliance on Google’s advertising, but there aren’t a lot of good options. Frankly, if Mozilla’s market share keeps declining, we are basically allowing Google and Apple to control the future of the web and define the web standards, so I will keep supporting Mozilla rather than Chromium derivatives.

Here is my take on what happened with Ubuntu Touch:

Canonical started out in the right direction by allowing Ubuntu Touch to run a standard Linux terminal and the Qt and QML software that would appeal to Linux enthusiasts and focusing on providing convergence, which was a feature that Android didn’t provide. Canonical initially focused on selling the Ubuntu Edge, a high spec phone, to the global Linux enthusiast market, but Canonical had no experienced producing hardware and it thought that it needed $32 million in crowdfunding to make the Edge work, and it abandoned the campaign after it raised $12.81 million in just one month. The Edge demonstrated that there was plenty of global demand for a expensive, high-end phone geared to Linux enthusiasts, but Canonical didn’t have any experience developing and marketing hardware and thought that it needed to raise far more money to make the Edge a viable project.

One of the problems was that Canonical decided to reinvent the wheel and paid for the development of its own Mir display server and the Ubuntu Touch desktop environment. Then, Canonical focused on creating a whole set of new apps for Ubuntu Touch and creating Unity 8 to unify the desktop and mobile environments. It poured massive investment into creating a giant siloed codebase, with no partners to help develop it. Probably one of the reasons why Canonical decided that $12.81 million in crowdfunding wasn’t enough was the calculation that it needed more funding to pay for its large programming team.

After Canonical gave up on producing its own hardware, it couldn’t find phone makers who were willing to market Ubuntu Touch to the global market of Linux and tech enthusiasts. BQ and Meizu produced Ubuntu Touch phones as experiments for the European and Chinese regional markets, but they were never fully committed to the OS, and did little to promote it to Linux enthusiasts.

That is an interesting breakdown. I had forgotten a whole ton of the details. I appreciate that.

I forgot that you could actually buy from a limited quantity offered in the EU.

Although some of what you wrote suggests that they couldn’t find partners to make the phone, I think it mostly means they couldn’t find partners to make the phone at volumes and global placement that would make it a competitor to Android / iPhone, if even in a limited fashion. Because, honestly, around that time there were a good number of homebrew projects pumping out essentially the same mobile devices minus the modem in various form factors including things like the Pandora (or was Pyra started around that time already?).

It’s just intersting to think that they had a phone already built and ready to address the enthusiast market space, yet somehow they dropped the ball. It took a much narrower focus of the Pinephone team to bring a device to reality in the niche with Purism looking to provide an experience a little closer to what some of us need.

However as implied above, “he who pays the piper calls the tune”. Money talks. Right now Amos is free to express his own opinions. Of course it is also possible that he already has a day job. :wink:

1 Like

@amosbatto could even simply collect all his posts from here in that blog. Currently, they are only seen by the Purism community; one has to dig in these forums.