Why I will be cancelling my Librem 5 preorder (different reasons from other users)

Hardware Librem 5
8

@Bum112, Let us know how your attempts to get a refund go.

Considering that the crowdfunding started in August 2017 and most people who preordered aren’t expecting to get their phones until Q1 of 2022, I can understand why people would cancel their orders for the L5.

However, I’m guessing that you want the community to discuss your points, so I’ll take a stab at it.

I agree that Purism, PINE64 and Jolla are fighting a very uneven battle by trying to develop mobile Linux, when Google and Apple offer well developed and very polished systems with 3.5 and 2.1 million apps in their app stores, respectively.

However, there are a number of reasons why I believe that mobile Linux will eventually succeed to become a viable alternative to the Android/iOS duopoly:

The work to add libhandy/libadwaita and Kirigami classes to desktop GTK and Qt applications so they can run in mobile Linux is far easier than creating new FOSS code from scratch, and we can see that it is already starting to happen.

Once you get serious about using only free software in an AOSP+F-Droid system, you will start to see a lot of holes in the system that need to be developed, just like in mobile Linux. For example, if you want decent speech recognition, text-to-speech voices, OCR and augmented reality that is FOSS, you will have to develop it in both AOSP and mobile Linux. Sure, there are a few areas, such as camera apps and virtual keyboards, where AOSP is ahead of mobile Linux, but on the other hand, desktop Linux actually has more FOSS games than AOSP+F-Droid, and it is a lot easier to to adapt the interface of a desktop Linux program to run it in mobile Linux than to port that same program from Linux to Android.

However, I don’t think that crowdfunding the development of Linux phones and going through the trials of being an early adopter is for everyone. If the goal is to get a working privacy phone as soon as possible, clearly an AOSP-derivative is the better solution for most people, and that is what a number of us on this forum have recommended that people buy.

I have used LineageOS+F-Droid on my phones since 2015, and for people who just want a privacy-respecting phone, an AOSP-derivative phone really is the best route. Today with /e/ Foundation, jolla-devices.com and Rob Braxman all selling phones with an AOSP-derivative preinstalled, it is easy for anyone to get these phones, even if they don’t want to install it themselves.

There are a number of organizations (/e/ Foundation, CalyxOS, GrapheneOS, etc) already working on AOSP privacy phones, so I don’t think that Purism would add much to current efforts if it decided to switch to making an AOSP-derivative phone, whereas Purism is playing a vital role in developing mobile Linux. Also, Purism has little expertise in AOSP and would have to fire most of the 11 developers it current pays to work on the Librem 5’s software, and it is often a bad idea for companies to try to do something outside their core competency.

AOSP may be better developed than mobile Linux and F-Droid has a decent app store, but it does little to address the 3 core problems that I see in the mobile phone industry, which are surveillance Capitalism, planned obsolescence and monopolization. I don’t see much hope of us eventually getting to a better place with AOSP, whereas I do have some hope that mobile Linux will solve all three of those critical problems.

With an AOSP derivative, we are still dependent on Google, which is the chief company in the world promoting surveillance Capitalism, and even if Purism forks AOSP, it is going to be very hard to find volunteers or companies that are going to take over the development, so that Purism doesn’t have to keep drawing code from Google’s latest AOSP release. If AOSP derivatives ever became a serious threat to Google’s business model, than Google could simply decide to not release the next AOSP version under a BSD licence, and then we would be screwed, because we don’t have maintainers for the AOSP code and it would be incredibly hard to recruit an army of volunteers to take over the maintenance of all that code. It is worth reading this article about how Google has progressively privatized essential bits of Android and AOSP has been stuck with outdated code. It is already a problem, and Google could make it much worse if AOSP derivatives ever start to pull any sizeable market share.

If you use F-Droid, you will see a lot of apps that have been abandoned and haven’t been updated in years, whereas Linux applications have a much better history of code being maintained, and it has an ethos that encourages volunteers to join existing projects, rather than creating new apps as in Android/iOS. There is the constant temptation for Android/iOS app developers to privatize their code or to start working on another app whose code is proprietary and/or monetizes users’ data, whereas it isn’t nearly as easy to do that in the Linux world, since there are few Linux distros that will accept those kinds of apps and there is strong community disapproval of that kind of code.

Most phones running AOSP-derivatives don’t get driver updates because Qualcomm, MediaTek and Samsung only support their mobile processors for 2.5-3.5 years (and most HiSilicon and Apple processors have locked bootloaders don’t allow the installation of AOSP). Google uses outdated kernels in Android and it can be very difficult to upgrade the kernel since the chip manufacturers often don’t release firmware/drivers for new kernels, which is why each new version of Android supports 3 different kernel versions so the kernel doesn’t have to be upgraded. Look at all the problems with the Fairphone 2 using a kernel from 2012 in the year 2019, and you start to realize why Android/AOSP makes it hard to fight against planned obsolescence: Frequently Asked Questions · Wiki · Librem5 / Librem 5 Community Wiki · GitLab

Yes, Purism could have used AOSP with the i.MX 8M Quad and it would not have had the same planned obsolescence problems as other phones, since NXP promises to support both Linux and Android on the chip till Jan. 2033. However, that is a special case with NXP, and the goal as I see it is to create a new market for phones that have long-term support, and that will be much easier to do with Linux than Android/AOSP.

I don’t understand your point about “no apps or widgets on the homescreen” of Phosh:

1200×800 56.7 KB

Which screen are you talking about?

Considering that Phosh uses its own compositor and it is an entirely different shell from GNOME3, it shouldn’t be that hard to convince Purism to alter its interface if you file issue reports and explain what changes you think it should have.

Considering that my poll on the PinePhone forum found that 56% selected a distro using Phosh as their favorite, versus 17% for Lomiri and 14% for Plasma Mobile, I’m not sure that most mobile Linux users agree with you. (My poll has an admittedly small sample size of 36 voters, so it may not be representative.)

I agree that $799 and the coming $899 price hike is a lot to ask. The specs were OK in 2017, but in 2021, they are very outdated. I saw the Tecno Pova on sale for $140 last week with 6GB RAM and 128GB storage.

The i.MX 8M Quad processor limits the video processing for the camera to 1080p@30 and the RAM to 4GB, so Purism is limited in how much it can upgrade the specs. My advice to Purism is that if it is going to raise the price to $899, it should offer 4GB RAM and 128GB storage. My web searches find that the cost of that upgrade is far less than the $100 price hike, and I believe that better specs will generate enough extra orders to justify the extra expense.

One thing that many people may not realize is that there wasn’t any better option to design a phone with hardware kill switches in 2017 than the i.MX 8M Quad, even if Purism had decided to use proprietary blobs. Every processor that Purism could have used either sucked way to much power (like the RK3399) or had built-in WiFi, Bluetooth, GNSS and cellular modem that would not have worked with the hardware kill switches. Today, we have the RK35xx processors on the market, so Purism has more choices for future models, but if you look at what Nvidia, Allwinner, Amlogic, Rockchip, Broadcom, etc. were offering in late 2017, I think that Purism made the best choice available at the time.

Considering that my poll found that Plasma Mobile is less popular than Phosh, Lomiri and Sxmo among PinePhone users, and only the last one of the six PinePhone Community Editions shipped with Plasma Mobile (vs three with Phosh), it is clear that PINE64 didn’t choose Plasma Mobile because it was the best or most popular interface. I believe that PINE64 chose Plasma Mobile for the PinePhone Beta for the following reasons:

  • Plasma Mobile works on a community development model that PINE64 promotes,
  • Manjaro, which is PINE64’s partner, is willing to provide support for Plasma Mobile,
  • PINE64 didn’t want to promote the interface of its chief competitor,
  • It allows the PinePhone to distinguish itself from the L5 in the market.

There are a lot of advantages to Phosh over Lomiri and Plasma Mobile in my opinion. See: Frequently Asked Questions · Wiki · Librem5 / Librem 5 Community Wiki · GitLab

12 Likes