Why Librem 5 will never succeed (in my opinion)

Why Librem 5 will never succeed (in my opinion)

Nils Gerarß

First let me start with, I work for one of the (android) fabless mobile manufactures out there, and i work with ODMs and the software stack on a daily basis.

I first have to say I love the effort purism is making and I really hope this is the way the mobile landscape is going. This is something that the (especially smaller) mobile manufacture our there want, we don't want googles locking, but it's currently the only thing that makes us money right now. I don't think purism have any evil intentions with what they are doing. But what they are doing is screaming failure.

  1. Using an Iot CPU that is not designed for phones is and will be a huge power draw and over heating issue. Just look at the massive heat sink on the devkit that can barely cool the thing in open air. This on a closed plastic case will overheat on startup. Plastic without metal infusion is a really poor conductor. Looking at the picture of the pcb (https://social.librem.one/@purism/102779164041504324) how are they even gonna get space for a heat sink there? normally this is done by placing the cpu on the side of the screen and using the screen as a heat sink, but that would not work on such a power hungry cpu as the nxp i.mx 8m. They might be able to fudge this using software, but then it will still be a huge problem doing any sort of task on it. Or they limit the cpu, but why use this cpu in the first place then...
  2. They are tying to make both software and hardware? are they insane? I don't think they realize how hard hardware is, but software for the mobile space is even harder. Sure you can easily get a nice looking gui up and running. but there is so many parts to a mobile, sleep more, screen saver, alarm clock, gps, bluetooth, wifi, touch, wakeup, background services, touch ui, permission handling, vibrators, notification, leds, wake locks, powermanagment, cameras, apps, just to name a few. Just look what canonical tried to do, they had over a 100 people working on it at some point that spanned over years. They got really really far, but still not complete. I have to say i'm impressed what the community has done with ubuntu touch, we at the office are always watching it but they could never have done something like that without the 10's of millions that canonical put into it! If they only focused on the hardware and for example put Ubuntu touch on it, it could be possible! But doing software that on top of it all is using gtk/gnome that already run poor on chromebooks using arm hardware is a huge task that cannot be done by a small team in the time they want (need at least 5-7 years).
  3. Shipping "development units" to people that paid for the full device (and not giving them the full device in the end) just to hit the deadline is so disrespectful to the backers. If they just where honest with the backers and not wrap it into a marketing bullshit. Just say "We cannot reach the deadline, we too way to much on our hands, but we have these early development devices, we can ship now if you want one. but the finish product we need to delay sorry".
  4. Be clear on what the current status on the hardware is. Since there is no image of the "phone" running anything we can only assume it's not anywhere near running. Since a pcb to a working pcb is a long hard road (unless you got lots of money), testing, failing, testing failing repeat. and when you get the pcb working there will be MANY hardware bugs that needs to be worked out.
  5. Trying to be the "apple" of open source so early is a bad move. Don't wrap everything in a marketing package to people that love tinkering with hardware. If you want to be the "apple" of open source, you need to first build up a brand, build up reputation and deliver hardware on time.

If i was the project manager, i would use Ubuntu touch as software and snapdragon cpu. Make it run mainline with freedreno. Use an existing ODM template to include the hardware modification I need. And start there. Build your first phone and then grow from there making new versions that is more and more complex as you gain the experience, trust and capital.

So in the end, this all is a sad thing that we often talk about in the office, as we could see this happened with our experience and i believe any project manager should have been able to see this, we are more looking at the pinephone as a great example of how to do this, they have been building many arm based hardware both boards, tablet and laptops, they are also using a mobile cpu and existing software. We hope to release a phone with gnu/linux ourself, but we need to see how it plays out first. We don't have the money to experiment with OS that might not sell. We all love linux, but sadly we need money to survive.

I really wish to be proven wrong, but what many year experience in this exact field, its not looking good.