Hacker News Reader

Comment by grizzles | Hacker News Reader

I wish these guys would just use android from the get go. It's already running on i.mx6, so by switching now they would at least have a chance of making a decent product.

They are literally throwing away hundreds of millions of dollars of excellent work that has been put into power optimization for no good reason. I've told the ceo there this like at least 3x but he is stubborn.

> “Android is so frustrating! Trying to remove Google’s privacy invasion bit-by-bit removes functionality bit-by-bit, and you end up with a non-working phone. Purism will solve this by putting your privacy protection and security first.” — Zlatan Todorić, CTO

This quote is nonsense. There are at least 3 projects (replicant,copperhead & another) that are shipping trees like this.

It will be so much harder for them to create a phone that stays alive for more than an hour or two running a full linux desktop.

View on HN

Whilst I don't agree with the reasons either, I do agree with the conclusion: I think Android is a dead-end architecturally.

If we want a secure and user-friendly phone, it won't be Android.

> It will be so much harder for them to create a phone that stays alive for more than an hour or two running a full linux desktop.

This is of course nonsense. I've run a "full linux desktop" on 1000mAh for more than a few hours, and I did it a decade ago.

>If we want a secure and user-friendly phone, it won't be Android.

I don't think you realize what it takes to keep an operating system secure. If you really think this project is going to be more secure than Android then that's a wager I would take all day.

By using a real GNU/Linux distribution like Debian you can have up-to-date software with all known vulnerabilities patched. On Android, all the apps bundle dependencies which house tons of unknown vulnerabilities and then the phone vendor stops updating your phone when it's no longer in their business interests. Any GNU/Linux can be more secure than that terrible baseline.

And who is going to patch all of the userland exploits? The radio firmware exploits? The driver exploits? Keeping a phone secure is so much more than just downloading the latest Linux kernel security patches. Unless you have a large team of security engineers that dedicated to patching and enhancing your OS then you're going to be in for a world of hurt once your platform gains any popularity and, thus, nefarious interest.

>On Android, all the apps bundle dependencies which house tons of unknown vulnerabilities and then the phone vendor stops updating your phone when it's no longer in their business interests.

The same is true for any device that has been determined to be end of life by the OEM. Yes, it would be nice if devices were supported to perpetuity, but even Apple ends support for their devices after 5 years and more than 10 percent of their current install base isn't even on the latest OS version.

>Any GNU/Linux can be more secure than that terrible baseline.

Perhaps, but saying it and doing it are two very different things. I hope it works out because I think there's a niche for a phone that is always up to date and never written off by the OEM.

This won't be more user friendly than android. Google has invested a ton in ui r&d. They will have done very well if it doesn't crash frequently and you don't have to endure long waits to do stuff.

I kinda feel you on this. I am attached to the Android ecosystem. I know two people who have attempted to use open source phones (a rom without gapps + fdroid). They're not _really_ open source as your still have binary blog drivers, but that's as close as you're going to get really. I will give it to them, they were dedicated. One of them had done this for nearly a year, but she started to give up when she found herself writing several apps to replace stuff in that was already available in the play store.

At the same time, I would really love an alternative to Android/eyeOS. I wonder if we'd have more open source mobile tools had the Ubuntu Edge succeeded in getting funding?

One of the biggest reasons we don't see a Linux explosion on Android in the same way we saw developers using it on their desktops in the late 90s/early 2000s is because you can't just install a base Linux distribution on any phone. Unlike the PC which is a full platform, ARM devices are SoCs where every manufacture attaches random shit to random pins and creates non-upstreamable Kernels:

http://penguindreams.org/blog/android-fragmentation/

Almost every rom we see out there has to have a customized build for each specific phone. Maybe some of this will start to change with Oreo, but I think a bigger barrier to alternative OSes on phones is simply a standard platform.

The only reason I am willing to fund this phone is because it's not android. If I wanted "just another android", I'd get one. They are a dime a dozen, and all terrible, imo.

> This quote is nonsense.

Heavily disagree. I've worked with those projects, and they are terrible and severely limited systems.

Purism has been pushing for actual foss systems, which I would strongly argue Android is not, and is moving farther from every day.

Me and many other people have been begging for an actual linux phone, and I'm excited to see where this goes.

You can run debian on android today. You can make android look however you want it to look. What they are losing is doze, app standby, the compiler tech, they will have to redo wakelocks, the resource manager, etc.

In android, nothing is ever actually running other than background services and the foreground app.

Since I assume they don't want to tell their users "never start a second app" I wonder where all this infrastructure is going to come from. They will have to reinvent a whole new app paradigm.

Sure you can run debian on android. I don't want debian on android though, I want debian.

I can run debian on windows as well, that doesn't mean I'm going to do it.

Why settle for a hacky second layer system, when you can just have the whole system?

And sure, I can make android look the way I want today, but I can't make it do what I want today, which is the important part.

Appeal to futility...

You can always just use the same old same old, but you aren’t gonna create something new - the thing of your dreams - by doing that.

They might succeed in this, and if they do, it might be awesome!

> You can always just use the same old same old, but you aren’t gonna create something new - the thing of your dreams - by doing that.

One could argue the opposite — that you can build the thing of your dreams far better and faster by leveraging non-differentiating work done by others.

For example: Even a decade in, iPhones use lots of materials and components from other vendors.[0][1]

Maybe Android isn't the perfect OS choice. But building a phone without access to an ecosystem of apps relegates it to a curiosity for most, which seems like a shame.

[0] http://www.macworld.co.uk/feature/apple/where-are-apple-prod...

[1] https://images.apple.com/supplier-responsibility/pdf/Apple-S...

To build a phone on top of Android, you would have to believe Android is a good system to build upon. I don't believe that is a fact though. There are lots of problems with Android. Sure, your phone gets access to a lot of apps from day one, but 99% of those apps are shit, so why bother.

We've got so few mobile consumer OSs in use today, and I don't think any of them are particularly great, although I believe one of them are much better than the other. I welcome completely new approaches to it instead of the way it is today where we have Apples stuff, and then we have numerous companies that makes shitty Android stuff, and a few that makes some okay great Android phones.

Lets get new stuff on the table, please.

>There are lots of problems with Android.

Were you planning on naming them or keeping us in suspense?

I did mention one, but otherwise no.

The power management app that comes with KDE/Plasma does a pretty good job with my laptop battery, certainly no worse than Windows did. It throttles back the CPU, dims the screen depending partly on user activity, has sensible policies and notifications for stretching out that last 10% of the battery. It also allows apps to request an exception from power management. What's wrong with the current OSS PM tools?

In my experience, the most effective power management tool on any device is actually just airplane mode.

On my laptops, tlp on Ubuntu/Debian/Arch (when I used those) and tuned on Fedora (which I've used since F25) have been awesome, so power management is unlikely to be the biggest issue with this phone.

Off-topic, but how do you feel tlp and tuned compare? tlp is available in Fedora repos, along with tuned, and I wonder which gives a better result.

I've used both on Fedora, and found tuned to be easier-to-use and more effective. tlp also does not work well with SELinux, and as such causes a flurry of errors every time you change the power state (sleep, resume, plug, unplug), which may be part of why I found it ineffective.

The way I set up my tuned profile is pretty simple:

    sudo dnf install tuned-utils
    sudo powertop2tuned -n laptop

Then, I edited the profile in `/etc/tuned/laptop/tuned.conf` and enabled things. On my Thinkpad X220, simply enabling everything worked great, but my Asus ROG laptop and my friend's Razer Blade had issues with that, so YMMV wrt what to enable. Then, I added USB autosuspend support by adding the following to the profile:

    [usb]
    autosuspend = 1

After that, I ran `sudo tuned-adm profile laptop`, which sets it and tells tuned to start on boot, and have had vastly improved battery life, good performance, and no hassles since.

Excellent, thank you. I've followed your instructions and am waiting to see how it all goes. My laptop is an X230, so should be similar to your experience.

I feel similar. If they provided a fully open phone where they worked closely with copperhead, it would be significantly more interesting.

Make it dual compatible, work with copperhead/AOSP android and whatever they are trying to invent.

Open hardware with hardware switches and a separated baseband is fairly interesting just by itself TBH.

There is no freedom in using Android. For the better part of a decade I've wanted to install phone software on a regular GNU/Linux distro and install it onto an ARM smartphone. A phone is a general-purpose computer and you ought to be able to use whatever distro you want on them. This project seems like the best shot of that happening.

Reply on news.ycombinator.com