Pine64 November Update: KDE PinePhone CE and a Peek into the Future

> The new RK3566 SoC looks really promising - once it eventually makes it into the next generation PinePhone, it's going to make it very attractive as a daily driver.

I'm not looking forward to that. For example, mainline lowlevel software stack for RK3399 used in Pinebook Pro still doesn't support suspend to RAM, and RK3399 is quite popular, with many devices based on it. There's barely any recognizable community around RK SoCs.

Not sure what you consider 'viable', but here is a RISC-V mobile development platform: https://www.crowdsupply.com/sutajio-kosagi/precursor.

> I do wish they had a Pinebook Pro and Pinephone with a bit more oomph.

The new SOC they say they have planned for their 2021 lineup seems to be just that.

I wonder if I will be able to buy a simple mainboard upgrade to my existing PinePhone once they have something shipping.

I certainly wouldn't mind more "oomph" in my PinePhone and it would be nice not having to buy a new device entirely.

I agree .. I've been hacking away on the PineWatch as a platform for the last few months, mostly to teach myself baremetal Rust, and it is an extraordinarily satisfying circumstance to be able to tell time on a watch running my own code. The hardware is decent for the price (very cheap), and it is so very, very refreshingly open.

I will pick up a PinePhone as soon as possible, based on this experience, and I'm probably going to be a Pine stalwart for the next few years, at least, as a general platform. These guys are finally delivering devices that we developers can be happy to get involved in.

I say that as a dev with years of iOS experience, where I am currently fatigued with having to keep up with yet more proprietary bollocks from Apple: the future of these devices is open.

I don't want them to focus on RISC-V. Only time will tell how appropriate RISC-V can be technologically in a smartphone application, and why would you want the company that focuses on open source to invest themselves into it? Are you aware that Riscv being an open ISA has absolutely nothing to do with open source (either software or hardware)?

Their upcoming Pinecil soldering iron runs on RISC-V, so there is that. :)

Really looking forward to it, it has been schedule-slipping for as long as I've been tracking now (this fall, basically).

https://news.ycombinator.com/item?id=25101115 One of the dev behind plasma mobile commented on that a few days ago.

For the majority of distros:

  - calls -> yes (to include VoLTE on a lot of US carriers)
  - SMS -> yes
  - battery life -> not very good (3-4 hours screen time?, less than a day standby)
  - mobile data -> yes
  - wifi ->
  - (optional) MMS -> no

They mention suspend from wakeup on incoming calls drastically improved this month, as did call audio quality. The author is now using it as a daily driver for his sim card, but carries an Android alongside it.

They’re optimistic that it will be a recommendable daily driver for some people (probably like you) by early 2021.

No MMS last I checked, which is always a bit surprising to me.

Are the developers all outside the US, in countries where MMS isn't used as much so this isn't perceived as a deal breaker? Or is it just a difficult feature to support?

The list of apps depends on the operating system. There are >13 OS available [0]. On Debian, which I prefer, Telegram works very well.

[0] https://wiki.pine64.org/index.php/PinePhone#Operating_System...

I don't know about other OSes but it's available on UBPorts, and it works fine.

https://open-store.io/app/teleports.ubports

There is a list available: https://mglapps.frama.io/

Telegram is installed by default on Mobian (an attempt to bring pretty vanilla Debian to the Pinephone, but running the Phosh UI).

Telegram Desktop works fine when tweaked as described on the Mobian wiki. For other apps may be https://LINMOBapps.frama.io can help you to get an idea.

Check for AArch64 binaries if you want to run something closed.

Personally I would avoid that, so in that case it's mostly whatever is in the distro you're using plus maybe whatever the developers responsible for the port added. (ex: postmarketOS is based on alpine, but includes some extra stuff like gnome maps and phosh.)

Telegram is available for Ubports Ubuntu Touch.

It has to start somewhere, with hardware where it IS possible to run open-source software without too much reverse-engineering and fighting NDAs and so on.

Supporting them NOW means that later, what you want, will be possible. It's an investment into a better future..

Maybe you will like Librem 5 smartphone then: https://puri.sm/products/librem-5

The kind of ecosystem that Pine are trying to foster tends to produce much, much leaner software though. Of course, at the moment there are simply some apps (like Telegram IIRC) that are thin wrappers around a web app. But taking f-droid as a reference, compared to the play store, I have great hopes to end up with a less bloated stack.

Cynics would call the apps on f-droid feature poor. But for me, the FOSS apps that I use day to day tend to just do the thing you need them for, no more. There's no telemetry overhead, they tend to use fewer dependencies, are less likely to be built on heavy frameworks like electron. Of course there's heavy sampling bias involved in this since I specifically select apps that feel snappy and light. But this has always been true for linux desktop apps, too. My favourite comparison is the size of pdf viewers. I have never lacked a feature from the Adobe pdf reader. The 400KB binary of evince totally sufficed, and it used much fewer resources at runtime, too.

I really hope that this ecosystem can push the boundaries of what's possible with lower-spec hardware, even if just to hold the mirror to the big platforms.

The Pinephone thus far is mostly for tinkerers. The super low price point is what will enable the community to form. It's not really a mainstream daily phone yet.

Maybe the Fairphone is something for you? It's 3x the price but still far from premium pricing gives you a decent 4G/64G Android phone without fuss. I had their first phone which was good enough for an everyday phone and support in Lineage was great. Removable battery, 3.5mm jack, and they try to keep spare parts around for at least a new years.

So buy an android phone AND a pinephone? Without support, they won't make it far enough to bump the specs.

The amount of RAM fitted is a limit of the SoC, the vendors that do make chips that can address more than this don't provide documentation.

>2/3 GB RAM is not enough in the Android world

It's not running android, it's fine for GNU/Linux. I did a lot of homework in college on a raspberry pi with 1G of ram.

Mobian is as fast-moving as any OS on the PinePhone. Every week I see substantial improvements.

It's not my daily driver yet. Soon. Probably before I get my Librem 5, which I have already waited 2 years for, and expect will like less than the PP.

There aren't open source graphics drivers for most Qualcomm SoCs.

> Snapdragon

> Open

Because these two are incompatible.

The new SoC for 2021 is a big step up over the RK3328, though!

I've seen this explanation coming up quite often in discussions around Qualcomm or other similarly sized players: you have to be ordering (very) large numbers to be able to get their attention. Documentation is provided under NDAs I hear, and if you'd let's say, decide to buy a small(er) number of boards off the grey market (from one of their direct customers), you'd still have the problem of lack of support (firmware/code, docs, etc).

We should definitely do that, but given that e.g. EU politicians still don’t understand why crypto-backdoors are a terrible idea, I have very little hope for succeeding on that level.

Personally, while that sounds appealing, I think we also still need the RE efforts. Hardware that's built to run free software out of the box is important, and it would be ideal if all those blobs were gone and we didn't have to worry about them. I'd absolutely be on board with an effort to enforce free drivers for the hardware we buy.

But... in the short term, given that's not currently the case, we still need a way to move forward, right? And it's much easier to demonstrate why free drivers are important if you have working software to show for it. ("Hey, there's this whole phone OS and app ecosystem that we could run on this hardware if it had free drivers!")

That said, yes, the endless reverse engineering required to keep the devices you've bought secure beyond the 2 or 3 years that the manufacturer supports it is ridiculous. I wish it weren't so. But given that that's unfortunately the situation we're in, I applaud anyone contributing to the RE efforts. (But I sure am keeping an eye on hardware that doesn't require it!)

What legislation? Mandating that all OS level software/firmware has to be under FOSS license?

Mandating a release of detailed HW documentation (datasheets for all the parts + device schematics) prior to being able to sell devices would also be an interesting experiment.

> international effort to reverse closed blobs

https://news.ycombinator.com/item?id=24916086

1.) You can't just open up every driver. I say this as somebody who works on proprietary firmware and software that is basically the future of our company. We operate in a market with thin margins, our control strategies are our advantage, and divulging any info past the bare minimum concerning them would result in mass counterfeiting and ultimately the death of our company. A lot of hardware development is like this. I wish it were some other way, because I know some people would be over the moon with what it would enable, but most of the time it's just not possible.

2.) I think you misunderstand what the Pinephone is -- if my memory serves me correctly, minus the modem blob, there is no need for reverse engineering, as the documentation for the hardware is available. You're talking about sending good money after bad ideas while advocating stopping the thing that has a clear path to victory (starting from a clean slate with something that has real documentation) in favor of something else that has a negligible chance of ever happening (having source be released for drivers but not getting any documentation for it).

I think dino has patches for it to run better on small devices. the default texting app runs on libpurple so it has support for jabber, i dont know if it constitutes decent though, I have nobody to talk to on jabber :(

It does not.

What I would offer is it has 6 back bogo pins that speak I2C, so someone who wants to build it in can.

Buy a hub. Instead of 3 USB, I'd like USB 3. :D

You can already use a USB hub/dock?