@amosbatto Would love to hear your opinion on it as always, you’re take on these devices is really interesting
this yt-video is what comes in my mind when I hear PinePhone and I know a lot of struggling with their PinePhones.
Kind of glad I went with L5!
I find the video dubious and lurid. If he had received a Librem 5, the reaction would probably have been similar or worse, since Purism had really advertised the Librem 5 for end users, unlike Pine64.
By the way there is already a thread on the PPP: PinePhone Pro just announced (Oct 2021)
The PinePhone Pro (PPP) is going to have significantly better CPU performance than the L5, due to its two extra Cortex-A72 cores. The T860MP4 GPU in the RK3399 lacks Vulkan, but the L5 can’t support Vulkan either due to the lack of support in the free Etnaviv driver, and the two GPUs should have roughly similar graphics performance. The LPDDR4-3200 DRAM in the L5 is twice as fast as the LPDDR4-1600 DRAM in the PPP, but other than that, the PPP is going to equal or better the L5 in terms of performance, and it is going to be 4.4 mm thinner and 45 grams lighter.
The RK3399S has hardware video encode (although not in H.265) and an image signal processor, which are both features that the i.MX 8M Quad in the L5 lacks. I expect that the PPP will be significantly faster than the L5 at capturing images and video and run cooler and use less battery when filming, since it doesn’t have to process video in software.
The RK3399 is a power hog that generates a lot of heat and I thought that it was impossible to make a phone with it, but Rockchip created a special power efficient version of the chip for PINE64, which surprises me, because I didn’t think that PINE64 has enough volume to make custom chip orders. Apparently PINE64 thinks that it only needs a 3000mAh battery, which also surprises me, but they are downclocking the CPU from 2.0 to 1.5 GHz, downclocking the GPU from 600 to 500 MHz, and downclocking the RAM from 933 to 800 MHz, so it may be possible to keep the PPP inside the same power envelope as the PinePhone.
Another thing that surprises me is that PINE64 managed to cram the RK3399 which is 21x21 mm (441 mm2) inside the same size of PCB as the A64, which is half the size at 15x15 mm (225 mm2). Looking at the PinePhone’s main PCB, I don’t see a lot of empty space to stick a 21x21 mm chip, so they probably had to do a full redesign of the board to make space, and that probably means that PINE64 did things like switching to BGA packages and using smaller 0201 and 0402 packages which make it harder to tinker with the hardware. Of course, the L5 already did many of these tricks to cram 1200 components on its two boards.
My take after a quick glance at the specs is that the PinePhone Pro (PPP) is bad news for the Purism, because it will match or exceed the L5 in almost every area. I thought that Purism would have a couple years where it could justify its high prices until the PinePhone 2 with the RK3566 arrived, but now it is going to be very hard for Purism to market the L5 at $899, much less the planned increase to $1199 on Nov 1, 2021 and $1299 in March 2022.
I really think that that Purism needs to rethink its pricing strategy, because the PPP is going to offer better convergence as a desktop PC (due to better CPU performance, more RAM and more Flash storage). Many of the areas where the L5 was better than the PinePhone (cameras, CPU/GPU/VPU performance, USB port speed), the PPP will be equal to or better than the L5, and the RK3399 has very good Linux support, so it is going to be very hard for consumers to justify paying a premium for the L5.
I assume that the AMPAK/Broadcom WiFi in the PPP will need proprietary firmware stored in the Linux file system (in /lib/firmware), since all 802.11ac chips on the market need it, and I doubt that PINE64 is willing to go to the extreme of adding an extra SPI NOR Flash chip to store blobs just to meet RYF requirements like Purism. The last I read, it was possible to boot the RK3399 with only FOSS, but it doesn’t run well, so I assume that the PPP will include some blobs like the other RK3399 devices that PINE64 sells.
In other words, Purism will have to sell the $800/$900 difference between the PPP and L5, based on free software and better hardware kill switches that are accessible and can turn off all sensors. I assume that PINE64 is planning to sell a retail version of the PPP for $100 more, just like the PinePhone, with a 1 year warranty and software support from Manjaro, so even that advantage of the L5 is going to disappear.
There are some free software and security/privacy diehards who are going to look at the L5 at $1299 and the PPP Retail at $499, and decide to pay the $800 difference. There are also a few like me who really want to support Purism’s software development and will also pay the difference, but the L5 is going to be a very hard sell for Purism in the future.
My recommendation to Purism is that it not increase its prices, or if it must increase its prices, it has to match the PPP by including 4GB RAM and 128GB Flash storage. We are probably talking about a $25 increase in its bill of materials, so this is entirely doable on Purism’s part, but really I think that Purism should increase the storage to 256GB, so that it can market the L5 as a better convergent PC than the PPP.
This situation annoys me. Purism needs to be able to charge high prices to be able to pay for its software development. In my opinion, Phosh has the best chance out of all the available Linux interfaces of actually getting mobile Linux to mainstream users. The software development that Purism is doing is vitally important for the future of mobile Linux, and it needs to be supported. PINE64 could have given Purism a couple years to market to the higher end of the market before it released the PinePhone 2 with the RK3566, but now it has taken away a large portion of Purism’s market. I wouldn’t mind if PINE64 was paying for software development to get mobile Linux good enough to reach the mainstream, but PINE64 isn’t paying for software development in any meaningful way.
Maybe PINE64’s hardware will attract enough volunteers to the Plasma Mobile project, that it will start improving at the same rate as Phosh has, but I am skeptical that Plasma Mobile is the vehicle to take mobile Linux to mainstream users, for the reasons that I laid out in this blog post. I wish that PINE64 would charge a higher price for its hardware and use the profits to pay developers, or at least leave the higher end of the market for companies that are willing to pay developers.
Nonetheless, I can’t get too mad at PINE64, because it is just taking advantage of the fact that Purism has been pricing itself out of the market and taken so long to deliver. PINE64 is going to bring a performant phone to market at a mid-range price that many more people can afford, and will dramatically grow the total number of mobile Linux users, which will help bring more volunteers to Phosh, Plasma Mobile, UBports, etc.
Experience has shown that volunteer labor only gets FOSS so far, and rarely produces software that appeals to mainstream users. Most of the heavy lifting at projects like the Linux kernel, LibreOffice, Firefox, Chromium, PHP, MySQL, MariaDB, PostgreSQL, Apache, GNOME, etc. are actually done by people whose work is paid by an employer. A good example is how many of the chief contributors to PostgreSQL are employees of EnterpriseDB, Crunchy Data, Microsoft, NTT and VMware.
If PINE64 gobbles up the entire Linux phone market, then we are left with just volunteer labor to develop mobile Linux. However, I have to admit that KDE is one of the projects that has managed to flourish with almost exclusively volunteer labor (aside from a few paid devs from Blue Systems and Qt being provided by the Qt Company). It is my honest belief that Plasma Mobile is not as good of a vehicle as Phosh to bring mobile Linux to the mass market, however, if Plasma Mobile can manage to do it with only volunteer labor, then I will be the first to cheer.
The long-term success of mobile Linux isn’t just some vanity project for geeks. The kind of Surveillance Capitalism being promoted by Google may end up laying the seeds for a dark Orwellian future, and it is certainly destroying the concept of personal privacy as we know it. Planned obsolescence is hastening climate change and the depletion of vital natural resources. The current mobile industry is promoting a dark future based on locked-down spy devices that we don’t control and eroding digital rights, and the most effective means to prevent that dark future is to create an alternative in mobile Linux that eventually gets good enough that it can appeal to ordinary people without technical skills.
The question is whether we can create that alternative based solely on volunteer labor and devices like the PinePhone, Volla Phone and Pro1 X, which don’t pay for the development of mobile Linux. I think that we have much better probability of success if we have some paid software developers, which means the Purism has to be able to make profits, because I don’t see any other company willing to pay developers except Jolla, and it is clear to me that the community will never adopt Sailfish OS and its proprietary Silica interface, nor will it ever make it to the mainstream.
Unfortunately that’s my fear as well. I want purism to succeed, as from the software perspective they seem to be the main driver for linux phone software.
Granted, unless you’re a developer you can’t get a pinephone pro yet, but I’m sure that will change in the next few months.
I stay with Librem 5 because it has texas cpu, truly modular, truly HKS, truly longevity, truly more…
Librem 5 next generation Fir should have been shipped in “shipping window Q4 2020”. Purism needs a hardware update (at least more RAM and Flash) and a price downgrade. But I think the former will be delayed at least for two years, the latter will not happen any time in the next years. I stick with Librem 5, because they do software development, but it gets very hard.
correct me if I am wrong, but isn’t the Librem 5 built to be modular? I mean in the future could your current Librem 5 be updated with swappable hardware in the board? I mean its not the FairPhone level of modular, but aren’t there certain parts that can be upgraded.
Anyway I own a PinePhone and a Librem 5, both I bought off of EBay, still waiting on my original Librem 5 order from Pursim. Pinephone is fun to play with in testing out different distros, but so far none of the distros are as usable as what Purism has done with their base hardware/software. Not saying the Librem 5 is necessarily daily driver ready yet, but its closer than the PinePhone as of writing this. My biggest drawbacks to the Librem 5 right now are battery life and texting, fix that, I will make it my daily driver.
Yes the BT-WLAN. GSM. can be upgradable.
The L5 is modular to some degree. Its battery can be replaced without using an tools. It’s easy to disassemble and uses standard Phillips screws and no glue or soldering is required. Its two M.2 ports make the WiFi/Bluetooth and cellular modem replaceable and upgradable, and it has a extension port on the back for adding mods. It has a separate board for the USB port, which makes it cheaper to replace if the port gets damaged, but the frame and display are fused, so it costs more to replace if the screen gets cracked.
Purism says that it plans to sell replacement parts, but so far it only sells replacement batteries and cellular modems. The big question is whether Fir will be designed to be backward compatible with Evergreen, so that Evergreen’s main PCB can be upgraded to Fir’s main PCB.
The only phones that I know of which were truly upgradable were the SHIFT5/5.1/5.2/5.3 that were designed to backward compatible, so a SHIFT5.1 could be upgraded with a SHIFT5.3’s board.
In my opinion, most of the Fairphone’s modularity doesn’t make much sense, since the only parts which are upgradeable are the cameras, and most of the modules are parts which are very unlikely to break and need replacing. Making a phone modular actually adds to the economic and environmental costs of a phone, so it only makes sense to make something modular if it is upgradeable or if it liable to break and need replacing. For these reasons, I’m more impressed by SHIFT than Fairphone as a company, but Fairphone gets all the attention.
Still no realistic estimation of the delivery date of my Librem5 despite the announcement of last September (it could be more than a year from now considering the catastrophic management of the CPU supply by Purism) and now the Pinephone pro which is more powerful, has a CPU adapted to a mobile use, has dimensions closer to the current standards, with a Gorilla glass screen and available for less expensive from early 2022.
I’m still waiting for the next few months, but if I still haven’t received anything by the first quarter of 2022, I’ll give up my Librem for a Pinephone.
Fairphone being modular might serve another purpose though, like I am not sure which models are 5G ready, if any, so in the future if 4G is ended by carriers, perhaps the Fairphone can be upgraded to be a 5G device? Although Librem 5 might be able to get those kind of hardware replacements too? Also imagine a future where we could swap phone parts like a gaming PC. It could be like a completely new hobby rather than buy a new phone every couple of years.
4g can’t be ended by 5g as 5g requires 4g and 5g is only practical over short distances(ie cities only).
There is no way to upgrade the Fairphone 3 to support 5G, without replacing its main circuit board, because the cellular modem is part of the Snapdragon 632 processor, which is soldered to the circuit board. There is nothing in the Fairphone 3 which is currently upgradable except its front and back cameras, and I suspect that Fairphone purposely designed them to have weak specs in the original 3 model, just so it could offer upgrade modules.
The only way to get significantly better cameras in most phones is to upgrade the SoC at the same time that you upgrade the image sensors, so modularity inside a phone doesn’t make much sense in my opinion. Realistically, if you want to upgrade anything significant in a phone, such as memory, storage, port speed, etc., then you need to replace the whole circuit board, which is why SHIFT has the right idea of making backward-compatible circuit boards that can fit in previous models. Unfortunately, the screen sizes and aspect ratios of phones changed, so SHIFT couldn’t continue its SHIFT5.x series, but the current phone form factor with a 6" 9:16 screen and a USB-C port is likely to be stable for at least the next five years (until folding OLED gets cheap enough), so it should be possible to design phones with upgradable main boards.
A truly modular phone with many upgradeable parts Iike in PCs isn’t possible in my opinion, because almost none of the available chips are designed to be used in sockets, and we would need a smaller class of connection standards, than SODIMM for RAM and M.2 for SSD, WiFi and cellular modem. One solution is to use an upgradeable SoM (system on a module) like the MNT Reform does, but all the SoM’s that I have seen are too big for a phone (although they could work for a tablet).
I think the L5’s design is truly remarkable, but when you consider the fact that it requires 2.5 times more electronic components than the standard smartphone and is twice the thickness and 35% heavier, it seems far more practical to me to make phones where the main board is replaceable and upgradeable, than using upgradeable parts inside the phone. To get higher resolution cameras, a higher resolution screen, faster USB port, more than 4GB RAM or better image/video processing in the Librem 5, the SoC will have to be replaced, so most upgrades would require replacing the main PCB and modularity is of limited utility in mobile phones. The one component where modularity makes sense in my opinion is in cellular modems, because it allows a phone to use the LTE bands in different parts of the world, but that can be solved by using a global modem and the PinePhone community has already done the hard work on the EG25-G driver to make a global modem work well. The cooling and processing requirements of 5G will make it hard to put a 5G modem on an M.2 card, so I’m not sure if we will ever be able to buy an M.2 G5 modem for the L5, although we probably will be able to upgrade the WiFi card to support newer standards that require proprietary blobs to use.
Is there any way that Purism could create a new software license for paid software development work that Purism pays for in the development of the Librem 5? Such a license could still allow the software to be open for review by the public. But it would only seem fair that Purism prevent its competitors from using their own software to compete with them. Let them open it up fully after they reach shipping parity. But it seems like they should prevent other phone makers from using their own software on shipped product until at least, they get to use it fully themselves. It just seems wrong for one company to pay the expenses so that another company can cash-in on the other company’s work. What Pine is doing appears to risk my Librem 5 pre-order by siphoning-off Purism’s profits, possibly to the eventual point of bankruptcy. Myself and others like me who have pre-orders are paying and taking the risks to help Pine cash-in.
I think the general question of more restrictive licensing has been discussed before.
Isn’t the above-quoted text the essence of open source? Someone does the work and incurs the cost and in some way isn’t reimbursed for the cost - and everyone gets to benefit.
Even Richard Stallman said “free, as in freedom of speech, not free beer”. If the goal is to keep the source code open for public examination up-front and indefinitely, then most of Stallman’s concerns would seem to be addressed. A temporary license restriction to prevent others from cashing-in for a reletively brief time seems reasonable. Ultimately, after some recovery of expenses, the code could drop in to the regular GPL. Perhaps future crowdsourced projects might at least consider temporarily restraining outside commercial interests from using the code to compete with its authors, at least until shipping parity is reached.
When I look at what Pine is doing to Purism, it’s more like “free, as in freedom to not have any respect what-so-ever for those who wrote the code”. Anything can be abused.
Nope. Free as in freedom means you have the right to read, share, modify and share modified copies of the software. Without those freedoms you are not free. Deciding who can or cannot share the software violates the principles of software freedom. Asking a “reasonable price” for source code is fine, as long as who bought it is free to share it gratis, if he decides to.
They are doing what they are allowed to do. And remember that if they need to change the code, they have to share the changes under the same (e.g. GPL) license the software is released with. So, everybody profits. That’s the advantage of using licenses like the GPL.
It can, and it probably eventually will, because
this is not true. High frequencies (NR) is only part of the standard. The rest reuses 4G freqs and covers similar use cases.
Not to mention that would destroy Purism’s image, either as “free” or “funded.”