zlacker

[parent] [thread] 21 comments
1. Pragma+(OP)[view] [source] 2022-08-17 14:05:38
While I’m sad to see the author leave such a project and grateful for his contributions, I have to say I think Pine64 is actually in the right here.

One of the biggest challenges in running an open source hardware project like this is catering to two very different audiences: The majority of your customers (99.9% or more) just want the hardware to get up and running quickly so they can get to their specific need or application. They don’t want to have to read endless IRC or Discord backlogs to figure out the current best distribution to use or read potentially outdated half-finished Wiki articles describing the tradeoffs of various distros. They want it to work and to get started quickly.

The open-source developers have an entirely different set of desires, preferring endless tinkering with the internals and actually enjoying the process of trying different distributions, building and testing bleeding-edge board support software themselves, playing in someone’s experimental fit branch to get one thing working, and other time-consuming activities relayed to the board itself.

If you let a project cater too much to the developer community at the expense of the 99.9% customers, it starts to become a huge problem.

For the best example, consider the huge success of the Raspberry Pi and their Raspberry Pi OS, while even the biggest competitors (such as PINE64) remain relegated to mostly obscurity. The hard truth is that if you want to make a product like this successful and mainstream, you need to narrow the focus and be ruthless about cutting costs, simplifying, and getting your users up and running with one easy, primary way to get started. I have several Pine64 products and they all suffer massively from the fragmentation and compromises they’ve made. Fun if you’re a kernel developer who spends tens of hours every week keeping up with your friends in the small developer community. Not fun at all if you just wanted to use the product for something and you realize you could spend weeks or months sorting through all of the disparate information sources and developer communities before you can have the product working enough to get started on that thing you actually wanted to build with it.

replies(4): >>detaro+t >>johnkl+j2 >>ddevau+T2 >>cycoma+Yq
2. detaro+t[view] [source] 2022-08-17 14:07:41
>>Pragma+(OP)
How does making booting alternatives harder and pissing of people developing software components they themselves want to use in their "get started quickly" distro help Pine64? Prioritizing their own efforts towards one distro doesn't have to mean harming the rest of the ecosystem they benefit from.
replies(2): >>Pragma+O2 >>megous+XI1
3. johnkl+j2[view] [source] 2022-08-17 14:15:36
>>Pragma+(OP)
You're oversimplifying tremendously and disingenuously.

Those 99.9% can go buy a phone. They don't need PinePhones.

What if the Raspberry Pi suddenly tried to be an everything computer for everyone? Now it needs a case, and a faster CPU, and expandable memory, and SSD, and a bigger power supply, and so on, until it's practically a NUC that costs $400.

Part of the success of the Raspberry Pi is that you can load whatever OS on it you want. Imagine if you could ONLY run Raspberry Pi OS!

replies(1): >>Pragma+c3
◧◩
4. Pragma+O2[view] [source] [discussion] 2022-08-17 14:17:53
>>detaro+t
When you’re trying to make a mass-market hardware product at bare-bones pricing, you have to be ruthless about simplifying the hardware and cutting costs. A single SPI chip isn’t a huge investment in absolute terms, but it adds one more sourcing complication (in the middle of a chip shortage) and creates significantly more RMA complexity (for the reasons that the Pine64 organization accurately explained from their past experience) to cater to a relatively small number of users.

Contrast this with the Raspberry Pi organization, which has seen massive success by having an uncompromising stance on simplicity and focus, even if the ideological purity of the project isn’t up to certain people’s standards. Like it or not, it’s what made them successful while projects like Pine64 continue to be niche products that require a lot of work and research to use.

replies(2): >>detaro+34 >>CivBas+Hj
5. ddevau+T2[view] [source] 2022-08-17 14:18:10
>>Pragma+(OP)
This is one of the biggest strawmen I've seen on HN, a platform famous for building them. There's a lot wrong with this but the simplest is this: Pine's stated strategy is to deliver hardware and let the community deliver software. And the community did deliver working software under the earlier community model Pine was pushing, and it's thanks to these efforts that "getting the hardware up and running quickly" is even possible. Now they're adopting a different model which completely undermines what worked about the last one.

Pine64 is making enthusiast products for hackers, not mass-market devices for non-hackers. Non-hackers have access to plenty of phones which just werk. Part of the promise of Pine's platform and the appeal to the target audience is the commitment to community.

replies(1): >>nextha+sO
◧◩
6. Pragma+c3[view] [source] [discussion] 2022-08-17 14:19:35
>>johnkl+j2
> Part of the success of the Raspberry Pi is that you can load whatever OS on it you want. Imagine if you could ONLY run Raspberry Pi OS!

That’s not the issue with the Pine64 ecosystem, though.

The direction they’ve chosen is actually similar to what Raspberry Pi has chosen: You can boot alternate OSes, but the primary focus is Raspberry Pi foundations own needs and everything else comes secondary. This is what it takes to keep a project like this alive, and they know it.

replies(1): >>yjftsj+jG
◧◩◪
7. detaro+34[view] [source] [discussion] 2022-08-17 14:23:20
>>Pragma+O2
I kind of doubt that adding an extra chip was the only way to preserve the ability to boot from microSD, given earlier hardware revisions did do it too - i.e. its something they decided to take away. (RPi btw nowadays can boot from multiple sources too ;))

And even then, it's only one of the complaints, and this isn't the only places I see backlash against pines treatment of the dev community - again for a company that relies on said community for a lot of software work across their products. My impression is that RPi foundation with Raspian relied a lot less on the community.

◧◩◪
8. CivBas+Hj[view] [source] [discussion] 2022-08-17 15:39:30
>>Pragma+O2
> When you’re trying to make a mass-market hardware product at bare-bones pricing, you have to be ruthless about simplifying the hardware and cutting costs.

But is that really what PINE64 should be trying to do? So far their support hasn't come from the "mass market". It's come from a niche market of open source hackers trying to build and support various Linux distros for mobile devices. Why does improving mass market appeal have to mean alienating your existing supporters?

replies(1): >>nextha+aJ
9. cycoma+Yq[view] [source] 2022-08-17 16:12:36
>>Pragma+(OP)
While I can agree that if you want to make a device with mass appeal you probably should focus your efforts, this doesn't seem to be happening. As far as I can tell Pine does not put any significant effort into the software, but instead hopes that the community will do it. Now if you rely on the community, focusing only on a minority (manjaro users) of a minority (Linux mobile developers) of a minority (phone users wanting a Linux phone) might be a bit too much focus I'd argue.
◧◩◪
10. yjftsj+jG[view] [source] [discussion] 2022-08-17 17:22:34
>>Pragma+c3
> The direction they’ve chosen is actually similar to what Raspberry Pi has chosen: You can boot alternate OSes, but the primary focus is Raspberry Pi foundations own needs and everything else comes secondary.

If I put a SD card with ex. Alpine Linux into a pi, it'll boot into Alpine Linux. If I put a SD card with Alpine Linux into a Pinebook Pro, it'll boot into... Manjaro.

replies(1): >>zozbot+7I
◧◩◪◨
11. zozbot+7I[view] [source] [discussion] 2022-08-17 17:30:39
>>yjftsj+jG
I don't think the Raspberry Pi does UEFI boot by default? You'd have to put in a SD card with Raspberry Pi-specific boot support, including proprietary blobs.
replies(1): >>yjftsj+7J
◧◩◪◨⬒
12. yjftsj+7J[view] [source] [discussion] 2022-08-17 17:35:47
>>zozbot+7I
Yes, that's true; my point is that the pi will happily boot whatever, unlike the Pinebook Pro which has a default bootloader on internal storage that ignores the SD card. So a distro/OS has to add the pi-specific bootloader to their SD image, and that's a pain, but they can't do anything to make the PBP work, because the machine won't even pay attention to their SD image.

EDIT: Actually I guess the Pi 4 added an onboard flash chip with an early bootloader, but I can't figure out if it impacts the boot order or changes how hard SD boot is: https://www.raspberrypi.com/documentation/computers/raspberr...

replies(1): >>megous+nJ1
◧◩◪◨
13. nextha+aJ[view] [source] [discussion] 2022-08-17 17:35:53
>>CivBas+Hj
You come up on your niche, and then when you have access to the broader market, you pivot to the group that will help you grow market power [1]. Similar dynamics exist in a lot of different ecosystems, and Pine seems to be responding to the challenges that have come with becoming big. It's sad that they won't be supporting OS hackers anymore, but they have to pivot if they want to bring onboard more customers (which seems to be the goal behind this decision).

[1] https://www.cgpgrey.com/blog/rules-for-rulers

replies(1): >>CivBas+fO
◧◩◪◨⬒
14. CivBas+fO[view] [source] [discussion] 2022-08-17 18:02:41
>>nextha+aJ
Can you really consider the "mass market" to be a key to power for Pine64 with their current lineup of products, though? The OS hackers seem to be Pine64's only key to power right now. And does supporting them really consume extra resources that could otherwise be better allocated?
replies(2): >>zozbot+7P >>nextha+bR
◧◩
15. nextha+sO[view] [source] [discussion] 2022-08-17 18:03:54
>>ddevau+T2
> Pine64 is making enthusiast products for hackers, not mass-market devices for non-hackers. Non-hackers have access to plenty of phones which just werk. Part of the promise of Pine's platform and the appeal to the target audience is the commitment to community.

Sounds like either Pine64 has grown past this and decided to pivot, or has been losing revenue due to a lack of customers from this niche market. Personally, as a hacker I love playing with different OSes. However, if I was to use any open source device like a PinePhone or Pine64 board to build something, I'd prefer a stable environment backed by an established foundation. Environment setup is hell, and figuring out which open-source OS works best, if it will be supported in the future, and how to install it would slow me down immensely.

◧◩◪◨⬒⬓
16. zozbot+7P[view] [source] [discussion] 2022-08-17 18:08:33
>>CivBas+fO
> Can you really consider the "mass market" to be a key to power for Pine64

Not as long as they ship Manjaro as the default OS for their hardware. Rolling-release distributions are not fit for mass-market use.

◧◩◪◨⬒⬓
17. nextha+bR[view] [source] [discussion] 2022-08-17 18:20:11
>>CivBas+fO
My guess is that OS hackers could be considered as one niche, while hackers and builders who prefer not going through a custom Linux install/config for their project (i.e., a weather station or a mobile smart home dashboard) could be a larger one. Definitely not "mass market" or replacing Android levels, but at the same time a significantly larger portion of revenue for Pine64. The switch to Manjaro would provide them with a key backer that allows them to unlock this market. People have been discussing the software quality of Manjaro, so maybe it has a good foundation or connections?

Also: I've seen some hidden costs of supporting custom OS installs being discussed, i.e. procuring extra chips to allow open boot. This may have factored into Pine64's decision.

replies(1): >>detaro+HJ1
◧◩
18. megous+XI1[view] [source] [discussion] 2022-08-17 23:31:55
>>detaro+t
Tow-Boot is pre-flashed on SPI on Pinephone Pro. How is that making alternatives harder? (aside of course for people wanting to use an alternative bootloader, like levinboot)
replies(1): >>zozbot+SO1
◧◩◪◨⬒⬓
19. megous+nJ1[view] [source] [discussion] 2022-08-17 23:34:59
>>yjftsj+7J
I haven't seen U-Boot in a while (I use levinboot on PBP), but if it has LCD and USB support already on PBP, then the user should be able to just self-erase the bootloader from eMMC using a command or two and then be free to boot from SD card or USB. Or just erase it from the booted default Manjaro.
◧◩◪◨⬒⬓⬔
20. detaro+HJ1[view] [source] [discussion] 2022-08-17 23:37:32
>>nextha+bR
> while hackers and builders who prefer not going through a custom Linux install/config for their project

That's a false dichotomy, nobody is demanding that users be forced through "custom Linux install" (whatever that means). The problem is also not primarily that Pine64 have chosen a "flagship" distro, but how they and said distro behave towards the other options. I'm sure the quality of the flagship distro is massively improved by making life hard for the project that did useless things like making the camera in the phone work...

◧◩◪
21. zozbot+SO1[view] [source] [discussion] 2022-08-18 00:14:10
>>megous+XI1
The discussion is mostly about PineBook Pro by now, where the SPI was left unflashed.
replies(1): >>megous+BU1
◧◩◪◨
22. megous+BU1[view] [source] [discussion] 2022-08-18 01:01:54
>>zozbot+SO1
Well, that's even easier then to rid of the bootloader from just eMMC and have it booting whatever from SD card.
[go to top]