Live data from Hacker News

Librem 5 dev kits are shipping

puri.sm

91–98 of 98 posts

Re: Librem 5 dev kits are shipping

#91
post #72

Earlier quoted context omitted.

a.) This is all very much WIP. b.) As long as you develop a standard GTK app for Linux, it should be easy to adopt it for Librem 5. c.) Are you seriously comparing their WIP docs to someone who has put billions of dollars and over a decade into their platform and has a JetBrains powered IDE?

Is GTK intended to be the programming environment aimed at typical developers? Writing object oriented code using lots of boilerplate C macros seems quite a bit more tedious than using a language designed for object oriented programming from the ground-up. At the very least will GTKmm be a first-class citizen?

There are GTK bindings for Python, JavaScript, Rust etc. It does not have to be C.

In fact it could be Qt too, that'll run perfectly fine too, if GTK isn't your thing.

Re: Librem 5 dev kits are shipping

#92

I have one of these on the way. The main surprise for me is they're shipping it without the lcd panel working. They can control the backlight. But no video. Now I'm participating in the gamble that can be made to function without rework, I'd at least like complete details, not "we are talking to the panel vendor".

That's why the support channels were setup. Also, there still isn't a lot of details to share in the first place. The problem comes down to sending the initalization codes to the panel, which for some reason is not working. The conversations continues. Now you're as up to date as the rest of the librem 5 team. I didn't want to drop into overly technical detail as not everyone wants that. You are of course welcome to…

One of the unusual and desirable things about this is your customers for the dev board will largely be very experienced... you acknowledge it in the letter you sent out.

What surprised me is until it's fixed, there's no certainty what the problem is. Eg, dsi lanes mixed up, polarity of dsi lane wired wrong, some semidocumented reset signal on the panel got forced to 0V or forced inactive but panel requires it strobed, psu to panel must come up in specific order, psu level or quality or wiring mismatch... there are many possible ways just from having no video it can be a hw rework issue to solve it. Of course dsi is also complex, it can easy be a sw-only config issue, or some combination.

It's not up to me as your paying customer to go ask on a 'support channel' why, it's up to you to explain the reasoning why you believe that's a software-only issue.

Re: Librem 5 dev kits are shipping

#93

Earlier quoted context omitted.

The librem5 may as well be subtitled "The GNOME phone", I think it goes without saying that it will use systemd.

You are free to install any operating system on it that you like. Just like your desktop. We will probably also see 100% free android versions (like Replicant) to be released for this phone when it's out. But you could also install another GNU based Linux system of course, or just another DE.

Indeed, that's part of the value proposition for such a device.

However, it's obviously preferable to want/use what's being shipped on the device to all recipients.

In this libre-oriented space in particular, what users are likely to want as a priority is the ability to easily reproduce the bits shipped on the device from source, and be able to restore the device to a state no different than shipped from the factory using those self-reproduced bits. The freedom to run some hacked up half-baked alternative stuff they're technically free to install should they wish, is not the top priority.

A major component of the value conferred by this freedom is that those shipped, reproducible bits, contain a desirable foundation for the community to converge on and iterate from.

Telling people "you can install whatever you want" before even delivering is signaling "potentially unusable, controversial vaporware is shipping, community fragmentation ahead" to anyone paying attention.

I'm not particularly averse to systemd/GNOME, but I do fear that it's probably too immature an ecosystem for production mobile use and it's unnecessarily threatening the project's overall success.

It would have been more prudent to collaborate with the Jolla/Sailfish folks and ship basically a secure Nokia N9 successor but with an easily reproduced and flashed image, which the N9 lacked, while the GNOME community grinds away on getting their stack mobile-ready for a potential future device. I don't know if the licensing could have been hashed out to get the N9 stack fully available in reproducible form, but I'm inclined to assume that if there were components legally obstructed from such distribution, it probably was less work to reimplement them than build all the components needed for a GNOME phone.

Re: Librem 5 dev kits are shipping

#94
post #69

Earlier quoted context omitted.

I agree with you on all counts but in order for the phone to be supported for a long time (both by Purism and by third parties) it needs to be reasonably successful. Otherwise nobody is going to even remember it existed two years from now, much less port and maintain apps for it. That's my main problem really, I have Spotify, WhatsApp and a couple other closed-source applications on my smartphone that I need in order…

That's sort of the issue the OpenMoko faced. It was designed to be a device with long-lasting support - but support died quickly because so few people had one.

TBH community distros were pretty long lived (and let's be honest, those were the only ones that mattered). At least SHR was actively maintained to around 2012, with last generated image being from April 2013, and I seem to remember QtMoko being maintained even tad longer. IMO that wasn't bad for a device released in 2008, especially for today's standards.

Re: Librem 5 dev kits are shipping

#95
post #84

Earlier quoted context omitted.

> an iMX8M laptop 8M is really weak. That would be pretty much a Pinebook with Vivante graphics. The i.MX 8 QuadMax at least has two A72 cores, similar to Rockchip RK3399.

Have Purism explained their processor choice anywhere? It would have been nice to have used the i.MX 8 QuadMax - it would have given a little more hope for the phone lasting.

I'd guess that power consumption characteristics of those two i.MX8 variants are different enough to explain that choice.

Re: Librem 5 dev kits are shipping

#96
post #61
post #49

Earlier quoted context omitted.

I'm impressed and very much respectful of what the Purism folks are trying to achieve but I just don't think their business plan is viable as it is. As such I really don't really believe that getting a few geeks on HN to support the project will really make much of a difference. They need a much broader market share if they want to have a chance to survive in the long term and I don't see how they can get it. As it s…

The difference could be : a phone that you're not supposed to throw away after two years. Something that would be maintained over the long run. Lots of people would buy a phone that can last 5-7 years. I am one of them. My mother is one of them. My father too. Not everybody is fond enough of phones or tech to want to change often. If it runs apps and has a good camera, it's fine. A phone that lasts 5-7 years would al…

Everyone that I know has phones that can last 5-7 years, basically they only get replaced when they die or get stolen.

Most non-technical users don't care about updates, their care about their low prices for pre-paid phones and the respective monthly expenses.

Re: Librem 5 dev kits are shipping

#97
post #84

Earlier quoted context omitted.

Have Purism explained their processor choice anywhere? It would have been nice to have used the i.MX 8 QuadMax - it would have given a little more hope for the phone lasting.

I'd guess that power consumption characteristics of those two i.MX8 variants are different enough to explain that choice.

Well, the mainstream phones have much more powerful processors than Cortex-A53, they just power manage them well…

Re: Librem 5 dev kits are shipping

#98

Earlier quoted context omitted.

I'd guess that power consumption characteristics of those two i.MX8 variants are different enough to explain that choice.

Well, the mainstream phones have much more powerful processors than Cortex-A53, they just power manage them well…

I wouldn't expect a SoC that's not even targeted for mobiles to "power manage them well". Mainstream phones are completely irrelevant there.
Post reply on HN