Hacker News new | comments | ask | show | jobs | submit login
Librem 5 Hardware Update (puri.sm)
186 points by fghtr 18 days ago | hide | past | web | favorite | 110 comments

My current question is - will there be silicon cases and tempered glass screen protectors available? I've pre-ordered, and I don't expect the sturdiest build quality, so I'd want to have some tools to protect agianst it. I'm not meaning any offense towards Purism and their craftsmanship, it's just that the first iterations of projects like these have more important things to optimize for than hamfist-compatibility.

For glass, it's different, but if it's glass, you can always cut from a bigger glass.

For cases, 3D print one, for comfortably less than 200€ you can get a Open-Source Hardware Certified Ender 3, and that's with DHL shipping already.

You can't cut tempered glass, you can cut then temper though thats not really a diy option.

Laminated glass is an option.

Didn't knew, thanks.

Phone cases printed from regular plastic aren't that great in my experience, you really need to use plastic with some flex (like TPU). The Ender 3 can print flexible filaments to some degree, but it's not very good at it without some significant modifications to constrain the filament path.

> For cases, 3D print one,

I know someone who 3D-printed a phone case with a consumer printer he got for Christmas, and it fell apart within a few days.

As a hobbyist that built my own printer I assure you he could've printed it at a better quality then that. Prints can vary wildly in quality, and the price of the machine is largely irreverent (1000$ printers can easily match a 10,000 $ in quality in many cases). He may not've had the most optimal settings for the print.

That said I don't think 3D printing phone cases is worth doing for any reason other then the cool factor. I'd expect 3rd parties to start selling cases for it fairly quickly, there just might not be a huge range of colors/options/prices

> I know someone who 3D-printed a phone case with a consumer printer he got for Christmas, and it fell apart within a few days.

Then that's his fault, because millions of people made durable flexible material things with their consumer 3D printers.

A bad workman always blames his tools...

My biggest worry would be that they don't seem to have chosen a modem for the phone.

they can't find an open modem? that would not surprise me....

Unless you're worrying that they will be shipping without a modem... what's the problem?

It's just that I have seen projects delay making an important decision by rationalizing that there is not enough information and that the decision can be made later.

Then it turns out that the vital piece which fits the requirements does not exist at all (or does not exist for small time customer), or is incompatible with the architecture which has a lot of resources invested in.

Not saying that this will be what happens and most likely Librem team are already well aware of this risk, but it is certainly a possibility.

Yes, just on power evelope alone so many unknowns that could influence other parts of a ‘done’ design.

I wish them well but I’m sceptical. Still, it would be great to be surprised.

The problem is not knowing if it will work in my region and on my network. It's a rather big caveat for something with open preorders.

Based on playing with my little adafruit cell chip I have, I can talk to it pretty simply by sending it standard modem commands. I don't know how this exactly lines up with a more modern phone but it's pretty cool. You could theoretically talk to it with something like this too


I hope they are able to create a free RIL for whatever modem they choose (in time).

That's actually not a big deal, as there are already multiple to choose from. FSO, oFono, ModemManager, and even more if you count ones from Openmoko times. Most of them should work with a module like PLS8 with minimal or no modifications.

I want the Purism 5 to be a good 1st gen phone; I am not expecting Apple-level build quality or the newest hardware. But even with what I consider reasonable expectations, I am very uncertain with purism's ability to reach them. Making even just a decent smartphone is really difficult, and purism appears to be struggling already. I wish them the best, but I won't consider the phone until monthes after laumch.

I put up the cash as soon as it was launched. As someone who has funded many “Linux on an X” projects over the years, my expectations on timelines and polish are very low.

I don’t expect this device to be good enough to replace my iPhone immediately, but I desperately want something running Free Software that can, in the medium-term. I’m very very happy to fund and hack on the project in the meantime.

This was my position as well. I haven't pre-ordered the phone but the closer it gets to April the more I'm excited to have a freedom respecting phone. Even if the reviews are bad I'll probably still buy one to hack against and contribute back to.

This hardware update does not appear to contain any update on the hardware.

Any update that doesn't say "delay" is an update in and of itself.

It does link to an FAQ page on specs, which was news to me: https://puri.sm/faq/what-are-the-phone-specs/

I'm considering preordering this, but I'd like some confirmation it will work with my Verizon account. The FAQ says:

  The Librem 5 is an open network phone, not locked to any particular network.
  We aim to support 3G and 4G for the most common international frequency bands and carriers, with an interchangeable module. Exact specifications will follow as we are evaluating the data+voice modems that will be used.
  We expect it to work on 3G and 4G networks, and hope for it to also work for GSM, UMTS, or LTE-based network services.
which is not quite convincing enough.

"The Librem 5 is an open network phone, not locked to any particular network.

We aim to support 3G and 4G for the most common international frequency bands and carriers, with an interchangeable module. Exact specifications will follow as we are evaluating the data+voice modems that will be used.

We expect it to work on 3G and 4G networks, and hope for it to also work for GSM, UMTS, or LTE-based network services."

Just because it's technically compatible doesn't mean the network operator will allow the device, does it?

I remember hearing Jeff Jarvis' story about trying to get Verizon to let him use a Nexus 7 tablet. IIRC was compatible with the network but Verizon dragged their feet for a long time.


T-Mobile usually have no problem with custom devices. Avoid Verizon if you can.

Verizon still has by far the best coverage in many parts of the US, so it may be the only option if that's important to you.

That was the case in the past, but T-Mobile were actively building up their network in the recent times so it should be a lot better than before. I didn't compare it myself, so can't say how much it has improved in areas where it was bad before.

I switched from Verizon to T-Mobile a couple years ago, and where I live (Oregon), T-Mobile's service is absolute garbage compared to Verizon's. It was fine in the dense parts of the city, but anything outside of that was a total gamble. Driving to the mountain, the beach, or really anywhere outside the metro area nearly guaranteed a lack of service all or most of the way. I even had poor reception in parts of my house, and would sometimes lose calls in my kitchen!

I recently switched to AT&T and am much happier, but there are still occasions where my wife (on Verizon) has service but I don't. I've also heard similar anecdotes from people I've met in other parts of the country, so I don't think Oregon is the only state for which this is true.

(Before you say that this was a problem with my phone, I changed phones once while I was still on T-Mobile, and the service did not improve.)

I live (Oregon), T-Mobile's service is absolute garbage compared to Verizon's

It's all regional. There are places in America where T-Mobile is streets ahead of Verizon.

The upper Midwest, and the Gulf Coast, for example, because T-Mobile bought VoiceStream, and PrimeCo, which were dominant in those areas years ago.

Between personal and work devices, I carry AT&T, Verizon, and T-Mobile with me almost everywhere I go across America (usually two or three different states a month). Verizon is best in populated areas. T-Mobile is best in lower density places. AT&T is something of a crapshoot. Some places it's incredibly good, and some places it's like trying to make a phone call on an angry hedgehog.

> There are places in America where T-Mobile is streets ahead of Verizon.

Nonsense, it's not even a comparison. Tmobile is complete garbage coverage compared to Verizon or AT&T. I don't believe there is any place in the US where Tmobile has better coverage than Verizon.

I don't believe there is any place in the US where Tmobile has better coverage than Verizon.

All I can suggest is you get out more. And carry one of each device with you simultaneously, as I do.

How do you know?

Yep, exactly. So if you happen to live somewhere that's only well covered by Verizon, and consistently good reception is important to you, you don't really have a lot of other options. It's similar to the situation with ISPs in the US.

Most phones/tablets are a little different of a story, because the chip is integrated with the phone. The Librem 5, using a separate removable module for LTE, is going to be a bit more like a connected laptop. Verizon doesn't certify laptops for their network, they certify the cards/modules that you put in a laptop.

If you use an approved module, there should be no issue: https://opendevelopment.verizonwireless.com/design-and-build...

I had a phablet on AT&T for years - a Samsung Galaxy Tab 2, which I ordered from Indonesia. I didn't have any issues with it at all. I didn't even contact them, I just swapped the SIM from my existing phone to my new tablet and that was that.

I'm in the same boat, waiting on Verizon support. But as a note, Verizon has a bunch of pre-approved modules (https://opendevelopment.verizonwireless.com/design-and-build...) and as long as you can pop one of those in the phone, it should be able to be activated on Verizon. One of the modules is very similar to the module Librem's code already supports too.

The most likely (but not 100%, so it's not announced yet) module to be put there is Cinterion PLS8. You can easily look up which frequencies it supports.

If they go with that module I will be very disappointed since it has basically 2015-era LTE support (in terms of both bands and speed). Here are the variants and the bands they support[1]:

> PLS8-E: LTE (20,8,3,7,1); 3G (8,3,1); 2G Dual Band

> PLS8-US: LTE (17,5,4,2); 3G (5,4,2); 2G Quad Band

> PLS8-J: LTE (1,3,19); 3G (1,19)

> PLS8-X: LTE (13,17,5,4,2); 3G (5,4,2); 2G Quad Band

> PLS8-V: LTE (13,4,2)

> LTE Cat. 3 (DL: max. 100 Mbps, UL: max. 50 Mbps, 2x2 DL MIMO)

I'm a T-Mobile US customer and while I could live with slower speeds the lack of band 12 support really has me questioning if I can justify spending $600 on this device. The lack of band 66 and 71 is forgivable since those are very new, but band 12 has been around since LTE was first deployed in the US (even if devices didn't start supporting it until late 2015).

Also note that the European variant of this module only supports dual-band GSM and no US UMTS or LTE bands so it will not work for roaming in the US. The US variants will at least support 2G roaming outside the US.

[1]: https://www.gemalto.com/brochures-site/download-site/Documen...

Lack of band 12 would make the device unusable in Seattle, for example.

That one only supports LTE Cat. M1 which is only suitable for IoT devices. As the datasheet indicates max downlink speeds are 300 kbps and there is no 2G or 3G support.

I find the last quote extremely odd and contradictory; my understanding was that UMTS is 3G and LTE is 4G.

I guess "3G" and "4G" here might be referring to some idiosyncratic American standards?

Having at least some band support is looking promising, but whether Verizon will make it a hassle to BYOP or not... is TBD. They're getting better with that overall so I'm hopeful, but I'll believe it when I see it.

The only carrier I see mentioned regularly that it should definitely support is AT&T. Other ones, I am not so sure about.

I received my dev kit and immediately broke it.

The SOM connects to the mainboard with 4x 80-pin tiny connectors. If you miss the little warning on the silkscreen and lift from the wrong side, BAM, two broken connectors.

Purism said they would sell me a replacement SOM, but they're not really communicating now.

Why did you try to remove the SOM? I have a devkit, and have ported an OS to it, but have not found any reason to unseat the SOM.

I always like to take note of the ICs on new dev hardware I'll be working with.

For one, I may find that something isn't working and the chips may be relevant to that. For two, if I want to re-purpose the kit for something in the future, that list of chips will let me know if I'm going to need to write a driver for my target OS.

You can gather a lot of this info from a running OS, but it's much quicker and more accurate to just look at the chips in my experience.

(Purism did finally send me a return shipping label for the SOM today)

> Replaceable Battery: Yes (with tools)

I am Looking forward to the Librem 5 since a while, but those specs aren't particularly promising. In general, there are three features I am missing most often in today's smartphones:

1. Replaceable Battery (without tools)

2. SD-Card Slot

3. Wireless Charging

Sadly there are very few models which have all three features. Requiring tools to change the battery is okay if you want to replace an old one with a new one once in a while, but not if you are on vacation, using GPS all day long, and just want to quickly change it.

It's just one model of phone, so it's going to be really hard to please everyone. The items you mentioned are important to you, but don't matter to me. There are other things I care a lot about that won't be there either.

A good market for phones would include products aimed at lots of different bases, but unfortunately for FOSS phones we're not there yet. So my attitude is to support this phone, and hopefully move toward a world where there will be multiple FOSS phones with different feature sets.

Yes, I am well aware of that. Sometimes I wonder if I should set up a page with a poll to collect what most people would want as a smartphone. Somehow I have the feeling that phone manufacturers don't know anymore what the consumers want and instead build phones with 4 cameras just to present something new (and I am one of those who actually enjoy having a good camera within their phone).

> Sometimes I wonder if I should set up a page with a poll to collect what most people would want as a smartphone

you will get the specs of an high-end kitchen sink for the desired price of a Brondi.

It depends on how you set up that poll. You could attach a price to every feature and let the people decide which feature they want to pay for and what kind of price sensitivity they have (e.g. if they would accept a feature if it would be for free). After you collect the data you analyze the data (e.g. with a cluster analysis) and find out which combinations would have the largest market success.

It would be still hypothetical, but I would expect better (from my point of view) results than the currently available models.

You would be surprised. As long as "demand" is analyzed in the hypothetical, your results will be skewed drastically.

There was a grocery or meal shipment start-up based in London that had all analyzed demand by studying their potential user-base, yet their product never took off. Their result: a lot of people said they would buy, but never did.

The market itself gives you the results you're asking for: Manufacturers are definitely studying and trying everything feasible, since this is such a saturated market.

What sells is where people's priorities are.

Obligatory reference to Homer's car from The Simpsons.

This is why people think big companies like Apple, Microsoft, Samsung, etc. are arrogant when they say they're making products we didn't know we needed until we got them. It's not arrogance; it's that their market research is infinitely superior to a simple poll.

With a poll, we're only constrained by our imagination. Real life has more constraints. The thing the big companies do: balance their research with real life.

Set it up! Then you can use the findings to influence manufacturers.

I agree, a lot of products seem that way.

Replaceable batteries are kind of a lost cause these days. You could buy some model-specific batteries and deal with rotating them around and charging them up and carrying them and rebooting your phone when you want to swap, or you could instead get a power brick that works with most/all of your devices and doesn't require swapping, and doesn't get obsoleted by a phone upgrade.

It doesn't solve every problem that a removable battery solves, but it solves enough of those problems that you're unlikely to see much demand for a removable battery anymore.

And remember that engineering the removable battery means engineering for a couple extra mm in thickness for the battery enclosure, which isn't needed in a tools-only or soldered-in kind of deal.

> Replaceable batteries are kind of a lost cause these days.

But, this is the signature open and hacker-friendly phone! Sure removable batteries have fallen out of style in the mainstream, but this isn't a mainstream product.

But I'm in their target market and I don't want a replaceable battery. I just want the battery to be serviceable when it wear.

For a given phone size, a replaceable battery require some additional space which means it will be smaller than a non-replaceable one.

I don't want to buy one or two additional battery to carry with me when I can get a battery that last a bit longer and that I can charge with any power bank. (mine or the one of a friend as there is no compatibility problems)

Users who want replaceable battery are niche in the niche.

I highly doubt this phone is going to win the thinness wars anyway, so the space savings aren't going to be as relevant. It would be much smarter to differentiate themselves on features such as replaceable batteries, which the mainstream market has ignored precisely because they want to chase thinness.

The real question is, how many users want replaceable batteries, and is there enough overlap between that group and the audience for an OSS phone. I would imagine the overlap is significant, but without market research on the subject I don't know if I'm right.

But you still want to sell it. And more so, you want to please the majority. Oherwise you might be killing the product by picking a niche.

You're just as likely to kill the product by not targeting the niche. Trying to please the majority puts you in competition with the big players. This phone's specs, with the possible exception of the CPU/GPU (haven't looked into it), are a downgrade from the 6 year old Note 3. The only saving grace at the moment is GNU/Linux, separated baseband, and hardware kill switches. Is that going to be enough? If the battery removal tools aren't as simple as a screwdriver, it's already lost me.

I think the thickness is just a minor factor, but there are two major reasons why they are out of style:

1. Making the phone water and dust resistant is much easier without replaceable batteries

2. The manufacturers want to sell smartphones and replaceable batteries are probably the single feature which extends the lifetime of a smartphone the most

The first reason alone would be good enough for many consumers to give up on replaceable batteries but combined with the second reason we come to the situation we are currently in.

I'd rather carry a USB battery pack instead of extra phone batteries. They're bulkier but can be used to charge anything.

As far as I know a microSD card slot would be present.

Things like wireless charging will likely come in subsequent iterations, but there's enough things to do already with the OS, software and the like.

If this sells enough, I am sure more things will be addressed in v2, but if we wait for the 'perfect' phone to buy one, there will likely never be a viable FLOSS phone. I pre-ordered to support the cause. If it's decent enough, I'd be happy.

Re 1: I would also rather have an easily replaceable battery, however for your use case I would think that a powerbank is more sensible than a second battery: no swapping/downtime and you can recharge both at the same time.

Sounds like fun taking photos with a powerbank connected to your phone ;-) In the past, I didn't have a problem with the one-minute downtime. Instead, I enjoyed having a fully charged phone within seconds.

Phone case with removable battery. I had a cobattery (now discontinued) case with my iPhone 6 - added a ton of bulk, but was useful to just swap batteries instead of charging.

Depends how hard is it to replace. If you're carrying around extra battery, tiny screwdriver should not be a problem.

Let's hope it won't be complicated, but somehow I doubt that it will be as easy as those normal replaceable batteries.

Does "with tools" just mean a screwdriver? If so, what's the big deal? Everyone's got screwdrivers.

It's mostly a matter of convenience and practicality. I can change the battery of my Nexus 5 with tools (mostly a screwdriver and a plastic prying tool), but I wouldn't want to do it away from home since it requires me to deal with small screws and needing some time to deal with small connectors that could be damaged if I was in a rush.

A common scenario is you're on vacation and taking a bunch of pictures and using GPS. You could bring your tools with you, but changing your battery involves 10+ minutes of careful disassembly/assembly and you need some space to work. You want this to be "changing AA batteries in a toy" difficulty levels.

How do you keep the second battery charged though? Surely a power pack which charges the phone over usb is way more convienent. You can get a small one if space is an issue or a massive one if you want it to charge multiple times. And you can charge it from any wall socket at the same time as your main battery, and it can be used for your other devices.

Batteries are more convenient because you pre-charge them at home, and then you just need to swap them in the field. With an external battery pack, you need to keep it attached to your phone for long enough to recharge - that's longer than several minutes, and it kinda gets in the way of actually using the device.

My Nexus 4's fragile cover (glass back and bits of plastic on the side) doesn't seem to have been designed for repeated wear-and-tear battery replacements. A 10 minute procedure does a little more damage each time.

So yes, the procedure should be "idiot proof".

For the battery, get together with someone else that wants the same and create a 'jacket' with an additional replaceable battery. And maybe wireless charging also. It would connect to the USB port. Such things already exist for various popular phones from the usual Chinese suppliers.

Are there any phones that support wireless charging and replaceable batteries? I would think that the wireless charging would make an easily replaceable battery much much harder to implement (and thus not worth it when so few customers care).

If "tools" meant some sort of screw driver I'd welcome that. If there are any parts that have to bend to open the case these will break in the long run and are less reliable in my experience (e.g. the phone's back would slide off when pulling it out of the pocket, experienced that with my old BB Q10).

Some really small torx screws that can't fall out of their holes even when loose would be optimal. If they also shipped a little screwdriver that fits onto one's key chain I'd be more than happy with that solution.

In the hardware spec list it says:

"Smart Card Reader: Yes (OpenPGP-compatible)"

The same list says it doesn't have NFC, so I'm just wondering what sort of interface this is going to be, so what sort of cards can be used... Anyone have any idea?

[edit] They have the Librem Key - https://puri.sm/products/librem-key/ - but that requires a full sized USB port, which I assume this phone wont have.

It would likely be a literal smartcard reader. E.g. you put something in there that looks the chip in your debit card.

What do you mean when you say "a literal smartcard reader"? Do you have a link? All the smartcard readers that I have seen are way too large to be placed in a smartphone e.g. http://www.dualboost.com/index.php?pos=photos

All you need is a little slit on the bottom of the phone that you can insert the smartcard into. It doesn't even have to go in all the way, just enough to get the contacts touching.

That seems incredibly unlikely on a smartphone, no?

I don't see why it would be unlikely. The standard identity smartcard (DoD PIV card, openPGP card, etc), like the standard credit/debit card, uses the same physical and electrical interface as a SIM (because a SIM is also a standard smartcard): the ISO7816 pattern. Dual-SIM phones are pretty common — even Apple has one! — and there's no technical reason you couldn't make one of the SIM sockets hotpluggable and present it to the applications on the phone instead of to the cellular module.

If it's the size of a SIM or MicroSD card sure, but if we're talking about credit card sized cards it strikes me as unlikely for packaging reasons. The phone already requires tools for battery replacement, and now it's going to accomodate a durable credit card slot?

Apparently this phone is going to be much bulkier than I imagined.

Apologies, my description was poor. I meant something that looks like a SIM card.

I was interested in this line as well.

It's probably one of these - https://en.wikipedia.org/wiki/OpenPGP_card

If you look between the battery and the microsd card slot you'll see the the slot - https://puri.sm/wp-content/uploads/2019/01/figure18b.png

Looks like it's an update on that the hardware started landing in some people's hands, not a change in hardware.

Maybe the community around this, both skeptics and non-skeptics alike, should come together and ask Purism to Open Source document the phone and it’s process of building a device like this (through as many mediums as they choose)? These blog posts are nice and fill a basic need of the public, but it would be great to give the idea and opportunity legs if they do or don’t succeed.

Reader beware, I preordered.

I ordered one a couple of months ago. I am hoping it turns out usable. Anything beyond that is gravy.

I am a bit disappointed to find them programming it in C "in this day and age". I would rather find C++ code, which leaves less room for dumb bugs, and is quicker to write and refine. But C++ code written by C programmers would be no better than C, and in the end you go with who you have.

given how insanely bloated the language is, I doubt it.

GTK is written in C.

When you have to resurrect 1990s falsehoods to make a point, you have no point.

There are good C++ wrappers to make GTK tolerable to use, although they cannot make it less buggy.

Then run plasma mobile on it.

I'm not sure why I haven't preordered one yet. I think I'm worried I won't be able to use it for daily use due to missing key apps (signal, authy, anki, password manager)

Realistically, I think that's my only concern. Might still get one though to tinker with


Seems like the wiki has not been updated properly.

>"Don't list problems"

Well, I would rather be aware of those.


Edit: Jumped around the site too much and lost the context...

That page is not about the Librem 5

As a iOS/Android developer, I'm looking forward to more information on dev tools. Last time I went through their docs for this information it wasn't very clear where to start.

It's not real clear, which is a shame because the development story is what makes me so excited about it. Basically you want to look at the gnome developer documentation (https://developer.gnome.org/) this is also often incomplete or outdated but it's a start. A standard gtk app will probably run fine on the phone, maybe with some minor tweaking. For compiling, linking, etc you just use the standard c tools, or the python ones.

I can't think of the name right now but there is a widget gallery installed with the libgtk-3-dev (I think) package that includes a heap of code samples.

There's also this whole gui xml designer thing they keep trying to push, but the best thing about librem development is that you can start with a single c file and expand from there.

Basically, download GNOME Builder and then it's your standard GTK development. Or you can use any other tools that you would normally use when building for Linux.

> Or you can use any other tools that you would normally use when building for Linux.

That's not super helpful if you've never built anything for Linux before.

Ok, let me try to be a bit more helpful then.

On Linux there are two big desktop environments, (GUIs, desktops etc. on macOS/Windows there's no choice, only one).

These environments are called GNOME and KDE.

The Librem5 will by default ship with a GNOME-based environment, customized for phones. On GNOME, you program GUI applications using a widget library called GTK, now at version 3. Think of this akin to Cocoa/WinForms.

There are two main programming languages to develop GTK apps, C and Vala, (Vala's akin to C#), however there are bindings to many other languages including Python and JavaScript, so unlike on macOS/Windows, you're not as restricted when it comes to your development environment.

GNOME has an IDE for building applications intended to run on it called Builder, akin to Xcode/Visual Studio. You'd normally use this IDE to develop your app. You'd want to select the Flatpak project type. Flatpak is a way to ship apps that will be supported on the phone. There's a library called libhandy you can use to make sure your app fits nicely on the phone ie is responsive.

The reference GNOME docs can be found at https://developer.gnome.org, the Librem 5 dev docs, which are WIP are at https://developer.puri.sm/Librem5 and the libhandy API docs are at https://honk.sigxcpu.org/projects/libhandy/doc

Now because this phone is GNU/Linux and is open, you're not restricted to what I laid out above, that's just the most sanctioned way to develop apps.

The phone will be able to run any Linux binary that can be compiled for ARM, so as long as the compiler you're using can do that you can use pretty much any language, framework etc. You can even draw your graphics via OpenGL or ship a command line app or anything else you may wish.

You're not as restricted, but also not as guided as on iOS/Android.

Sorry, I realized my comment was probably snarkier than I mean it and made it look like I was fishing for a "more helpful" answer. :)

That said, this is a great comment! Thanks for summarizing!

No problem, honestly I was just a bit lazy with my first comment, you pushed me to stop being lazy, so thanks. :-)

iOS and Android basically provide a sandbox for Apps (where the user can set permissions). Does Librem5 do the same, or does it install programs like, say, apt-get would do, that is, without sandbox per application?

Flatpak is what provides you with a way to package applications with all of its dependencies and also provides a per-application sandbox. You'd normally install an application via the Flatpak app store that is going to be on the phone.

apt would remain an option for those who want it, but for regular users they'd install the Flatpaks via an app store like interface.

Wouldn't you go the Builder route then?

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact