Librem 5 Shipping Announcement | Hacker News Reader

Librem 5 Shipping Announcement

puri.sm

188 pointsgnufx88 comments

This is a... puzzling announcement.

It certainly reflects a departure from their previously announced shipping window, which back in May they "adjusted" (the word they used at the time) from Q2 2019 to Q3 (see https://puri.sm/posts/massive-progress-exact-cpu-selected-mi...). And certainly back then they weren't saying anything about Q3 2019 being the start of a year-long iterative release cycle, it was just that the product would ship in Q3 2019, full stop.

A cynical interpretation: it sounds like all the batches until "Batch Evergreen" (due Q2 2020) will be of less-than-general-release quality, with each batch getting a bit closer to what the average person would consider a finished product until Evergreen reaches that mark.

So one reading of this announcement would be that release of the actual, finished Librem 5 product has now slipped again, this time to mid-2020 -- but they didn't want to just come out and, you know, say that. So instead they've created this "iterative release schedule" so they can say they're shipping something in Q3 2019, even though the actual thing they'll be shipping then isn't going to even be close to what the average person would consider a finished, release-quality product ("loose fit, varying alignment, unfinished switch caps... basic web browsing, early power management").

I like Purism and am generally supportive of what they've been trying to do with this device. But this persistent pattern of (1) slipping deadlines and then (2) announcing that the deadline has slipped in ways that seem calculated to obscure that fact is really troubling.

Missing deadlines, by itself, isn't that big a deal on a project like this -- they're trying to do something big and unique, it's understandable if that takes longer than expected. But if you're going to miss a deadline, I'd rather you just say you missed it and give updated guidance, rather than trying to convince me that isn't what actually happened. It just leaves a bad taste in my mouth.

I guess it might be necessary for marketing reasons, but I'd prefer if they'd just not announce deadlines. You know you're not going to make them, these things always slip, again and again.

I also do not really expect them to be "what the average person would consider a finished product" by Evergreen either. As far as I understand it, this is really aimed at power users and enthusiasts willing to suffer some pain for the ability to tinker and the feeling of control over your device. Which is great, but they might've been a bit clearer on that.

But ah well, I suppose mistakes are inevitable, and there'll be many more - it's a young company.

Even for enthusiasts, who I agree can be willing to forgive a lot of fit-and-finish type issues, the first couple of batches seem really underwhelming.

I've never seen any statement from them that the current state of the software for the device is at the level of "basic web browsing, early power management," for instance. That sounds like an early alpha at best.

Software issues are less concerning, though, since those can always be fixed/improved with updates. Accepting alpha quality hardware if you're in an early batch seems like a bigger deal. If you're in an early batch and get a hand-assembled box with parts rattling around inside, is that just what you're stuck with forever? When all the previous guidance was "this thing you paid for will ship in Q3 2019," why do you now have to make a choice between accepting an unfinished device now or getting a finished device sometime next year? I can see how offering that choice benefits Purism, but I can't see how it benefits the Purism customer.

A lot of this just boils down to expectations management, which Purism hasn't been great at.

And it sounds like most of the hardware in early batches is exterior hardware that won't be ready, so I almost wonder if it could be upgradeable later (perhaps for free/cheap?) if the internal hardware is all there. But they don't really clarify whether or not that will be the case.

On Matrix it was mentioned that a FAQ that answers questions like that is coming.

Probably, should have seen this coming and had that ready with today's announcement. I actually find this concerning because it's a glimpse at the bubble they must be in if they thought this announcement would stand on it's own, or if they thought that it would be welcomed with applause.

> I guess it might be necessary for marketing reasons, but I'd prefer if they'd just not announce deadlines. You know you're not going to make them, these things always slip, again and again.

Fairphone did this with the Fairphone 3, and people complained they had to idea what the specs were going to be. Some even wondered if the company was dying. Finally, there were complaints that the marketing was "like Apple", a blasphemy.

There's something to say for both methods, and there's always going to be people unhappy with any decision made.

Haha, I guess. What's interesting there is that they did announce timelines for the Fairphone 2, and had to postpone that a number of times - to many complaints.

As an enthusiast, I did prefer the way they approached it this time, and got pretty hyped by the FP 3 announcement. i also didn't see as many complaints about the uncertainty as I did before about the delays, though that's probably because there were no delay posts that people could comment on.

In any case, it'll be interesting to see what approach they will take for FP4 - i.e. which they liked more.

>Missing deadlines, by itself, isn't that big a deal on a project like this

You would think the people responsible for attempting said project would have the foresight to understand this in the first place.

I think this is the normal hardware development process, opened up to public scrutiny.

I think it's reasonable - they seem to be matching expectations to groups of people.

People who want better hardware and software should wait for a polished phone.

People who want to develop software, or who want out of the chooise-ios-or-android problem can get an early phone with reduced expectations of perfection.

Personally, as someone who paid for one during crowdfunding, I'd be fine with one of the earlier hardware revisions. (Though probably not the very first!)

I'm honestly not as worried about the state of the software as I am the hardware; the software can always be fixed later. As for the hardware... so long as it's put together well enough and runs alright, I'm pretty happy.

I've been holding out for a worthy successor to my old N900 for years, and quite honestly I'd rather have something sooner that's a bit less polished than keep waiting!

This announcement reminds me of thw neo900 project. I hope for a better outcome, but this announcement sounds eerily familliar. Purism seems better positioned to ship, eventually. Seeing as Q3 is almost expired, I expect a similar announcement from the Pine64 group,too.

Edit: Pine 64 will ship

https://www.pine64.org/2019/09/05/september-update-the-pinep...

If you look closer, you can see that Purism and Pine64 are actually at similar stages to each other and promise to ship similar things at similar time.

I think there is an unfair amount of negativity in here and would like to offer an insiders perspective:

I have worked 11 years in product development in various roles (Hardware, Test, Production, Firmware and Mechanical often simultaneously). I can confidently tell you that this is how product development is in the real world (even amongst the big players). This schedule puri.sm publised is something you would never normally see, but I guarantee you exists for every product.

This is why, even though I am a lover of technology, I am not an early adopter. I normally buy a product when it has been on the market for 6-12 months as, in my expirience at least, you are on about revision 3-5 by then. The last minute changes for compliance are no longer done by hand, the first major issues from the field are fixed, mechanical tools are stable etc.

I know this place is majority software engineers so here's an analogy I'm sure many can relate to: How often have you shipped software to a customer that is beta? Or even alpha? Product development is no different :)

The expressed negativity is not directed towards hardware development process, rather towards what ended up being insincere promises the team gave about the state of the product that would be shipping in Q3 2019.

Oh I understand that. I was just trying to communicate that: their latest promise was to start shipping final hardware to people Q3 2019 and Bach Aspen is exactly that.

I actually give them kudos for allowing people to choose what batch they receive. I specifically did not pre-order because I don't want an early batch of hardware of any product. Had I know I would get a choice I would have pre-ordered.

You still can.

Wish them all the best, but I tried the latest firmwate image on my dev kit last week and although there's plenty of progress, many things are not there, at least on that image + platform combination.

No lte, no wifi, browser scrolls at ~10fps, browser crashes, display has some hw jittering if you look closely.

I guess some of these are solved on the production hw (they use a different lte module) but still, only consider it if you have a pain threshold or have booked to go into hypersleep for 18 mo.

That sounds about right for where they are in the development process right now. Anyone expecting a remotely 'consumer' device experience should be waiting for the Evergreen release at least. Only early adopters willing to deal with some pain should be jumping in right now.

I paid for a "dev kit"... it's not like I have unreasonable expectations. I went back to check it to figure out if I should buy an actual phone... I kept my money in my pocket.

I can handle quite a bit of dust and pieces coming month by month but I can't handle no wireless comms and a crashy and slow browser. At least the internet and the browser operation have to be reliable and fluid or nobody can use it for normal operations. I'll try it again in a few months.

But LTE and WiFi work just fine on my devkit?

What did you do to get them working? I flashed the latest image and no sign of them, even with a valid SIM.

The UI is being reworked right now to better fit the screen, but it's generally there in GNOME Settings (latest changes broke the password field and I think it's not restored yet, but you can fill it in by editing NetworkManager's config anyway). For LTE you just have to set the correct APN. I've just connected to the internet with LTE to test it and it worked; and I'm using WiFi connection daily.

Make sure your kill switches actually turn WiFi and LTE on. Also, the modem is generally way more stable when used with battery inserted, as USB might not be able to supply enough current.

Turning on power switches to the chips is pretty important!

It gives a hint about how manufacturing is, that the switch tops come out so late.

>LTE and WiFi work just fine on my devkit?

I don't know, do they?

Which browser are they using?

It's the Gnome webKit browser Epiphany (aka Gnome web).

I'd prefer firefox but I don't have any real beef with it. Except on the dev kit the scrolling is more or less unusable. It might be a 'feature' of the dev kit since the imx8 runs hot enough to burn your hand on that due to hardware problems that shouldn't exist on the real device.

Is there a reason they wouldn't get Mozilla to help out with a FF mobile port? Seems like the obvious and symbiotic way to get this done.

I'm sure Mozilla would accept Purism's work if they did it, but even a lot of work for desktop Firefox (tabs in the titlebar, wayland) is done by Red Hat people.

That's true, there's a lot more to a partnership though then just the coding and it would still be beneficial for both.

Epiphany, GNOME's WebKit-based browser. Firefox doesn't resize small enough to fit on the screen.

Does that mean no extensions and thus no ad blocker? If so, do they have another ad block solution at the device level?

Epiphany has support for Easylist-based adblocking -- at least on the desktop version, I don't know about the mobile interface.

There's no distinction between the "desktop" and "mobile" versions of GNOME Web. It's literally the same code (compiled for the appropriate processor architecture), just at a different window size.

Since I am personally unfamiliar with GNOME’s toolchain for compiling apps for both mobile and desktop, I didn’t want to speak too confidently about the matter above.

So, all options shown in the preferences of the desktop version will automatically be shown in the preferences of the mobile version?

You can test it by shrinking the Epiphany's window while running its latest version on desktop.

Epiphany has a built-in adblocker (using normal adblock style rules).

> Firefox doesn't resize small enough to fit on the screen.

What exactly does this mean?

* I can make the FF desktop window super tiny

* I can do browser zooming down to 30%

* even if the chrome would obscure the webpage viewport I can go into fullscreen mode

* FF itself has a responsive design mode and ships with Iphone and other small form factors in the dropdown

I understand that it may be easier to tweak Gnome's browser for adding navigation gestures/url bar appearance, etc. But I can't imagine how FF wouldn't be able to display a web page in a way that fits the Librem 5 screen.

It can display a web page, but there's a long way from that to making the UI pleasurable to use on small touchscreen.

Can I ask what you mean by that? Firefox's core has been used on the mobile phone in multiple applications that function just fine and Firefox's desktop version is perfectly capable of resizing down to mobile phone sizes, even emulating a phone if you want.

There's no mobile version of Firefox that would work on GNU/Linux. There was one, but got ditched when Fennec got rewritten into being a native Android app.

Developing good smartphone userland and OS is very software engineering resource intensive. You cannot do it on a shoestring budget. Even Nokia with its Maemo was only 80% there - and they were able to throw hundreds of software engineers on it. I was also part of one early smartphone development team where you had 150 people at its peak and they still could deliver the product. Not to mention OpenMoko.

Even when Librem delivers, it will be subpar UX to what people are used to. It is a trade off, but I feel only very hardcore enthusiastic are willing to take it.

Not sure what module they're using for the devkits, but you can check to see if your bands are supported here:

https://puri.sm/faq/supported-networks/

To see which bands your provider uses, check here:

https://cellmapper.net/

One of my concerns about this is fragmentation. They've essentially announced 6 different phones released over a staggered time period.

Will all batches get updates for the same duration? Will they all get same day software releases? How about support? Will there be subtle differences that cause headaches? If you opt for one revision over another and that has issues, what's the process going to be to handle that case?

Rather than just having a single shipping device with a fixed design they're going to incrementally change it, and I fear it's going to be an absolute disaster.

I think if you visit this comment in exactly a year's time I'll probably be right, and it'd hardly be a surprise.

I absolutely agree with smacktoward, they're doing this to hide the fact they're simply not ready, but instead opening a can of worms that could devastate the entire project.

Not a fan of this, you can only hope this is the right move forward.

Most phones are incrementally changed like that (aside of the case maybe, that's unusual), it's just usually not announced publicly. Aside of the v2 at the end of the schedule, this is all about multiple batches of a single model (and even v2 probably won't have a separate software distribution - I would expect just a different bootloader to flash).

That still doesn't answer the question of the fact there _will_ be different models and how that affects software support.

For example, Chestnut says: "improved power management", will all previous models get that?

In Birch it says "Next run of board", are there changes that change the software in any way?

Evergreen is the "LTS" release, do the others not get "LTS"?

In Q4 2020 there's a new CPU, does that affect drivers or software in any way? Will they provide software updates for all devices for a number of years?

You'd hope opting for Model 1 2019 doesn't punish you but I can't see how they have the resources for testing every single model for every single software release.

It's not a good idea in the long run. Not even Google would do something like that and nor should a smaller company, it's idiotic.

> That still doesn't answer the question of the fact there _will_ be different models and how that affects software support.

There will be different revisions, not models. It doesn't affect software support.

> For example, Chestnut says: "improved power management", will all previous models get that?

Yes (unless there's some serious hardware bug fixed in Chestnut, which is unlikely).

> In Birch it says "Next run of board", are there changes that change the software in any way?

No.

> Evergreen is the "LTS" release, do the others not get "LTS"?

You just apt-get upgrade to LTS. It's listed in the milestone because, well, it's not going to arrive before that.

> In Q4 2020 there's a new CPU, does that affect drivers or software in any way?

Unknown at this point, but even if yes, then today with kernel device trees it shouldn't be any issue.

> Will they provide software updates for all devices for a number of years?

This software is made to work on GNU/Linux. You can already run it on other devices, even your laptop. Just look at GTA04 - it's also kept close to mainline and in turn, it's still being supported.

> Not even Google would do something like that and nor should a smaller company

Google does something like that. Apple does as well. There are multiple hardware revisions of single iPhone models, some even include as big changes as different LTE modules.

I'm tempted to get the earliest batch I can, and then purchase a second phone for v2.

I'm a little disapointed the process hasn't been as smooth as I would have hoped, however I feel that is completely offset by the new type of product we're getting.

I was thinking just the same! :) My main reason for buying this phone (which I have) is to support this movement. To support a hope that the phone platform (which arguably will grow grow grow to be the main future platform) can have a proper FOSS platform. I HATE that all I can choose from is iOS (apple) and Android (Google). Just imagine where the world would be without Linux.

Some pictures (/mock-ups?) here: https://social.librem.one/@purism/102740815257018934

Looks a bit chunky…does anyone have a good side view shot?

Looks a bit modular. Seems you can change the WiFi card. I wonder what's that other exchangeable bit to the right of it?

The WWAN/LTE card is replaceable. An incredibly nice trade-off for a little extra thickness.

It will definitely be very chunky if they fit 2 MPCI modules on there

Looks chunky indeed but I like it. Tired of the fucking thin and light phones everyone else is selling that easily want to slip out of my hand. I hope this has a heavy and metal feel to it.

For those of us suffering from Noassatall, who cannot comfortably carry anything in our back pockets, a thin phone fits nicely in the same front pocket as a wallet, without getting scratched up by the keys in the other front pocket. I too would like a heavy phone, but I think it could be thin and heavy if made from the right metal, instead of glass and plastic.

I carried a palm pre in my front pocket 10 years ago, it was probably thicker than this phone (though also much smaller diagonally).

That, and my Treos before it, went in my pocket with the keys, but that was the last one to do so. They didn't build it as tough as the Treos, so it didn't last long.

I guess this is both a way for them to pressure themselves to finally release no matter the cost because with a thing like this I believe they could go for years without considering it "ready ready". They have way less resources than other big corps when they entered the business. Also I guess this will provide many new testers to make the repro-identify-fix loop.

I am very excited for this phone, but I am not sure about this approach. I think it will cause a lot of people to wait for the v2 release.

As someone who placed their order recently, I am totally fine with that if it opens up the possibility of people who ordered late to get their hands on one sooner. I don't mind having to tweak component alignment and update via the terminal (and maybe update keycaps?).

I think people who are afraid of effectively being beta testers are going to have a tough time with a full-on linux phone whether they get one now or after another year of development.

I am perfectly fine tweaking software, but getting an unfinished hardware build is pretty hard to swallow for the price. This isn't a dev build either, it is supposed to be the final end product. To me it sounds like they are trying to release an unfinished product so they don't miss their ship date again. I would much rather have the ship date pushed out and get the finalize hardware.

>I would much rather have the ship date pushed out and get the finalize hardware.

Then isn't this the best of both worlds since they give you exactly that option?

This sounds like beta testers is too generous. It looks to me like the finished phone next year will be beta testers by any reasonable measure.

I am in this boat. However, I've been waiting for this phone for so long, that I've regressed to a flip phone since my previous phone broke and I was unwilling to buy another smart phone since the Librem was coming out in only a few months. So I appreciate the published schedule, it allows me to weigh the pros and cons of getting the phone earlier vs having the benefit of getting some of the kinks worked out. It seems that the hardware won't change much except for mechanicals, so at this point I'm thinking I'll wait for the second iteration just because there's always something with the first one :)

I too have reverted to a flip phone. Well, actually it's a "candy bar" phone -- the Nokia 3310. I've been using it for one year and one week. I'm definitely looking forward to the Librem 5 release.

Osborne Effect:

https://en.wikipedia.org/wiki/Osborne_effect

Close but I don't think it applies until they actually announce v2. The effect is about how the announcement effects the psychology and thus sales of the current version.

They basically did announce a "v2", it's the last batch:

> Batch Fir

> Hardware: 14nm Next Generation CPU

> Mechanical Design: Version 2

> Software: Long term support release

> Certifications: FCC and CE

> Shipping window: Q4 2020

Why include a significantly upgraded model in the release schedule for v1, though? What's the logic in that? Why would you want to give a fraction of your user base a big CPU upgrade essentially for free?

My reading: they're way behind schedule, and they know people will be upset by the fact that the device they thought would be shipping in Q3 2019 won't actually ship until "Batch Evergreen" in Q2 2020. So they want to give people an incentive to wait -- something they would feel makes the wait worthwhile, something to wait for. And telling them that if you're willing to wait longer than anyone else, they'll throw in a significant hardware upgrade for free, creates that incentive. It gives people a reason to hang on instead of asking for a refund.

This. I was already planning to buy a PinePhone to check out the "Linux Smartphone" scene and then take a wait and see approach on the (much more expensive) Librem 5 but now I will definitely be waiting until the units with the improved processors ship late next year.

Yeah, I'd rather not order early and be caught with a very unfinished build, only for someone who ordered later to get a better one for the same price.

I'm excited about what's happening here, I'm tempted to get like a PinePhone to play with, at a lower price point, but go iPhone this year, and hold off with switching to a Linux model until there's more robust manufacture out there.

I appreciate folks who can invest early in important projects like this, but I'm not in that category.

They said in the announcement that those who ordered early can choose to wait for a later batch if they want to.

That's probably a good thing: a lot of people wanting a relatively smooth(er) experience should be waiting for v2 or even v3.[1] v1 of most hardware, especially a Linux phone, is going to involve some bumps and bruises. Keep in mind this is a small team running with a minuscule budget compared to consumer phones so you're not going to have an iPhone or even Samsung device/software experience. They're doing some great work and but it's going to take time to work through the issues.

[1] Would you really want to encourage a bunch of enthusiastic non-technical/unprepared customers with high expectations jumping in on v1 of this?

I find it amusing how perfectly they align with PinePhone[1] - first batch in September (with some drawbacks), final product Q2 2020, both blog posts released on Sept 5th.

[1]: https://www.pine64.org/2019/09/05/september-update-the-pinep...

So much negative in here! I mean, anyone who thought these were going to hit the market like an iPhone was always in for a rough surprise. We're talking small companies trying to run in a marketplace almost exclusively occupied by mega multinationals.

I'm really excited to see not just one but potentially two open phones around the corner. I really wish I wasn't about to move overseas so I could buy one (more likely, both), but as soon as I know where I'm living I will!

Protip: when putting out a shipping announcement, put some product shots somewhere near the top. Attention spans on the Internet being what they are, I'm not reading three screenfuls of text to find out what it is you're shipping.

Well, I'm out. This is disappointing and is starting to have a pretty bad vapor smell. I shelled out the change for a presale but thevpruce difference isn't worth it for unfinished hardware. I'll check back in for v2 but for now I'm thinking a dumb phone will be the way to go. I with them all the best and appreciate the announcement but with all the delays and now a timeline showing just how far from a finished product they are, my confidence is low.

"vapor" ... they just said they're shipping the first phones in 19 days...

I'll clarify:

The batch of phones going out in 19 days is not a finished product. It's expected to be loose fitting, not have endcaps on the switches, not have properly functioning power management and updates have to be done using a terminal. That's beta hardware, at best.

The next few iterations are improvements on the beta hardware but still not yet a finished product until Evergreen. That speculatively Q2 2020 but they've been off by miles with all the previous delivery dates so I won't hold my breath. Until that hardware comes out, which is what a consumer grade product should be, the only hardware delivered will be various levels of beta/release candidate.

By then, you might as well wait a few months for the hardware refresh. I appreciate the goal and the work done and if the hardware was polished and all the pain points were software, I'd be singing a different tune. Software can be fixed after the fact but poor fit and finish on hardware is forever. (And I've been Linux-as-a-desktop since the latter half of the 90s, so I can take a lot of pain. I'm all in on a Linux phone but I'll go buy a dumb phone while I wait for v2.)

I fear this is going to be OpenMoko all over again.

Funny coincidence (?) naming the first batch Aspen, which was the code name for iPhone OS 1.0.

Unfortunately it looks like they picked Aspen, the wood, rather than Aspen, the ski resort. (Also, wasn't iPhone OS 1.0 code-named Alpine?)

Iirc Alpine was the default root password on early versions of iOS

"alpine" is actually the current password for root (and mobile) on iOS. The earliest versions of iOS used "dottie" as the password, AFAIK.

> And we are compliant with, and submitting for, the “Respects Your Freedom” certification from the Free Software Foundation.

This is huge. Let’s see if they get certified.

does the contacts app support carddav?

I am thinking birch or chestnut sounds fun

It's just GNOME Contacts. Yes, it does.

Reply on news.ycombinator.com