This is good! Hopefully Purism can build on their recognition to funnel resources toward FOSS-mobile applications because this will help us all.
This reminds me of something else. What would be the best model for a FOSS "app store" with mechanisms for compensation? Lowering the barrier for transactions to FOSS projects really helps their survival and sustenance by individuals who cannot contribute code.
Here's a very half-baked thought: One can imagine users supporting community servers (matrix, activitypub/mastodon, etc.) and applications through a cryptocurrency wallet that is loaded in the same way that people load their g--gle/a--le wallets. One can also imagine applications that can tie into this wallet to compensate publishers based on some model (a transaction API like g--gle/a--le or something like what brave is doing with the BAT). My question is what are the failure modes of such a model?
For example, this still wouldn't stop the perverse incentives of:
- Content publishers going for the types of bait that generate the most clicks. Clicks=compensation regardless of it's from ads or some sort of attention token.
- Application developers casing addiction mechanisms that go after whales. The same loot-box mechanisms can be implemented with a FOSS game that uses a crypto wallet API.
I'd be happy to use an app store that tracks active screen-time per app and disburses a monthly "subscription" proportionally to the developers of all the apps I use.
I don't have any concern about the type of currency used to pay.
A quick check of my screentime over the past few days shows only two apps which wouldn't benefit from this: DAVx and Syncthing. Both of these are essentially daemons I configured once, and ideally never see the app interface of ever again until I get a new phone. Otherwise, apps I benefit from are represented in the screentime of the past couple days.
I've started developing something like that, and I started by noticing like you that basing on screentime isn't fair to apps that are purely in background, so I tried to compensate for that.
Then I found that some apps, even though they were useful, didn't actually do much under the hood. Like say Frost for Facebook which is mostly a webview with few extensions. So I wanted to add a section for things that are mostly a "viewer". An additional example of such a thing is a media player: if you look 10h a day at videos, it's thanks to the content, not to the viewer.
And then, there are behaviours you maybe don't want to trigger, like getting people addicted. 2048 using the screentime metric would have gained a lot of money (and it does deserve a lot!), but IMO people who got addicted to it should give them less.
And what about apps that have paying variant on Google Play?
Which deserves more money, the app that is simply free on fdroid and Play Store, or the app that is free on fdroid, but paying (or free reduced) on Play Store.
So, I spent few days trying to make something fair, and at the moment I believe that the most fair I can do is to give equally among all apps that I've used in the last 7 days.
This may result in applications trying to keep you hooked in, addicted, or having unnecessarily long processes.
Subscriptions make sense for dynamic content, and service. Fantastical and 1Password are examples of subscription-based platforms which were once buy to play (term from games industry).
However if you self host the data (you probably can and should) or sell your soul to the devil (Google etc) you paid for hosting. So a subscription for the software doesn't make sense from customer PoV. You essentially don't pay for service, compared to Disney+, World of Warcraft, or Netflix.
> I'd be happy to use an app store that tracks active screen-time per app and disburses a monthly "subscription" proportionally to the developers of all the apps I use.
Screentime is a really, really bad proxy for usefulness though. Weather apps, email clients, maps (without turn-by-turn directions), the alternative app I use to control my Hue lights, ... will starve, while the Kindle app and social media will get showered in money. That alternative camera app, hardly any screentime at all. My to do app shows me lots of notifications but gets very little screentime. Netflix however...
I wouldn't want any monitoring of which applications I use. In any case, such an incentive would make developers of apps that should have quick, fast actions do things to make the apps slower. For example, if it takes a user 1s to use the app, they'll add a 0.5 sec animation to open it and 1.5x their money and so on.
And those seconds will be dominated by the tens of minutes I spend in other apps per day. And Android and iOS already monitor which apps you use. The calculation of ratios could easily be done locally.
> And Android and iOS already monitor which apps you use.
I know that. I don't want it as I stated above.
> And those seconds will be dominated by the tens of minutes I spend in other apps per day.
That was just an easy to understand example. The real money will be in maxing out total screen time. Such examples exist as Facebook, Netflix, or advertising driven platforms that hyper-optimize for watch-time or "engagement" to our detriment.
Isn't apt-get, pacman, emerge, etc. more or less equivalent to an app store ?
I don't think there's necessarily a need for projects to be funded or paid for - giving back to the mobile ecosystem the same way we give back to the desktop ecosystem seems like a pretty fair way to go.
In what way? I don't see any fundamental difference between discovering GNU/Linux packages and Android ones (in the "stores" I've used -- f-droid, /e/ apps, and Aurora).
Algorithms that match intended words and synonyms over verbatim pattern matching. Seriously, apt search sucks because if you are searching for something that does not include the word you use, it falls over completely. Just like every single time I try and brew install a package I sort of remember the name of, it will just not give me anything.
This could be an area of improvement. E.g. I don't think adding metadata to packages would be particularly complicated and that could help (e.g. using tags) with discovery.
It's interesting to read the definition of a linux distribution on wikipedia:
"A Linux distribution ... is an operating system made from a software collection that is based upon the Linux kernel and, often, a package management system."
When I read that, I realized just how fundamental the package manager is and it makes sense.
I'd be interested in a Purism app store that charges a small fee per app "purchase". The apps would be Libre so nothing stopping people packaging/compiling etc themselves (or getting from an alternative source) but to me the convenience of having that done already and curated/integrated would be worth a couple of bucks per app and even if it was optional, just having the default be paying a small fee might give enough of a nudge in the right direction.
I'll pretty much pay for anything which takes the packaging out of the equation. There's value in not having to build and distribute my own stuff (and those mechanisms need financing).
On top of that, I'll gladly pay for hosted versions of self-hosting systems. Nextcloud, Mailserver with DNS, gitea, drone/jenkins, piwik, etc.
Sounds counterproductive, to pay someone to host something that is meant to be hosted by yourself, but most often, I just need the option to move to selfhosted when time arrives.
Sure, I can host my own piwik, or gitea, or nexcloud (I do) and even (painfully) manage my mailserver and DNS with mailinabox. Out of principle. But the TCO of that is mountainhigh, if I'm honest. Tucking someone €9.99 a month to deal with one of these for me, would save me a lot, in the grand scheme.
Edit: what I'm trying to say: I'm convinced there is a businesscase there. There's money to be made in this area of FLOSS, which can trickle down into the upstream projects or be a businesscase as funding for the projects themselves even.
I've always thought the F in FOSS was for "Free as in Freedom" not "Free as in Gratis" but yes, FLOSS was the intention here, thank you for the reminder.
This is just an annoying result of OSS projects squatting on FOSS, re-muddling the waters about what "Free" was supposed to mean (free as in freedom/"libre"). The original dichotomy was just OSS / FOSS. I guess the resulting division is precise (opensource you pay for / opensource and gratis / opensource and free as in freedom), I just find "floss" somewhat gross and just say Free Software.
The "L" part was an effort to include all the speakers of Romance languages who say "software libre" or "software livre". In my experience living in South America, very few people say "código abierto" or "fuente abierto" at the Linux conferences except if they are companies presenting business plans. For most, they don't see the need for the division between "free software" and "open source", so FLOSS is a more inclusive term to cover more of the world.
My personal belief is that we need the discourse of both the OSI and FSF because they appeal to different segments, and we need both the business community and people care about user rights to be in the movement if we want FOSS to succeed.
I dont care about the cost or any of the pysops from you google and apple employees. I want to see and support more free software mobile operating systems period.
The Purism Website failed to convince me that this project has even a chance of becoming an enjoyable alternative.
Not only doesn't it show many features or screenshots of apps, the things it does show are better at demonstrating its weaknesses than strengths.
For example, theres a link to a Video about running Firefox [1] that reads like a parody. At one point, the narrator says enthusiastically: "With the settings unusable in portrait, it's time to switch to landscape mode"
Thankfully, they also offer an optimised browser, but the fact that they mention Firefox at all, demonstrates a lack of understanding about what users want from a phone.
This phone is still work in progress, but it progresses at a fast pace. The very fact that you can smoothly run desktop Firefox with all plugins on a phone is a huge thing. There is no need to write new applications, just adapt the existing ones. If you want more pictures and videos, look here: https://forums.puri.sm/t/librem-5-media-photos-and-videos.
> The Purism Website failed to convince me that this project has even a chance of becoming an enjoyable alternative.
I carefully chose the word "enjoyable". I absolutely see the appeal, and I do want this project to succeed.
I also understand, that it's a work in progress.
However, from my experience, I believe that Linux is simply not enjoyable to use. Other operating systems provide advanced features out of the box, such as clipboard-history, clipboard-sync, PiP, powerful search and more. In Linux, you may need to use the command line to arrange your screens [1].
The Website should have convinced me, that this project is built by a company that cares about usability just as much as they care about privacy and openness. Instead, I get the impression that it's only for the most determined of linux users - you know... the kind that would wan't desktop firefox on their phone even if it's almost unusable.
> There is no need to write new applications, just adapt the existing ones
Yes, in theory this is great. But in practice, on my Surface, it never worked. The Interface was laggy, buttons were hard to hit, the touch keyboard didn't appear when it was supposed to, smooth zooming in Firefox didn't work. The difference to an iPad was like night and day.
[1]: Source: Watching a coworker use Linux for the first time. I don't remember the distro.
It's not a lack of understanding. I would like the option to run desktop applications on my phone and I also appreciate that they are honest with what works and what doesn't.
Open source projects works by targeting people who understand technology. When I first installed a linux distribution, maybe it wasn't what a user wants from a distro. A decade or so after that everyone and their aunt has a linux computer in their pocket.
Some have a Java based computer that happens to use Linux kernel, but could pretty much use anything else POSIX based.
Others have an Objective-C/Swift based computer that happens to use a mix of FreeBSD and mach hybrid kernel, but could pretty much use anything else POSIX based.
What I think is that Purism got it wrong by creating its own mobile distro instead of starting from a pure-FOSS Android (AOSP) and patching the missing things with thinks like microG, better permission system, isolation and ways to force closed source apps to behave and obey the user (app wants my address book and refuse to work if I deny it? Nice, take this randomly generated list of names and numbers! Location? Take this random place. Phone or cell ids? Take this random stuff...)
Advantages is that they can start with a bunch of free apps (F-droid is nice!), do not get too far from what SoC vendors already have and can move to another distro (even running both them in parallel) once they sold a good number of devices and the business is sustainable. All while respecting user freedom even when they really need some proprietary stuff (e.g. bank apps).
But, well, that ship has sailed. Let's see what we can do.
Eschewing Android for an actual Linux OS was the point. Alternative Android distros have been around a long time.
The link addresses one point about apps: "With the Librem 5 phone you are running a full-blown desktop computer in a mobile form factor. If you plug it into a keyboard, mouse, and monitor you will have PureOS and the many thousands of applications available as you would from any desktop computer. This means that we are starting with many thousands of applications that “just work” as a desktop application"
Besides web apps and dedicated apps, you can also just run anbox if you want something from F-Droid.
> Eschewing Android for an actual Linux OS was the point. Alternative Android distros have been around a long time.
Correct, but we have Android _distros_ not phones that support them well.
We need to get some expensive phone from a supported list, unlock it (losing warranty in the process) and flash the distro. Not a thing we can say to even technical users who are not really inclined to HW hacking. To say nothing about the non-technical people who are the ones more vulnerable to the current mobile environment.
> We need to get some expensive phone from a supported list
LineageOS runs on a range of phones from very cheap to expensive. I have owned two Xiaomi phones on which I installed LineageOS that cost only about 100€. Later I upgraded to a Motorola phone that feels fancy in comparison, but still wasn’t more than about 200€.
>If you plug it into a keyboard, mouse, and monitor you will have PureOS and the many thousands of applications available as you would from any desktop computer
But who actually wants this? Id rather want an enjoyable mobile experience. When I want a desktop, I will use a desktop; I dont need a half arsed experience in my pocket.
Desktop and mobile convergence means you can take your desktop with you in your pocket everywhere and use any desktop apps on it whenever needed. It does not have to make the mobile experience worse. More info: https://puri.sm/posts/investing-in-real-convergence/.
> This feature has been around for many years now.
If it's been available for years I'd love to know to find it. There's some attempts in the android ecosystem to do this but they pale in comparison to a proper Linux desktop environment - the main problem being that you're running android apps designed for a mobile screen instead of responsive apps designed for both small touch screens and large desktops.
Google makes such tinkering difficult with every release. I won't be surprised if they have teams that basically look at this part of making it difficult to alter core Android framework.
There's LineageOS (been around for sometime; even more if you count its earlier avatar) and they are usually one or two version behind Google. OEMs are catching up with Android release schedule now because they usually don't touch the core Google features and services (the privacy nightmare land) and usually just add to it and in some cases (unfortunately at that) tweak the UI (for the worse - Samsung has been at it since the day one).
One problem with Purism/PureOS I see is, it is too small and started alone. I hope I am wrong but it may end up not going anywhere other than remaining a vanity phone - that "other" phone - the "privacy" phone.
I can say with 100% certainty that I would never have backed that.
I may not agree with everything about the Librem 5 (in particular I care much less about hardware openness and would prefer better performance + battery life), but the whole point for me was to have "proper" GNU/Linux on a phone, none of the shenanigans that come with trying to use and Android smartphone as a general purpose computer.
I think our best bet to end the duopoly of Apple and Google is to have a cross-platform language that runs on all platforms (iOS, Android, and perhaps the web).
The reason is that there will otherwise be important apps that never end on anything other than Apple or Google OSes, like (for example) bank applications, government apps (for taxes, alerts, covid tracking), perhaps things like netflix or perhaps the BBC news app.
Also, why would developers spend time developing for an open platform, when they can make 10X the amount of money on the closed platforms?
So effectively, we need the spirit of the web, but in the app stores.
(I know, this sounds a bit like embrace, extend and extinguish; but this time it's from the "good" side :)
To end their monopoly governments could just make their business model illegal - that is giving something for "free" and then spying on users and showing them ads aka brainwashing or making sure devices cannot be repaired only discarded.
I think WebAssembly will (eventually) be that language. Write in any language, but compile to one target that runs on any device.
The bigger challenge is the integration with the host OS - we need a common set of OS calls that can be made available to applications running in WebAssembly (eg, WASI).
WASI is the tip of the iceberg for a UI application; its scope doesn’t extend outside of the kind of things you can access with your friendly neighborhood libc.
This sounds a bit underwhelming but I think web apps (in the system browser), Electron and Cordova are as far as we'll get in that direction. (If you exclude projects that are technically more "elegant" but never used in the wild.)
Right, there's a handful of cross-platform toolkits that are in widespread use as well. There's still a difference to full-on web apps though, in that web apps might run on platforms the developers never intended them to run on. The API is independent of the underlying system which means that any platform can implement a runtime environment (browser) and execute web apps without recompilation (kinda like Java but Java is barely used like that anymore).
Arguably this doesn't hold true for Electron/Cordova, so RN is probably comparable to that.
Also, such abstraction is probably difficult to port existing applications to. And there could be a mismatch with some applications (games?) that simply want lower level abstractions.
Yes, but this time it's "free competing with free", or almost-free (or paid for by your personal data, which sadly for most people is equivalent to free).
The power of the web is that it’s accessible from ANY platform and there’s already an app to use it which comes preloaded with whichever platform you’re in: the browser.
That said, native app performance is still unmatched (but not by a long shot!) the react-native project (both by FB and MSFT) looks promising.
There's also Flutter, interestingly by Google (not sure what their strategy is here, as their main interest should be to keep the world's information indexable which becomes difficult when people use more native apps as opposed to the web).
Flutter+Dart is a good choice for this because they already have a goal of supporting 5 platforms with a single codebase: Android, iOS, Web, Windows, Linux and MacOS.
I'm 100% behind Purism and the Librem 5 project (placed my pre-order three years ago), but the "Hold My Order" seems pretty ridiculous. I don't think they should have made that an option at all. I think this is actually my first criticism of Purism/Librem.
They should have just wrote something like:
When the software you donated for is ready, rest assured your Librem will receive the update. Librem 5 is the only phone that gets better with age... (some marketing they've used in the past)
I kind of appreciate that they gave the option. If having X app is a blocker for someone to using the phone, instead of having it sit in my underwear drawer until that app is ready, I'd rather wait for later versions of the phone and more kinks to be worked out.
In my mind giving someone options isn't a bad thing.
It also gives Purism some idea where to focus its development efforts, since Purism literally has a thousand problems to solve so it has to prioritize what is truly important and what can wait.
This is great news and I like that Purism are announcing this. They had previously announced (I can't find it on their blog unfortunately) a flatpak based App Store which is hopefully soon to come.
What I miss: What is the developer story here? Their team , even though it had 2 additions recently, is not very big with 9 people who focus on software. Also they aren't particularily open as to what already works and does not work, which – as the Librem 5 is not widely available yet, nor thoroughly reviewed (still waiting myself, having a PinePhone though) – is a problem.
If you want to have an overview about which apps already exist, I maintain a list at https://linmobapps.frama.io.
It's not very good at monitoring the actual state of the apps and does not look to good. If you have ideas on how to (or want to help) improve it, please get in touch!
If you want a list of the apps for the Librem 5, see:
https://source.puri.sm/Librem5/community-wiki/-/wikis/Freque...
(in the first link, it tells how ready the software is to be used in the Librem 5, and that list is maintained by the developers so it should be up to date.)
Thanks. I will check that list again (which should be linked on LINMOBapps already), there seem to be some games (and maybe more) that I have not added to my list (mostly because I just don't care about games).
What I meant with "state" is stuff like: Does it crash? Is it "feature complete"? Most rating systems (including the one on MGLapps/LINMOBapps) seem to look at "does this fit the screen" more than these aspects, and it is virtually impossible to fit this all into one score.
I love this. As a backer who will hopefully get my phone in the next few months - and who understands that the software is an ongoing piece of work - I am happy that I have the ability to put my money where my mouth is when it comes to asking for some feature.
I will probably mostly donate for battery life improvements, and other system wide improvements.
If anybody from Purism is reading this, I would suggest having "Software Keyboard" as an option - in particular I would love to see the keyboard become more suitable (than what I have seen in videos so far) for terminal usage/programming without an external keyboard. Think Hacker's Keyboard on Android, if you are familiar with that. Basically shrinking a full PC keyboard into a touch keyboard.
They're the same as native Linux desktop apps. For the most part they're using GTK3 with libhandy, but it's also possible to run Qt apps, or any other toolkit really.
They announced the phone in august 2017 and I ordered in september. By october they had collected $1.3m in orders. and my ship date has always been a few months out. :)
Yeah, I bought the DAVx/ICSx pair (as well as CalDAV-Sync and CardDAV-Sync pair) to get Android syncing. I found it more understandable than when configuring Fastmail on iOS given the way Fastmail makes you (or made when I initially set it up years ago) use separate app-specific passwords.
Isn't part of the beauty of app sandboxing in android how it leverages the concept of users in *nix OS? IRC, each app is considered a user to the OS.
Does anyone know what they mean when they say that Apps will be sandboxed? It can't be similar if the phone is running a desktop OS masquerading as mobile OS.
Would "porting" Whatsapp mean that the end result would be able to communicate with users using Whatsapp on Android, or would it just be Whatsapp on an entirely separate network, like Signal?
The former is a lot more valuable to me than the latter.
What is the difference between puri.sm vs Google regarding privacy? Won't user data end up on one of their servers?
Can I self host the phone ecosystem? (For example having my own app "store" or maps)
The best way to think of it is that Google is an identification service.
They make their money by identifying who you are, and creating a datastore with as much information about you and your activity as possible. This is both online data and offline data. Online data is collected directly via google services and indirectly services like google analytics, google javascript libraries, google fonts, recaptcha and more. Offline data is also purchased and stored, for example consumer credit card purchase data.
Given that, it is clear that every free service they offer can be funded by data collection. Are your kids using google docs? Do your friends email you from gmail? do your coworkers use google sheets?
Ok, so purism doesn't make its money from data. It makes its money from selling you a phone. This is a lot less money than google makes, because you pay it directly to them. They can't turn around and get millions from advertisers.
On the other hand, because data is not their revenue stream, they don't care if you store your mail on their servers or on your own or where you get apps or even what OS you run. So there's no need for dark patterns to get you to interact with them.
Yes, you could have your own store. You could have your own OS. You can create and maintain and share your own maps or data or whatever. Actually linux people have been doing this for years and years. check out nextcloud.
A lot depends on what online services you are using. The goal of Purism is to provide a suite of online services at Librem One (https://librem.one/) so you don't have to share your personal data with Google, Facebook, Twitter, etc.
Am I the only one who thinks this project will never succeed? I mean, don't get me wrong I want it to succeed, but in this ecosystem of apps/capitalism, and Apple/Google duopoly, I see no chance of an open project like this going mainstream, and the intersection between the tech enthusiasts (as I really don't think this project is aimed at the non tech-savvy) and who actually would want to replace their phone with this isn't enough (HN community isn't a good KPI for this, obviously. We need to check out the general audience) to support the costs.
Well, I've read it, but didn't even convince me a bit. All the reasons that they claim for succeeding against Android and iOS are very, very niche corner cases where 99.9% of users don't care about.
This looks pretty great, and shows progress towards becoming a replacement for my iPhone - but, one question - what are these apps? Am I missing the ability to navigate to a page describing each of these apps? I can probably guess what some of them are, but doesn't it seem a little strange to have a list of apps, but no descriptions for the apps available?
(On Safari, MacOS, in case its just an anchor bug..)
When I see things like this, and aweSIM, and the not-yet-materialized Purism USA Edition (and the price tag associated with that), and the exposé by a former manager, and their unethical resale of open source software... I can't help but wonder how well Purism is doing, financially speaking. It seems desperate.
" Many people believe that the spirit of the GNU Project is that you should not charge money for distributing copies of software, or that you should charge as little as possible—just enough to cover the cost. This is a misunderstanding.
Actually, we encourage people who redistribute free software to charge as much as they wish or can. If a license does not permit users to make copies and sell them, it is a nonfree license. If this seems surprising to you, please read on."
I agree on all points except "unethical resale of open source software". If they comply with the license there's nothing unethical about it.
Selling open source software is one of the valid ways open source reaches end users. Users/Consumers want to buy polished, maintained, and supported things. That costs resources to provide.
From a user perspective, having a preference for FOSS software shouldn't limit your options to unsupported abandonware you couldn't pay for if you wanted.
Recall the F in FOSS is Free as in Free Speech, not Free Beer.
This reminds me of something else. What would be the best model for a FOSS "app store" with mechanisms for compensation? Lowering the barrier for transactions to FOSS projects really helps their survival and sustenance by individuals who cannot contribute code.
Here's a very half-baked thought: One can imagine users supporting community servers (matrix, activitypub/mastodon, etc.) and applications through a cryptocurrency wallet that is loaded in the same way that people load their g--gle/a--le wallets. One can also imagine applications that can tie into this wallet to compensate publishers based on some model (a transaction API like g--gle/a--le or something like what brave is doing with the BAT). My question is what are the failure modes of such a model?
For example, this still wouldn't stop the perverse incentives of:
- Content publishers going for the types of bait that generate the most clicks. Clicks=compensation regardless of it's from ads or some sort of attention token.
- Application developers casing addiction mechanisms that go after whales. The same loot-box mechanisms can be implemented with a FOSS game that uses a crypto wallet API.