An unexpected revival of Firefox OS | Hacker News Reader

An unexpected revival of Firefox OS

tuxphones.com

275 pointsredbell158 comments

FirefoxOS was one of those hail-mary longshots that you always thought would be destined for failure but you had to respect the fact that they had the guts to attempt it anyway. (Though to be fair, Rust was the same sort of project, so sometimes long shots pay off.) It showed that Mozilla understood the existential risk of ceding control of the platform itself to your direct competitors. I'd love to see it come back and I'm sad that I never got my hands on one of the original phones.

Lot of people got pissed off they wasted resources on it

Although, in fairness, that could be said of literally every project Mozilla has launched that isn't either a Web browser or MDN.

I think they have a pretty good future in web services. I'd be happy to pay for a mozilla password manager/vpn/email hosting/cloud storage. But for some reason they only sell the VPN and not in Australia.

That’s a rebranded Mullvad VPN.

If you want the VPN in Australia, you can get it directly from Mullvad - they do have servers there.

Yeah that's what I do. But I'd be happy to pay an all inclusive subscription to Mozilla for a full set of web services and drop my subscription to google and mullvad. But they just don't offer a product I can buy.

That is missing a lot of context. They wanted everything to be Javascript and running on a $50 Smartphone. ( I think initially it was even lower than $50, but they rise it to $50 a bit later )

If there was ever anything that show how Silicon Valley and Tech, the modern meaning of "Software people" with little to zero understand of hardware and supply chain. This is it.

And Mozilla wasn't the only one, even Bill Gate said all hardware will eventually cost zero and be subsidised by running software and services.

The lack of understanding of hardware and Supply Chain, even after COVID and in 2022 still greatly baffles me.

Yep, that's why I grinded my teeth a bit when I read the post title. As a broke ass student back in the day, I counted my pennies to buy one of those devices. Even shilled it to my peers, some of whom also bought devices.

Inside Mozilla too. Some execs thought Firefox OS would never be competitive because Google could always make a cheaper Android phone that still supported the whole Android ecosystem. Others wanted to move desktop Firefox to maintenance mode and bet everything on Firefox OS.

They should’ve built a cross platform UI toolkit based on Servo, and built Firefox OS on top of that. That would have given it the efficiency it needed to be successful. They’d have taken a lot of market share very quickly if their apps could run on android/iOS too.

I owned a Firefox phone and it was awesome. It had understandably few apps, but it worked really well.

One thing I really appreciated was that it was conceived as an appliance rather than as an attention grabber. We tolerate our phones being attention grabbers, but unrightfully so, it doesn't have to be this way. The default UI even had the saturation toned down, to not be so much in your face.

It was pretty capable for an 80 Euro phone. It was the only smartphone I owned that I could feel vibrate in my pocket. I concede that Android ran better on it, but it's not like FirefoxOS ran bad, on the contrary.

I unfortunately lost it, and it was about the time it was being discontinued. I would probably get another one. Heck, I would get one now, if they were still around.

Anecdote: friend had a KaiOS and it was pretty aweful.

KaiOS is based on FirefoxOS and sells more phones than Apple does.

That's interesting. I went on their website and it was nostalgic browsing these phones.

Which phones use KaiOS?

There's a list on Wikipedia[1], not sure if it's complete; the company behind KaiOS has a list of devices too [2]. Important ones include the JioPhone, which was part of Jio's large push into the mobile market in India. There's also a nostalgic Nokia 8110 4G (Banana Phone); there's even some with modems appropriate for the US market.

[1] https://en.wikipedia.org/wiki/KaiOS#Devices [2] https://www.kaiostech.com/explore/devices/

Anyone want a Bananaphone? I tried to use it for a while, but AT&T shut down their 3G calling network and it doesn't have the radio for VoLTE, and AT&T isn't doing 4G calling - they jumped straight from 3G to VoLTE. It also doesn't do MMS for group texting. I'm easy enough to find, I can toss the thing in a box. On the plus side, I got a Flip IV out of it. I ran with it for a year or so before the keyboard got so bad I gave up - when you can't dial a phone number without having to correct multi-presses and such, texting is damned near impossible.

Isn't 4G calling and VoLTE the same thing? A 4G phone without VoLTE uses the 3G network when placing/receiving calls, hence why the 3G shutdown is causing pain even though non 4G phones are really rare these days.

I don't believe they're the same thing. I can connect to the 4G data networks with the Banana, but not make calls, and I'm not sure as to the details of why - going deep into cell network reverse engineering wasn't my goal here, I wanted to experiment with the state of non-iOS, non-Android cell phones in post-2020. I got a lot of data, but... well, I'm back to iOS, so I couldn't make it work.

There are KaiOS devices that support VoLTE and work on US carriers, but the Banana doesn't seem to be one of them. In addition to its many other pain points. Those keys suck to hit...

Doesn't it work on the 2G network?

There is no 2G network in the US anymore. Or at least, very little. AT&T shut down theirs in 2017; I thought T-Mobile shut their 2G GSM around then too, but sources say T-Mobile's 2G shutdown is end of year 2022.

Many EU carriers have planned to keep a small amount of 2G for industrial devices, old handsets, etc, but US carriers have not.

I had similar issues with TMobile. Bought the phone for backcountry use outside of the States, and I sold it on eBay after I got back because it was useless here in the US.

It does not appear that KaiOS is something that anyone should own.

"The first major commercial adoption of B2G/Firefox OS was seen around 2018 with KaiOS, a widely popular commercial fork of Firefox OS that promised to be a modern OS for inexpensive feature phones. Although KaiOS turned out to be disappointingly closed and oriented to tracking users and shipping advertising, it was quickly forked in GerdaOS, a custom ROM that promised to bring back some of the openness of Firefox OS to KaiOS devices such as newer Nokias, and kill the intrusive user trackers in the process."

Can generally confirm. Not from a "tracking" perspective, but from a "General usability in the modern world" perspective. I spent a year with a KaiOS Flip IV (about a $70 AT&T phone), and finally gave up and went back to a smartphone when the keyboard got so bad I couldn't dial a phone number without massive corrections, much less text.

It felt, first and foremost, like a device built by someone who had been told stories about the phones of the first 5-8 years of the 2000s, but had never seen one, never used one, and wasn't quite sure how it was all supposed to really work. Lots of really weird pain points like predictive text not capitalizing the lone "i" no matter what mode I was in. I started trying to write texts without "I" in them to avoid having to perform the 72 keystroke sequence to capitalize it - I don't like being lazy in texts about things like that.

The performance wasn't bad initially, but it felt like there were some seriously non-optimal data structures involved in everything. A text thread with more than 50 messages or so started being slow to interact with, God help you if it's got more than a few pieces of media in it. And if the phone had too many text messages in general, the whole texting app would lag badly (seconds to open, seconds to see a list of conversations, etc). I had to very aggressively trim old threads, and sometimes even current ones, to try and keep the texting app responsive.

It also got oddly prone to "Oh, yeah! There was a thing I was supposed to notify you about! You're looking at me now, I remember, you got a text! chime" sorts of behaviors I could never run down the cause of. I'd be texting someone, waiting for a reply with the phone closed, and... nothing. Magically, as soon as I opened it 10 minutes later, I'd get the notification that they'd sent a text. I hate to say things like "You have one job," but "Notifying me that there's a phone call or text message" is sort of the core function of a modern phone-type device.

The browser was a joke. I could browse Hacker News and that was about it. Anything more complex would fail to render, or just crash the browser. A modern web capable device it was not.

I think I gave it a fair shake for a year. But I was still doing things like "Carrying an iPhone, powered down, for when I needed something I couldn't do on KaiOS." It was mostly either drone ops or a particular building access app that I needed often enough, so I don't really fault it, but it was annoying not having them on my carry device. Same goes for "literally anything but SMS." Attempts to get a Matrix client on the device didn't go very far (there is one, but it doesn't work very well), Signal was a no-go, and nothing else was useful either.

It did have Bluetooth, though. And I did like being able to cart around music on SD card. Don't carry too much, though, the music app couldn't handle a ton.

I moved to it when Apple decided to push on-device CSAM scanning, and... a year and another OS revision later, they've seemingly not done that. And then iOS 16 added Lockdown, which I decided to play with, determined did everything I wanted in such a feature, and I'm back to an iOS device with very few apps installed. I never had warm fuzzies about the security posture of the KaiOS ecosystem either - Apple, I do generally trust to try to secure their devices. KaiOS, nothing about it really felt robust against a remote attacker, though I've no evidence that it's actually full of holes. It just never gave any real confidence that it was more than a best effort, YOLO'd sort of thing tossed over the fence.

There were some fun little games, but they were all very peppered with advertising. At least you could sideload stuff.

Anyway... it's a thing, I tried it, and despite my exceedingly high tolerance for tech pain, I couldn't make it past a year. Failures on my part, I suppose. I could have probably dismantled it and repaired the keyboard...

> The browser was a joke. I could browse Hacker News and that was about it. Anything more complex would fail to render, or just crash the browser. A modern web capable device it was not.

That's the one thing that surprises me the most in your post. Any idea if it was actual browser engine issues, or running on inadequate hardware?

Both, far as I could tell.

I believe the device has 512MB RAM and a quad core 1.1GHz A7 - so not exactly rocking the hardware. But the browser engine didn't have things like some basic Javascript support, near as I could tell - Discourse forums wouldn't load properly, grumbling about how I ought to enable Javascript. And on that little CPU and RAM, anything "faintly modern" just wouldn't work right at all.

Not that you can see much, it's a 240x320px screen.

There are some higher end KaiOS devices out there I debated, but spending a couple hundred on that level of functionality just didn't seem a good use of funds to me.

Can confirm, both. KaiOS has tons of bugs and vendors usually ship old buggy versions. But the hardware itself often dies or has flaws, because it's the cheapest phone you can make with modern features (bluetooth, wifi, 4G LTE, a multitasking OS that can "run apps", etc).

The problems I see with KaiOS are 1) lack of aggressive development to fix bugs/features, 2) lack of vendors upgrading the OS on new handsets, 3) lack of vendors upgrading the hardware to run newer versions of the OS. And the market doesn't really care (or have a choice), they just need the cheapest phone possible that can check the weather forecast or email.

Until both developers and vendors give a crap about building efficient technology, they will continue to churn out bloated, buggy software on cheap hardware. Even Palm Pilot was more advanced than the crap churned out today.

> Even Palm Pilot was more advanced than the crap churned out today.

I would pay a lot of money for a Palm Pilot designed for 2022. The team that created Palms were gods.

Palm didn't have a fully-featured browser but pre-rendered pages server-side (that solution was called "Access" I believe). Palm had nowhere near the power, emulating Thumb/68k instructions on a handheld. FirefoxOS was on to something, and even WASM could maybe provide apps without an app store, but with the incredible web bloat of the last decade we got neither lean nor portable web apps.

>Palm didn't have a fully-featured browser but pre-rendered pages server-side

iirc opera mini browser did something similar. The browser would tunnel all requests through their server and render and optimize the page before sending it back to the browser

That's unfortunate, if no less confusing. Any chance this happened before https://techweez.com/2020/03/11/mozilla-will-support-kaios-i... ? I'm just trying to square the statement that it has a fundamentally broken browser engine with the claim that it's still being supported directly by Mozilla. (Not, to be clear, that I doubt you for a moment.) I wonder if they did something like intentionally disabling features (say, JS) in order to make it fit on the hardware. I mean, ultimately kind of whatever, because I'm hardly going to use a proprietary stack like that anyways, but it's puzzling that browsing is the weak point on a Firefox OS fork.

Not sure. It was a Flip IV I ran from about Aug 2021 to Aug 2022, on KaiOS 2.something.

> Magically, as soon as I opened it 10 minutes later, I'd get the notification that they'd sent a text.

Android does this in power save mode. Plug in a charger and all of a sudden you get notification pings as background tasks are given a chance to run.

>Lots of really weird pain points like predictive text not capitalizing the lone "i" no matter what mode I was in.

My modern Android phone is like this now; it pretty much never capitalizes lone 'i's. In fact, the whole way auto-correct works is completely brain-dead; it's constantly trying to force me to use wrong words and wrong spellings, even when I try to back up and correct it. It's extremely frustrating. And judging by all the errors I see in other peoples' texts, it's not just me.

Same here - Samsung or Google keyboards same, they must have secretly agreed on doing this crap. I disabled correction and by looking at the word suggestions I'm positive I will never reenable it. Any usable Android keyboards out there which do also swiping and don't suck all your data?

I'm debating whether I should give up and disable all auto-correction. It does save me time when I fat-finger words, hitting the letters directly adjacent to the ones I intended, but then it wastes me time when it does what I described before.

Honestly, it's been decades now. How hard can it be to make auto-correct not suck? There's so many utterly stupid things it does which should be fixable with a few lines of code. For one, how about not auto-correcting proper names (which you can tell because they're in the middle of a sentence and start with a capital)? And how about making a lone 'i' capital? And how about not auto-correcting at all if the user has backspaced over the auto-correction and fixed it?

You'd think the eXTra sMarT kEYboArD would have learned in all these years at least how to spell my own name correctly but guess what...

https://f-droid.org/en/packages/com.menny.android.anysoftkey...

I've just about given up on autocorrect on phones. iOS, Android, doesn't matter: every single one I've tried is absolute garbage. Much easier and faster to just type everything manually than to rely on autocorrect shortcuts that seem to go out of their way to ignore everything that should be corrected while "correcting" everything that shouldn't be corrected.

I'm of the opposite mindset. I don't want any text correction to apply any form of capitalization because it indicates to the reader that I am on a mobile device. I type all in lower case on desktop and I want to be able to do the same on phones. Sadly my OnePlus disregards the Android setting to do this (which you may have inadvertently set).

Are you posting from mobile or desktop right now?

My Note 10 has corrected "you" to "U" - I'd seen the word switch before my eyes. Nevermind that it corrects "i" to "U" sometimes as well. And I've never written a text ever with that lazy contraction of "you". It's like the stupid is being forced upon us.

> predictive text not capitalizing the lone "i" no matter what mode I was in. I started trying to write texts without "I" in them to avoid having to perform the 72 keystroke sequence to capitalize it

Surely hyperbole and off by an order of magnitude...?

Not if you count that as cumulative time spent (wasted) over the course of a day… (video).

Video: https://nexus.armylane.com/files/kaios-I-test.mov

"Per 24h" is an arbitrary and specific period of time which should be included with the just as specific numerator.

I liked the idea of a feature phone you could install apps on. KaiOS wasn't it, it's just a terrible feature phone that isn't actually extensible in any practical way. You can't really write your own apps for it, and it's also not even good at being a traditional feature phone.

Only in India. Apple iPhones are way too expensive there.

>KaiOS is based on FirefoxOS and sells more phones than Apple does.

KaiOS is used on 160 million devices around the world.

Apple sell more iPhone than that in a single year.

They may have ment the are dozens of diffrent models to choose from.

[citation needed]

how many planets do you guess would it take if there were 9 billion people carrying, patching und using Apple phones?

I have read some terrible reviews on KaiOS, UI lags and freezes, shipping broken updates, terrible battery drains. I'm sure there are some devices have good versions of KaiOS, but with this track record it's like winning on lottery. SailfishOS from Jolla seems to be a much better alternative for a non-Android non-iOS device that is practically usable.

How is it that comments like this one can simply get away with such empty claims without any evidence supporting it?

Extraordinary claims require extraordinary evidence, and the burden of proof is on the claimant (the parent comment) to provide it.

Until then, your claim is simply baseless and can be easily dismissed.

I miss Windows Phone. Used to have a Lumia 920. Sleek, rock solid phone that rarely crashed and always gave the impression that it will be there when I need it. Apps weren't great but still I was quite happy with a few. And I started to like the Tiles UI.

Then the Microsoft-Nokia joke happened.

i miss the Zune and its UI. was the Windows Phone's UI similar?

Never used Zune. WinPhone's UI was similar to Windows 8 at the time (I think the term is Metro design language).

A little bit, yes. The scrolling menus were in the style of Zune, and the home screen was like that of Windows 8. It was fast and lag-free, no loading of UI elements, and all that on a dual core ARMv7 (I used the Samsung Ativ S).

A different generation of windows phones but I used HP Ipaq 514 back in the day and oh was it terrible to use.

I think that was Windows Mobile which is different than Windows Phone.

I used to like the Windows Phone UX too, I had it on a few phones over a period of time back in the day.

But, all of the phones I had used to crash pretty regularly, rock solid it was not! Still, I miss it!

Weird, I had one of these too and it crashed constantly- I would reach into my pocket and if the phone felt hot I'd know without looking that it had crashed and had already drained the battery

If Nokia wasn't gone use Android, using its own much, much better MeGoo OS would have been much better. That's what the later Lumina hardware was initially designed for.

And MeGoo was better then Windows Phone as well.

> Then the Microsoft-Nokia joke happened.

The real joke and borderline criminal act was for Nokia to adopt Windows when they had a much better open OS already.

> I'm sad that I never got my hands on one of the original phones.

I was at Mozfest in London and they gave everyone a phone, that was in 2014 or 15, so I would be surprised i nobody still had one kicking around.

Would love to know why this of all things was downvoted

I had one of the original phones, used it for over a year because I was broke, it was free (work gave it to me), etc.

It wasn't terrible, but it had a LOT of bizarre bugs and couldn't compete with even the cheapest Android (the kind that phone providers basically gave away for free at the time).

It was better than the old Nokia Symbian S40 device I had though.

I ended up moving to a super low end Android device that I got for free from a provider for porting my PAYG (prepaid) number to their service (also prepaid).

I also used one of the Nokia Windows phones for about a year as a student. It was good for calls and SMS, but fuck all else worked. Ended up lending that to someone who lost it.

It's strange to me that people would build a budget phone OS on top of web technologies - after all, web-based desktop software can grind full-fat tower PC to a halt.

I can't imagine it working out for low-end mobile hardware.

Firefox OS was arguably bad (especially when they redesigned everything, breaking half of the features I used in the process) but it was surprisingly snappy given how underpowered the hardware was: I actually had two phones running Firefox OS, the GeeksPhone which ended up broken after a year and a 15 meters fight in a staircase, and then the Motorola one, both with very limited hardware and yet it never felt slow.

My experience with Android always felt sluggish since then (given that I never buy expensive phones).

Nothing forces you to write slow websites.

I was grumpy Android user at the time FirefoxOS came out. To me it was the dream phone that will get me away from Google tracking so I bought a Flame device as my only phone. This meant I had to convince everyone in my family and network to move away from apps unsupported on my new open device, so moved to Signal instead of WhatsApp for example. I was very disappointed to see the project killed by Mozilla but still went ahead and got a Banana phone with KaiOS, only to discover it has as many trackers as Google but without much of the usable features!

This was the straw for me that broke my faith in open phones and went full iPhone since.

This is the saddest story I've read in a while. Especially the ending.

Frankly, this is the same as my story. Except the open platform in that case was Android - or at least it was the open platform when I started using it, 13 years ago.

After some point it started declining constantly, and I just had enough so I switched.

Indeed my journey of disappointments has too started with Android. I must say my experience outside phones (laptops, desktops and servers) are very positive with combinations of BSD and Linux and I hope it remains the same.

Mobile is a tragedy with no cheerful endings.

You either get your privacy heavily violated, get yourself protected in a costly walled garden or get yourself stranded in a high maintenance, low guarantees software universe for a device that has ever increasing importance in daily life.

I guess it was too early to get rid of the Google tracking. Now the situation is significantly better: it is possible to daily drive Librem 5 or Pinephone, as many people report (me included).

I moved countries last year and used the new phone number as an opportunity to get a banana phone. It doesn't have Signal, its Whatsapp doesn't support being connected to the webapp, its T9 implementation is absolutely horrific, and its 4G reception was very poor. After six months the 4G stopped working altogether and I reluctantly moved to a cheap Android phone by Motorola.

I wanted to be a believer, but it just…wasn't it for me.

The best semi-open phone experience you can get is a Pixel with CalyxOS or GrapheneOS. Everything will work because its built on android, but with full privacy and better security.

Not GP, but I looked into that option. The problem to me seems that most popular apps depend on Google Play Services in some way. So your options are (a) use microG (which seems pretty janky), or (b) install the real Play Services, which puts you back at square one.

Or c) use alternatives to those most popular apps.

Also, "seems pretty janky" is quite the shallow dismissal.

Re: alternatives -- see this comment chain a few days ago: https://news.ycombinator.com/item?id=33103321

Re: jank -- per the official microG wiki [1], most of the functionality is only partially implemented and/or has bugs. I don't see how it's shallow to consider that as janky.

[1] https://github.com/microg/GmsCore/wiki/Implementation-Status

I use MicroG daily, and never faced issue with apps it all. Everything works

Clayx ships with MicroG and almost every app works with no problems.

GrapheneOS has a GMS install option, but it installs it as a non root service, so you are still going much more private than stock Android.

Not everything, as I understand it - some apps (such as Google Wallet and some banking apps) rely on Google's SafetyNet API and won't work on those systems.

It's a shame because it's probably only a handful of apps but it's the ones that are important to me.

Calyx has a hack around SafetyNet, but it breaks often

It would be great to see a minimalist version of this that is just completely rock solid.

I don't care about many apps. I need maps, and a camera.

My streaming service can play from a browser tab. Someone could make a decent podcast site that does the same. The OS should make it easy and reliable for such sites to use 10's of GB of local cache space.

Beyond those requirements, I'm mostly good. (A calculator, flashlight app, etc would also be nice.)

Anyway, I hope that (unlike 100% of these open source initiatives so far), they focus on keeping it simple.

> I don't care about many apps. I need maps ...

As someone with a hobby project in this area I can confidently say that's not simple.

Either you implement a new simple map app (very hard), or you add a comparability layer to run an existing map app (never done it, assume very hard and also not minimalist).

Edit: Here's an example of an impressive 10k LOC attempt at a new map renderer. Look at what it supports so far. (This map is representative, the renderer just doesn't support things like text yet) https://webgl.demo.maplibre-rs.maplibre.org/

There is a map app for KaiOS that essentially just wraps openstreetmap https://github.com/strukturart/o.map

Thanks! I forgot that on KaiOS you can just use web libraries. That's using leaflet.js, a excellent older library that's perfectly fine for most use cases.

Wouldn't web actually be a good example for browsers to leave the native behind? A lot of native map use has never moved beyond the generation of mapsforge (vectors rendered in the fly into bitmap tiles that are then used as bitmaps), because apparently that's good enough. In the browser, there's everything that evolved around MBTiles and the only thing missing would be some offline repository API filling the gap between a system map and isolated PWA containers.

I don't think you're right about the evolution of map rendering approaches. Isn't the vast majority of native map use Mapbox GL?

You're right that linking to a vector renderer isn't too much work. I was commenting on the simplicity of writing a map app from scratch.

Sure, there's a lot of mapbox on the native side as well, but that is far more restricted to high profile customers (e.g. the Strava app) than in the browser. At least from my superficial impression, libre forks of mapbox legacy code seem to be far more alive in the subset related to the web than in the subset related to native.

One cause I think is the lure of offline capability: when I'm in the browser I don't mind loading on demand, but as soon as native enters the picture I'll want at least an escape hatch to downloadable maps. And then the Pandora's box of provisioning, backwards and forwards compatibility is wide open and you happily stick to mapsforge or whatever it is you already have.

(ps: and my post above was supposed to start "Wouldn't maps actually" instead of "Wouldn't web actually", but you seem to have read intention instead of words just fine)

> least from my superficial impression, libre forks of mapbox legacy code seem to be far more alive in the subset related to the web than in the subset related to native.

Just based on my superficial impression I think normal apps on the play store pay mapbox if they don't have a big budget and consider maplibre (the fork of mapbox's most recent approach) if they're big enough.

My surface impression is maplibre is used by companies like Facebook who can afford to hire low level graphics devs and fork mapbox's work rather than pay the new license.

For every feature, including offline, I think paying mapbox is the cheapest option with decent quality up to that scale.

Of course, there are projects like mine that aren't motivated by profit. (I'm using mapnik, an independent oss project that got significant mapbox investment years back but which they now consider legacy).

The e-ink Light Phone has directions and a map [0]. I wonder how they do it.

[0] https://support.thelightphone.com/hc/en-us/articles/44021234...

Based on a cursory look that seems closed & proprietary, so I'm not sure.

Their privacy policy reveals online directions are sent off to a mapbox competitor

> We've built the Directions tool, like all of our tools, with privacy as a priority. Light does not have access to any location info or logs of your trip history. These trips are handled privately without user identification using our navigation partner HERE.

I suspect everything maps related on-device is a thin layer over the HERE SDK.

I suspect that's possible because LightOS is actually a skin over Linux or Android.

Funnily enough, that sounds like a ChromeOS phone.

That's more or less what FirefoxOS was supposed to be: ChromeOS, but with Firefox instead of Chrome and on phones instead of netbooks.

That is selling it short. In ChromeOS, the browser might be front and center, but the internals of the OS are mostly normal OS stuff. FirefoxOS had JS/HTML/CSS used all around the OS itselft besides the browser app/experience.

No I don't mean they used JS to write the HAL or anything like that. I'm saying that whole user-facing part of the OS is a WebView. Everything that the user interacted with was part of an HTML5-based application. The window manager, the process switcher, etc. There were custom JS APIs to handle all that. ChromeOS is quite different in that respect.

Firefox OS inspired me to build my own Debian-based operating system that uses web technologies for its UI and apps. It uses Electron as the runtime[0] and works both on desktop (AMD64 arch) and mobile (on the PinePhone). It's very much a work-in-progress project and is by no means complete, but it's progressing nicely over the past few months and I'm sure will become quite a complete OS functionality-wise in no time.

Feel free to check it out here if it piques your interest: https://liveg.tech/os

(There's also plenty of videos that I've previously made that showcase some of the key moments in its development: [1])

Capyloon IMO is also an awesome OS platform that achieves very similar goals to this project, and it'll be great to see what they build next within the world of non-mainstream web-centric operating systems!

[0]: A tad controversial, I know — but the rationale is that using one Electron app as the desktop environment means that you don't have to run Slack/Discord/Spotify etc. in their own Electron apps; instead, they can run as web apps that share the same Chromium instance. That way, you don't have completely separate Chromium instances/processes running at once, where they'd otherwise all be consuming lots of RAM due to their startup overhead.

[1]: Playlist: https://www.youtube.com/watch?v=pmOR2J9fDZM&list=PL3aAeH1lqy... — Prism is a prototype smartphone that runs LiveG OS, and gShell is the desktop environment.

Looks cool. The fact that same shell runs on both mobile and desktop conceptually reminds me what Canonical was trying to do. Being based on web technologies, couldn't you make a browser runnable demo for the interface? Such as the Flutter-based shell https://web.dahliaos.io is doing.

Oh wow, didn't know that dahliaOS had this! The only barrier that I can see would prevent this from happening for gShell is that web browsers don't support the `webview` element that Electron lets you use (it's basically an iframe but without all the CORS restrictions and the likes, so you can see why browsers don't support it). I suppose one cool side-project would be to set up a service similar to Microsoft's Windows 365 that would stream a desktop environment to something like a `canvas` element — that way, we could get around all the CORS stuff.

Also, yeah what Canonical did with Ubuntu and Ubuntu Touch is pretty cool in terms of convergence (apparently it's really stable, too). I do find having the sidebar UI on mobile a bit odd, but I should imagine that it'd be something to get used to. Phosh and KDE Mobile I think are both convergent, too.

My first thought given the name was 'does it have anything to do with Google?' (sorry)

Because of the double O?

Probably because of the g in the subdomain, kind of makes it feel like "live g[oogle]"

Ahh yeah I can see how people might think that with the name "LiveG"! The previous name was historically "LiveseyGadget" (my surname + "gadget") a long time ago when I came up with some of the ideas, but for all intents and purposes, this got shortened to the now-official name "LiveG" and the meaning behind the name essentially dropped to be more of an open project.

I don't intend for this to be critical. You've created something pretty impressive and are giving it away which is awesome. But since you've acknowledged the decision as controversial I'll say it - web technologies for UI kind of suck compared to just about anything else. They are hard to reason about, they are fragile, slow, and heavy. They are a house of cards build on shifting sands.

Now I know there are people here who hold the exact opposite opinion. They've worked extensively with Carbon/NextSTEP/Cocoa/UIKit, or OS/2 PM, or Win32/WinRT, or even Swing/JavaFX and then moved on to web UIs and love it. I don't get it though. I sometimes wonder if the web UI fans are mostly people who want to work and play on Linux and that's about the only way software makes it onto or off of that platform these days - as Electron apps.

> web technologies for UI … slow, and heavy

It’s not the web platform features that are slow and heavy. It’s all the JavaScript on top of it. You can make a UI on the web with just HTML and a bit of CSS. Nothing beats that in terms of performance and small size*. But if you add several MBs of JavaScript frameworks on top of that, then it becomes what you described.

*Try the HTML version of Gmail in your browser https://mail.google.com/mail/u/0/h/. Just HTML and CSS. It’s the fastest email client in the world.

edit: I timed it. HTML Gmail loads and fully renders in about 1 second.

> Try the HTML version of Gmail in your browser https://mail.google.com/mail/u/0/h/. Just HTML and CSS. It’s the fastest email client in the world.

I use Fastmail on a regular basis and it's snappier than the HTML version of Gmail.

How long does it take to load and fully render HTML Gmail on your device? For me, it’s 1.15 seconds, and that’s from the moment I click the above link to the moment rendering is complete.

> Nothing beats that in terms of performance and small size

Native controls do. For example, Steve Gibson wrote a DNS benchmarking utility[1] and the executable is 159k. No matter what you do, an HTML & CSS is still being rendered by a slow interpreter.

BTW, when I clicked on that link for HTML GMail, Google made me click through a warning trying to scare me off. I'm not sure why they would care if I use the HTML version.

In the end, I don't think the speed of CSS + HTML controls matters much. It's limiting and not a great choice for an operating system UI toolkit.

[1]: https://www.grc.com/dns/benchmark.htm

I don’t understand why you’re mentioning a DNS utility.

You couldn't write a web-based version that was anywhere near as fast and small. A browser sitting there showing a blank page is bigger.

Yeah I understand, there are certainly a lot of interfaces out there that are built on the web and because of all the JS they use, they can feel really clunky. The reason why using the web as a platform for this system is because it is — in my opinion — the most viable tool we have at the moment to build cross-platform apps and services that work on all modern desktop computers and mobile devices from the same codebase (that is, without having to go down the React Native or Flutter route).

I do consider performance issues with the web to be quite serious, which is why I've avoided large JS frameworks such as React, Angular, Vue etc. and have gone for a lightweight one that was made specifically for the project (mainly as an experiment). The results do show: the system is rather responsive on an original PinePhone (and I should imagine is really smooth on a PinePhone Pro due to its much better specs), and you can get to the LiveG OS lock screen within 10 seconds of booting the system in VirtualBox.

But yeah, I do see what you mean by the web being... clunky in general. You say about the web being "fragile" and "a house of cards built on shifting sands", and yeah, the culture around developing web services has largely been to use waaay too many dependencies to achieve simple tasks.

I’d love to see FirefoxOS come back. I had a mini “candy bar” phone years ago that ran it, and although it was terribly low-spec, it had a better user experience than my flagship Android phone at the time.

https://capyloon.org/

Capyloon for freedom, runs most anywhere ( Desktop too )

It runs FirefoxOS apps.

Runs on top of Android on any phone as a GSI.

Comes with a SDK out of the box

Pragmaticly free in that it runs anywhere

I had a few FirefoxOS devices when it was new. The idea was very appealing, and I had my own open source handheld intentions for it.

One interesting aspect was that it was being marketed as a firmware for low-end phones, but obviously seemed desirable for midrange and high-end as well.

One concern was that the build process (built atop Android) seemed crazy-unauditable and possibly unsustainable. I hoped that the FirefoxOS parts would eventually be transplanted to atop a more manageable open source hardware/device layer.

Towards the end, there was also an unclear funding(?) partnership, involving the home screen and a particular company, which didn't seem consistent with my goals for an open and trustworthy device. I don't know how that partnership turned out, and can't immediately find info on it.

Today, with how far we've stretched "Web"-based JS, massive stacks, software provenance, and control of data, I'd be more interested in a device implemented 1/100th as huge, and 10000x as trustworthy.

> Furthermore, the Capyloon page proposes WebAssembly and IPFS integration as the main priorities of the project, but it is not made immediately clear how this integration works in practice.

Yeah, I remain interested in Capyloon because it's a plausibly-decent-looking smartphone OS that isn't Android, but the IPFS emphasis always struck me as... well, I can kind of understand the "putting the user in control" angle, but pragmatically I would not have started there. Of course, it's a FOSS project by volunteers so if that's what people happen to be interested in then that's what'll happen:)

Very honest question: where would you rather start and focus on? Feel free to join the matrix channel (https://matrix.to/#/#capyloon:matrix.org) if you want to chat about it!

Skimming through the comments about how bad is KaiOS has destroyed my pretensions of getting rid of my Android smartphone and use a dumbphone instead.

For me, it would be great to have a simple, privacy oriented phone with a decent camera and, in terms of apps, just maps, Spotify, Telegram, Whatsapp (sadly still having friends/family there) and MFA. (Banking app from the bank I use would be great but that's too much to ask for). I guess the sweet spot is gonna be in a middle range smartphone compatible with Graphene or Lineage and install a minimalistic launcher in it.

I enjoyed my Nokia 2720 flip phone for half a year. It’s got a Google Maps web app that supports voice search, enough for looking up hours or a phone number. The camera is bad, but enough for photos of products or things to remember.

If you know what you’re getting into, it’s not a bad (smart) dumb phone, especially if you are looking to get away from a smartphone. I rooted mine to install a QR code reader and app for ePubs.

CSS and HTML are both from the web but different things. It makes sense to use CSS to control the style (therefore it is named Style-Sheets) in toolkits like Gtk and Qt! While JavaScript is an interpreted language for the web but not for applications.

Yes. JavaScript can be used on the desktop, we name it usually Electron. It actually replaced Flash! It allows companies to copy their web-code to the desktop and relabeling it as application. This is cheap. The users suffer from slow applications, which require a network connection, which consume high amounts of resources (memory) and provide bad integration. This is expensive. Microsoft Teams tries currently to switch from Electron (which includes Chrome) to Chrome itself.

I'm rather happy with the GNOME-Shell. It comes with its own implementation of GJS which changes the situation. The GNOME-Shell itself suffered from memory leaks which are fixed for some years:

https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/64#note_...

I'm thankful for the authors description. Replacing language A by B or B by C doesn't improve the implementation. Every implementation requires a lot work and constant care. And the comment describes how they control memory usage rather well in C/C++ with ASAN (GCC or CLANG) and Valgrind, suffered from incompatibilities between GObject and JavaScript and memory leaks within JavaScript. The implementation with JS allows for immediate manipulations of GNOME-Shell through developers but I wonder if the would prefer a native language when doing it again. At least that would reduce complexity and remove the dependency on the interpreter.

I still own one of the Firefox developer phones pictured in the article (orange shell) that got donated to me on the promise to port some app to HTML5.

I too still own the same one. I did run it as my main phone in 2012.

I had a Sony Xperia running FxOS as my (only) phone until this year, when it finally succumbed to its swelling battery. But I'll admit I am a fairly minimal phone user.

Now would be the time. With more and more languages getting Wasm compile targets, apps can now be completely running in the browser.

Isn't KaiOS pretty popular on feature phones? Or has the buzz on those died down a bit?

It's like the leading OS for non-smart phones.

it's not. S3/symbian is still leading

Really? There's around 170 million KaiOS devices. I was unable to find stats for Symbian. Is it still so much? I thought the platform is dead for years.

tbh most 10 year old Symbian devices probably still work.

Yeah, I have a few in my drawer. The actual question is, does anybody turn them on for anything more than few hours of nostalgia? Last time I tried, it wasn't able to browse the web due to SSL errors. Also, did they ever get to support 4G? Where I live they already turned 3G off, so mobile internet in my Symbian devices doesn't work.

I'd be very surprised if somebody actually uses it as their daily driver. Maybe as a really-dumb phone, only for calling and SMS - but anything smarter than that doesn't work anymore. Even the maps are a decade without any further updates.

I went down the path of looking at "dumb" phones to offset my technology use as a SWE for personal reasons. Almost every phone was Kai.

The problem is that KaiOS is just a "big browser" so its usefulness entirely depends on how much developers care about the mobile web, well they don't.

The mobile web today is horrible to browse, websites are slow, most aren't designed with mobile first in mind, mobile has become an afterthought because companies want users to use their "native apps" where they have full control over ads, tracking and what not.

Go to i.reddit.com and compare that version with the regular mobile website, there is no debate, i.reddit.com is bloody fast and light while whatever the regular mobile website uses is slow as hell. Why is that? Because reddit new design wasn't designed to be efficiently browsed on mobile phone.

Same could be said with any ticket/hotel booking, eshops, or governmental mobile website. 99% of these aren't designed for mobile at first place. Slapping a CSS with some media queries isn't designing for mobile.

KaiOS isn't a failure, the mobile web absolutely is since developers stopped caring about it.

I wonder what all of the PWA booster developers feel about the mobile web.

It's not just mobile web. You need good accessibility--especially for keyboard--for this to work. But a lot of developers don't make their sites work without a pointer device. This is a problem on the web at large though.

For instance, I needed to get a BIOS upgrade for my mouse issue on my Lenovo laptop, but their support site a) required JavaScript and b) wasn't using tabindex or keyboard accessible buttons/links to navigate it. Why this is required to get me a simple ISO or zip is pretty ridiculous. I should be able to use a TUI browser to get such downloads for my device and it doesn't require any interactivity.

Beyond this, PWAs aren't the most popular because of Apple if I'm speculating. The PWA experience on Apple's OSs are so handicapped that it's often seen as a waste of time to develop for as Android users are the only ones that might bother installing the app and desktop discoverability is horrible (Fx stopped support for SSB before it even went to stable). Also there isn't a FxOS or KaiOS system that is actually used in the anglosphere where most developers still come from so the need to support is overlooked.

>For instance, I needed to get a BIOS upgrade for my mouse issue on my Lenovo laptop, but their support site a) required JavaScript and b) wasn't using tabindex or keyboard accessible buttons/links to navigate it. Why this is required to get me a simple ISO or zip is pretty ridiculous. I should be able to use a TUI browser to get such downloads for my device and it doesn't require any interactivity.

Call me a dinosaur as much as you want, but once upon a time those would have been on a FTP site, simple and easily accessible.

Then the web "evolved".

I don't disagree. Luckily I just checked and it seemed the model was added to LVFS this month so I don't have to think about it so much.

New reddit has a few different versions they seem to be A/B testing, some of them are lighter/faster, some restrict things like reading comments or viewing nsfw content to people who are signed in or using the app. I like libreddit for browsing, it's read only but I don't have an account so that's fine.

yeah, many web sites work better in desktop mode on my phone...

Which is weird, because the web is the one place where you can hit any OS mobile or not. I mean, yeah that adds complexity but not developing concurrent versions of apps has to be a win.

From the article: "KaiOS turned out to be disappointingly closed and oriented to tracking users and shipping advertising"

This article is weird. Looks like an important part was lost in editing.

I'm still not sure if Capyloon is open source or privative, TFA is not very clear.

It's open source, and all new code is released under AGPL. See the frontend repo at https://github.com/capyloon/nutria/ for instance.

I would be pretty happy if Mozilla would make the desktop Firefox (the one for Windows/Mac/Linux) responsive. So it becomes usable on Linux on mobile phones :)

No way this OS could have worked, html+js based OS on a resource constrained device is the stupidest idea ever, reminds me of the java based blackberry OS, a shit show

What's the most popular mobile OS? and what they created to power their interface and to gradually replace their old toolkit? yeah, it's not a html+js based OS ;)

If they ever want to revive Firefox OS, they'll need a better frontend story

I never even slightly thought FirefoxOS would be successful. We had MeGoo, a much better mobile OS. That was the real chance of having an open mobile OS.

Man, I'm still sad about MeeGo. Nokia was our best shot but fortune was not in our favor.

It was literally the best phone OS in 2011. Not the best in terms of apps of course.

What really pisses me off is that they literally didn't even sell the N9 in major markets. Like wtf are you doing. This was a high margin phone, ready for production. It should even have been out much earlier.

The follow up phone was getting close to release. They also had the N950 almost ready for release. And in theory they also had a lower prices phone with MeGoo that was supposed to be out, but it was killed because of their internal competition.

Nokia just really stuck around far to long on SynbianOS and that group had way to much power inside of Nokia.

I really wanted the N900 and I was just waiting and ready to buy some N1000 that I was sure would soon come out. But it never came. By the time the N9 came out it was already pretty late and they had already killed the platform.

It was a trojan horse by Microsoft. They sent one of their own to destroy Nokia from the inside and takeover the company. Everything went downhill right after he became the CEO.

https://www.wired.com/2013/09/microsoft-nokia/

The book "Operation Elop" comes to a different conclusion.

Its more a case of Nokia board-members WANTING this situation, that's why the allowed this direction and were more then happy to pawn of that whole business to Microsoft.

Hard to believe that a brand that was in direct competition to Apple and had a strong brand value just as premium as Apple if not more got dissolved so quickly.

You'd better support Ubuntu Touch. At least there's some community around it and it already runs on several phones.

Can it run android apps in an emulator?

Even if it can, Google Play services won't be allowed to function within it.

No, running android apps is not a goal for that project.

I still have my Firefox phones. I cannot bring myself to throw them away and let go of the dream.

> Nokia's Maemo (now Leste).

I thought the successor to Maemo was Sailfish...

Leste IS Maemo, just the free parts, with effort to replicate the closed parts. Maemo's successor was MeeGo, and Sailfish is the successor of MeeGo.

> Firefox OS, also known as Boot2Gecko from the "Gecko" rendering engine of Firefox, was a surprisingly clean and smooth experience, and its official deprecation in 2016 was probably more of an adoption issue than a technical problem. > > Even Firefox OS's user interface was very nicely designed, with visible inspiration from HP WebOS and Palm OS. Clean lines, a lively flat design, and an easy-to-use app development toolkit made it a compelling alternative to Android and iOS. Except nobody adopted it seriously, and it died in 2016.

…did the author ever try to use a FirefoxOS as their daily driver, or did they just play around with it for a bit?

• The SDK had constant breaking changes

• Update policy from hardware vendors was worse than Android's (Alcatel never updated their "flagship" FxOS phone outside the US)

• Together, it meant that most users couldn't use the app store

• The UI/UX was very obviously just half-assedly cobbled together. It was flat only because that's easier to do when you have no designers on staff, and none of it was clean compared to stock Android 4

I never tried writing an app, so I can't comment on the quality of the SDK, but since it broke constantly and most app authors apparently gave up, going by all the incompatible corpses rotting in the app store, I also question the "easy-to-use SDK" part.

The default apps were also kinda awful, I still remember the camera app having an overflow issue where too bright pixels were rendered as black, e.g.

Overall, FxOS failed not because people didn't understand its hidden genius, it failed because it was awful.

Reply on news.ycombinator.com