Google is finally killing off Chrome apps, which nobody really used | Hacker News Reader

Google is finally killing off Chrome apps, which nobody really used

theverge.com

377 pointsdoener259 comments

First of all, this was announced more than 3 years ago, and it still won't happen for another 2.5 years.

Second of all, this should be treated as good news for the HN crowd: Google is transitioning from a proprietary, Chrome-only app model to the open-standard, PWA model. Who doesn't support that?

There are a bunch of "typical Google cancelling a product" comments here. But for real, I simply cannot imagine any Chrome app which people actually use, whose developer is not going to transition to PWA and gain a bigger market.

Seriously, this is like complaining that Microsoft Edge doesn't support ActiveX controls. (And I certainly don't recall any complaints about that here.)

This is good news, people. Less fragmentation and more open standards.

This is very good news unless you bought a Chrome app (say, a game), and you won't be able to use it anymore.

As a person who uses the Hangouts chrome app daily, I have been expecting a day when I go to launch it and it wont open for the last 3 years. Google doesn't want to support "Classic Hangouts" anymore (and supposedly they killed it in October for GSuite, but not for my domain, I guess). Hangouts is one of the only ways I privately communicate with old friends and past co-workers since EVERYONE has at least one gmail account.

And trying to get 20 people to switch to something else is a complete non-starter.

> And trying to get 20 people to switch to something else is a complete non-starter.

Eh. I participated in getting all my friends to start using WhatsApp and later to stop using it.

Way more than 20.

Does hangouts.google.com not work for you? It works for me. You don't need an installable app to use it. And it works inside Gmail too, of course.

(True, for companies Google's currently transitioning to the alternative "Hangouts Chat" which has enterprise features. But that's different from consumer Hangouts, and in any case obviously neither depends on Chrome Apps.)

hangouts.google.com will actually launch the chrome app since I have it installed. Also, I only use hangouts for the chatting and SMS features, so all that cruft for video conferencing and stuff really detracts from the simplicity of the hangouts app.

So uninstall the app and then just use hangouts.google.com? There are buttons for video/audio calls, but if you don't click them then it behaves like a chat app and when you have 2 conversations open they get hidden anyway.

Over a year ago the Hangouts app had a notice at the top saying support was going away and to switch to the extension. That is still working fine.

We've been moving people over to Chat in our domain but a large (probably majority) still use Hangouts.

In addition to the Chrome app, there is the Chrome extension. They basically do the same thing and the extension has no expiration message. The main reason I wasn't using the extension before is that it didn't provide a desktop icon in Linux. There is a way to fix that by creating a .desktop file[1]. I'm don't remember if there is a similar problem and/or fix on Windows/Mac.

For what it's worth, I've found Hangouts to be the better consumer communications tools. I use other software like Signal and Whatsapp but Hangouts works better between devices. It also makes moving to a voice/video call easy for my less-technically-inclined family members. This is especially true when you want to have a call with multiple users, want to switch for your phone to your computer or want to start a screen share. All of these things can be very useful, even with family members.

Of course, there are trade-offs. There isn't E2E encryption. The video takes a lot of battery on many devices. You never know what's going to happen to the service in the future. What you discuss is being recorded and probably used to target ads.

[1] https://askubuntu.com/questions/771408/separate-icons-for-ch...

I still have SMS tied to Hangouts. I moved my phone number to Google Voice when I left the US years ago and kept it on there when I moved back. It's been migrated to hangouts, which is pretty much total garbage now.

I want to migrate my phone number to another service, but I also want to get off Android entirely at some point and want to find a service with a public API so I can eventually write Plasma Apps or a Matrix bridge. Are there any good VoIP providers I can pay for monthly that fit that bill?

Duo is their future replacement and works fine on the desktop.

Duo is about video calling specifically, no? Hangouts is a generic IM client.

I'm on Google Fi, and one major feature of Fi is that it lets you receive texts and calls to your phone number on your desktop, and access voicemail there. But it's all done via Hangouts now, and Google has been very vague on what the replacement is going to be going forward.

They have reverted back to Google Voice for this and I'm pretty sure it'll be the replacement.

When I open voice.google.com, as soon as I login, it immediately redirects me to the Google Fi account page. Which does not have any of this functionality.

Works fine for me. Also if you look at the GV Android app it has gotten a lot of recent updates compared to Hangouts, including Dark Mode. Additionally they have added the option to text from it which has newer features than Hangouts.

I believe Google wants to release Hangouts Chat to general consumers and replace Classic Hangouts.

A chrome app that people use? There’s a lot of hobby flight controllers which are/were programmed using a Chrome app. There was no alternative. And since most web apps can’t get access to the host USB (webUSB seems to exist, but still seems pretty nascent), that whole system is probably going to have to find a new app deployment method.

These should probably be native mobile apps, no?

On a Chromebook ? The Android on ChromeOS stuff isn't as complete and has a hard time seeing e.g. USB peripherals.

I've reached out to the people I work with at Google to see what this will mean for my Chrome Application [0], which is a Certificate Provider [1] that talks to a Google provided ChromeOS Application [2]. Technically I can turn it into a Chrome extension now that there's built-in PIN prompting support, but the built-in SSH Agent will suffer as a result, since it will have no way to prompt for a PIN.

[0] https://chrome.google.com/webstore/detail/cackey/lpimdiknnpi... [1] https://developer.chrome.com/extensions/certificateProvider [2] https://chrome.google.com/webstore/detail/smart-card-connect...

Not on a Chromebook, as it appears it is no longer a usable dev or runtime environment for such applications.

Flight controllers in question don't feature wireless connectivity - they are typically used in FPV racer drones, and only offer a USB connector for firmware flashing and configuration. There are Android apps that use USB OTG for the job, but the primary option is the desktop app built by the same authors/community as the flight firmware, which originated as a Chrome app and was eventually migrated to Electron.

Hasn't Betaflight (and iNav, and other relatives/clones) migrated to standalone Electron by now?

Web USB is going to more widely supported than Chrome's proprietary USB API. A Chrome app doesn't have to be written, its use of proprietray APIs just needs to be replace with the equivalent standard ones.

Complaining about Google is a sport these days.

Seriously... let it go people. Google is a business.

As often, the top HN comment is an answer to comments that very few people made. Almost nobody complains here, and even the article states that Chrome Apps weren't widely used.

The top comment when the GP comment was posted was the thread starting with "Typical Google move ... now I suppose my Chromebook will be soon a useless brick."

(Disclosure: I work at Google, speaking only for myself)

Any comment can be the top comment because the time decay factor dominates until there is a comment that amassed a certain amount of upvotes.

It's 5 hours old now, so the time decay factor is insignificant, and it's still the second highest comment.

So it's not accurate to say that no one is complaining about this, or that the response was not to opinions that are actually held.

This is not accurate in this case and feels like a response that is dismissive of but does not actually address the points made by the person you're responding to.

Anyone know the status of being able to install PWAs on the desktop?

The reason I use chrome apps is to have "Gmail" as an application on my desktop so I can alt-tab to it, put it on my dock, etc and treat it as a normal app instead of a tab.

The new MS Edge supports this.

"add to desktop" option has been available for a couple of years now I think.

In Chrome, open the page you want, then on the Chrome menu select More Tools / Create Shortcut...

Give it a name and check "Open as window". Now you have a shortcut you can use like any other application shortcut.

(I've only tried this on Windows and Linux, but it should work on macOS too.)

If you are a Windows or Mac user, third-party apps like Kiwi for Gmail support this functionality (plus support multiple accounts.)

Just make a simple bash script that goes:

open -n /Applications/Google\ Chrome.app --args --app="https://gmail.com"

Add shortcuts to that.

I was using those on an old Chromebook that never received support for Google Play / Android.

VLC was available as a Chrome app, so that filled a need I had.

Off the top of my head, these were the Chrome apps I used:

* VideoStream (allowed casting to Chromecast from filesystem) * PostMan * Internal tools at previous job that used MDNS

None of these applications work as a PWA. All of these moved to Electron instead. Which is heavier and just as non-standard and dependent on Chromium as Chrome Apps.

Postman has become so bloated in the past two years to a degree where I don’t enjoy using it anymore. Anyone know of an alternative that keeps it simple?

I use this thing: https://insomnia.rest/

I've been using Insomnia instead, but it's also Electron I believe.

I agree that this was a terrible idea from day one and am glad it has been killed and will soon be buried.

But this: "typical Google cancelling a product" is still an important message.

Google routinely floats out APIs, services, and platforms and encourages their use despite fatal flaws that mean they are doomed from day one. And some developers fall for it and later have to pay the price. Hopefully, the repeated warning will minimize this. Think of the devs right now that are mistaking Flutter for a platform that will last.

> But this: "typical Google cancelling a product" is still an important message.

But it's one that appears with such regularity in HN comments that it drowns out all other discussion.

I've frequently seen otherwise interesting posts related to Google derailed into yet another pile on - or even worse full of people that think they are being witty by posting "how soon before it's cancelled?"

You’re interpereting it as someone being witty, but consider that people have legitimate concerns and are supplying mountains of concrete evidence for their position.

I wasn't disputing the validity of their complaint. That's orthogonal to being unfunny through repetition.

Claiming that them saying it is them thinking they are being witty is disputing the validity of their complaint.

For example:

I could constantly remark on the current bad weather using an unfunny joke.

Somebody could complain about my terrible sense of humour without implying it was actually lovely outside.

You're still claiming that people are saying it just to be witty/funny. They aren't. It is a legitimate concern that people should be informed of, if they aren't aware.

> You're still claiming that people are saying it just to be witty/funny.

You see that word "just" in there? That's the bit I disagree with.

It is possible to both agree with someone and dislike how they express themselves.

Hacker News is not a person, and you will only drive yourself insane if you treat a conversation with a collective like it is a person.

> people that think they are being witty by posting "how soon before it's cancelled?"

I legitimately don't think most people are trying to be funny when they say things like this. I think they feel (as I feel) that it's not worthing becoming reliant on Google's new fancy products, because there is a high likelihood they stop working in a few short years.

> But it's one that appears with such regularity in HN comments that it drowns out all other discussion.

See, for example, the second ranked subthread in this very story, from a user who apparently thinks this change is going to brick all Chromebooks; and the flood of responses talking about things like refund policy drowning out the tiny handful of comments pointing out that this the premise was a complete misunderstanding of what is happening.

HN can't rationally talk about Google any more. It's kind of depressing.

> HN can't rationally talk about Google any more. It's kind of depressing.

Unfortunately, it's not the only company where threads turn into this. Every Apple product thread is about removing the headphone jack (still), every Amazon thread is about counterfeit products someone bought. There are others.

This is probably going to happen and soon... And then next they'll kill off extensions, and then Chrome will become the next "Internet Explorer"...

One of the things that people will take for granted is the effectiveness of extensions that work as Ad blockers which brought online tracking to a grinding halt. Google because of their advertising services has always wanted to suppress the ability of people in blocking ads. Most of the changes they make to Chrome, now that they have captured majority of the market share are geared towards generating profit, and preserving their status a dominating company.

What's probably going to happen and soon? The comment you're replying to does not suggest some event will happen.

I don't think anyone is saying that web based apps were a bad idea from day one... Wouldn't the success of PWAs and pushing the web forward to support this type of open computing a good thing? Essentially, Google took a bold move, the tech world liked the idea of a more app-centric look a the web and pushed for PWAs. By all accounts, this is a success for all the web. It's okay for a product to go away if better stuff comes of it.

> Google routinely floats out APIs, services, and platforms and encourages their use despite fatal flaws that mean they are doomed from day one.

For most use cases, there are better options now than Chrome apps. ServiceWorkers didn't exist when Chrome apps were introduced. Neither did wasm, and its precursors NaCl and PNaCl were scoped to just Chrome apps, not to the public web.

Chrome did a very good job at a) building a platform that developers could use right then and b) working with other browsers and standards committees to get enough features implemented to replace what they built. They said, no, we shouldn't have a proprietary thing called "Chrome apps," we should just build it into the web platform so that people who don't use Chrome can use them too.

And they set a very long (many years) timeframe for shutting down their initial version of the thing.

This is exactly what I want a platform provider to do. I don't want "Chrome apps" any more than I want Flash or ActiveX or "Best viewed in Internet Explorer." But also, I do want platform providers to build platforms - I don't want them to refuse to build platforms on the grounds that they don't know the right answer yet and they can't immediately standardize something right now. I'm glad Flash and ActiveX are dead, but I'm glad they existed because they proved that the platform is valuable, if only you could implement it in a reasonable way. I'm glad ActiveX's original version of XMLHttpRequest existed because it proved that there was demand to standardize the version we now have.

Is there any way to bypass CORS in a PWA? Because Chrome Apps were a really handy way of writing GUI utilities without having to go full Electron.

Unsure, but I think if you have an extension, you can make the request in an extension page, which isn't CORS-restricted.

(I don't totally understand the difference between a Chrome app and an extension that turns itself into a PWA, both are browser-specific and both are distributed through the Chrome web store, right?)

A chrome app has access to normal sockets, local filesystem, etc.

Not sure how they'll handle this... given the risk of losing a local db store's data in chrome, I'm hoping they normalize on something.

Ideally, a directory for access/storage enough to use SQLite as via WebAssembly should be available to plugins with a request/approval for read (and optionally write) access ... to me, that should be a goal of any file system interface presented via WebAssembly, is the option for something that at least supports that use case.

There's a standardized File API https://developer.mozilla.org/en-US/docs/Web/API/File and a nonstandard FileSystem API being driven by Google https://developer.mozilla.org/en-US/docs/Web/API/FileSystem

But yeah it seems network sockets are going away. Honestly that seems good from a security perspective but I will miss the mosh app for CrOS.

No, you can't. PWAs are just "glorified" web pages. Nothing more or less. Electron is the only viable choice for desktop apps with web technologies.

If the endpoint you are requesting says it does not want to expose itself to unknown origins, that should be respected in the default config.

For testing, security research and so on you can always modify the browser config or use a proxy.

Flutter is an open-source framework that targets Android, iOS, Web & Desktop. That's rather different to something that just targets Chrome. The developer experience is amazing. Personally I'm using it and not too worried about Flutter lasting. Just like I use react on the web and I'm not worried about that either.

That something is an important message is not a sufficient reason for it to appear on the comment page. Being important and tangentially related to the topic is also not sufficient. Otherwise, all comment pages degenerate into discussion of the closest available hot-button issue.

Wait what? Edge doesn't support ActiveX controls? Why didn't I notice until now?

I smell another round of consulting fees!

The problem is the Google Brand approach of replacing a used consumer product with a backwards incompatible replacement and forcing the expense/risk of a port onto the end-user as if they have the same resources as Google does.

There will be a lot of cases where this approach is reasonable. But because Google has built a strong reputation of constantly doing it, each event will be a new outrage causing more and more churn. You reap what you sow and all that.

Postman used to be available as a Chrome app but when this was announced they released it as an Electron app instead.

The end of support for Chrome Apps on Windows, Mac, and Linux is June of this year, unless you have paid Enterprise/Education support.

people supporting enterprise. enterprise software is bad to begin with and often has long shelf lives. ActiveX, old java, flash et al are all still flourishing at a large corp near you

So sadly true. Why does my company's spam filter require Flash in order to view a list of messages that were flagged? That's insane.

training sites(internal/external) will be using flash/java for a long long time

>> This is good news, people. Less fragmentation and more open standards. reply

Actually is bad news. PWAs are pretty useless and when I say useless I mean no better than a website. All the web restrictions(i.e cors) apply to PWAs. It's just a "webpage shortcut". The alternative to Chrome apps are really Electron apps not PWAs.

besides Chrome Remote Desktop i cannot think about any other useful Chrome app.

There were SSH and RDP (windows RDP protocol) apps when I used a chromebook that were pretty decent (getting the private key in the SSH client on the chromebook was, well, interesting).

Not sure of others.

Chrome Apps were a Chrome-only proprietary solution, and they're being replaced by Progressive Web Apps which work in other browsers as well. This is the sort of change we should be supporting!

(Disclosure: I work at Google, not on Chrome)

No I don't support pwas either.

Why?

Do you support X or support Y is not really what was being talked about, the question is from a starting point of X existing do you support it moving to Y (instead of remaining X).

Moving from a proprietary to a shared/common system is an obvious move forward.

So PWAs will have support for Bluetooth and all the other features of Chrome apps?

That's the broad idea of Google's Web Capabilities effort, “There are some apps that are not possible to build and deliver on the open web today. We call this, the app gap. The gap between what’s possible on the web and what’s possible on native. We want to close that gap. We believe web apps should be able to do anything native apps can.” https://developers.google.com/web/updates/capabilities

> We believe web apps should be able to do anything native apps can

It's a nice sentiment. Time will tell how that turns out.

https://developers.chrome.com/apps/migration

Everything except advanced File System APIs are supported by PWAs, including Bluetooth through the Web Bluetooth API. [1]

The replacement Native File System API is in origin trial and is pre-standardization, however. [2]

[1] https://developers.google.com/web/updates/2015/07/interact-w...

[2] https://web.dev/native-file-system/

Another big gap is the sockets API. Chrome apps have direct access to TCP and UDP sockets. PWAs can't have this for obvious reasons.

Good catch, and the migration page kinda glosses over that.

They are asking for input and use cases, though.

"Q: My app uses the chrome sockets API to do networking. Can this be done with the open web?

You might be able to do what you need with WebSockets. However, to use this, you will likely need to change the remote end of your connection.

If that isn't possible, we'd like to hear more about your use case—please let us know about what you're trying to accomplish via the Chromium issue tracker.[1]"

1. https://bit.ly/new-fugu-request

This is bad. Simple TCP and UDP sockets are the foundation of the internet.

Correct, but they are not the foundation of the Web -- HTTP is, and web specification writers are only concerned with creating APIs for the web platform (hence why Web Sockets is an upgraded HTTP connection, not lower level like raw TCP or UDP.)

What are the obvious reasons? Many game devs would absolutely love to have UDP available on the web in order to do proper networking.

Extremely good information and background context is available in this article: https://gafferongames.com/post/why_cant_i_send_udp_packets_f...

The big issue is trying to make UDP "safe" enough security-wise for the web platform, and such considerations are generally taken extremely seriously. If a vulnerability were discovered, the Google Chrome userbase could become a botnet capable of DDOS pummeling any website.

Workarounds apparently exist for UDP over WebRTC, but it's a very "bulky" solution.

Does this mean no longer SSH terminals for Chromebooks?

The last Chrome OS app I worked on for education used UDP to send broadcast packets in the classroom and find other devices to be able to communicate via WebSockets.

I don't work there anymore but if one of the schools using that app bought Chromebooks without support for Android apps it's going to be a shit show in 2021.

I feel like chrome apps were actually ahead of their time and very much better than embedding a full browser per app with electron.

Of course, they were not perfect but IHMO made much more sense than what we have today, at least on a technical standpoint and could have evolved into something truly great.

> I feel like chrome apps were actually ahead of their time and very much better than embedding a full browser per app with electron.

> Of course, they were not perfect but IHMO made much more sense than what we have today, at least on a technical standpoint and could have evolved into something truly great.

Killing Chrome Apps isn't about abandoning what they enabled, it's about moving it from Chrome App-specific capabilities to web platform capabilities.

https://developers.chrome.com/apps/migration

https://developers.google.com/web/updates/capabilities

I.i.r.c. Chrome Apps are compiled binaries specific for Chrome (NaCL), they're more like Flash, Java or Webassembly applications.

no not at all.

There are javascript+html+css apps that could use the Chrome API ( https://developer.chrome.com/apps/api_index ) which allow them to do more than regular javascript+html+css web pages.

There is no compilation.

Not NACL. JavaScript/HTML/CSS. But with access to local filesystems, sockets, etc.

Because of that, Google's suggested replacement of Android Apps (won't work everywhere Chrome does) or PWAs/Extensions (both heavily sandboxed) doesn't work.

Chrome Apps are more like sandoxed Electron apps which do not require the user to install the same runtime every time for each app installed, unlike Electron. No need for NaCL, they can be written in Javascript.

> I.i.r.c. Chrome Apps are compiled binaries specific for Chrome (NaCL),

NaCl was supported in Chrome Apps, but most Chrome Apps are just web apps using the chrome.* APIs, not NaCl.

Eventually we will get libchromium{.so,.dll} installed by default in all operating systems so that Electron apps can share the same runtime as your browser.

When Google first announced all the way back in 2016 that it would end support for Chrome apps on Windows, macOS, and Linux, it said approximately one percent of users on those platforms were actively using packaged Chrome apps.

The move was announced in 2016. Apps are used by 1% of users.

That's still a shit load of users.

Chrome claims it has 1 billion users, thats 10,000,000 users of chrome apps, 1/6 of the french population.

Of course I myself never used them, I learned not to trust any Google product to begin with. These are the kind of decisions that will damage Google on the long run. I've heard that Stadia is not doing so well...

Chrome Apps were Google's idea of packaging websites as self-contained apps. They only worked within the Chrome web browser.

Now, we have Progressive Web Apps, and Google have decided to promote using open standards to achieve the same goal, rather than offering a Chrome-only solution.

PWAs can be used to "install" applications on your desktop or Chromebooks that are almost identical in capability to Chrome Packaged Apps. In my view, this is a highly responsible move by Google in moving the web forwards via standards.

Google have even released a migration guide (in 2016, when this was first announced): https://developers.chrome.com/apps/migration

Typical Google move. I have a Chromebook. It was listed on the models that would support Android app, this support never came (it's still "planned", 2 years later...). Support for Linux apps didn't come, either.

So now I suppose my Chromebook will be soon a useless brick. Thanks, Google. Got the message, I'll never use any product with your badge again.

You say that, but you got a Chromebook despite everything Google did in the past. If people had gotten any message they'd have abandoned Google back in 2008.

First they came for the Grand Central users, and I did not speak out—

Because I was not a Grand Central user.

Speaking just for one, I haven't adopted a new Google consumer product in 7 years. I don't even click the links.

Pretty sure Reader had that effect on a large chunk of the technical base they previously relied heavily on.

Indeed... I don't even consider using their development tools now.

Just the other day I met a guy who told me he wouldn't consider learning and using Go because Google will cancel it eventually

What a reputation issue they're developing. Even if their tech is pretty good, there's a growing number of people who won't even consider it (myself included)

That doesn't really make sense, though. Go is entirely open source. If Google dropped it tomorrow there are more than enough non-Google people invested in it to continue development.

Maybe, but how many of them are compiler engineers? Go does not strike me as a language, that a compiler engineer would love.

Eh, if google cancelled go, there's a good chance a bunch of go compiler engineers would move to other companies using go...

It would continue development, but how popular would it remain without Google blessing?

Very reasonable concern. I like Go, but it's not sufficiently better than its competitors (Java/C#) to be professionally maintained or adopted. Then there's the huge chunk of people who use it just because of the Google halo effect. My guess, Go and its ecosystem would end up floundering like other ex-corporate languages like Object Pascal or Objective-C. Even more so if Google abandons Go because they made a brand new replacement.

Shutting down Reader really did an oversized amount of damage to Google's reputation. I always found it surprising that Google wouldn't devote the engineering resources just to maintain Reader and not upset the influential people who relied upon it.

No one is more influential than Google.

- Google

Was Google anywhere near as bad in 2008? I don't recall anything particularly awful back then.

No, I think that was around the launch of the first android, right? There was some serious interest in a Linux (!!!) based phone. They were seen as a savior of sorts, as I remember it. Now they’re a place where novel ideas go to die.

That's how I remember it - they were starting down their current path in privacy terms but hadn't hit the moral event horizon yet, and they were still in that unbridled optimism phase where they were launching moonshot projects rather than churning through web service offerings. I didn't get seriously concerned about them until later (around 2014-2015 iirc) and it took until the start of 2018 for me to completely de-google myself.

I don't remember all the little details, but it was the year I started to actively avoid anything they touch, starting with their search engine. I think the main concern was the privacy nightmare.

Chromebok was originally marketed as a simple, no-nonsense laptop and that's what it is. All iterations since have been useful but did anyone really believe it was going to be a ThinkPad killer?

Once they started selling $2k+ versions, yeh I kind of do think expect that actually.

Fair point.

Precisely. Though all the Google devs I've met weren't using them for development --> devs aren't the target market.

TBH I dont understand the developers preference for laptops, generally.

The laptops are good when you need to be mobile - but you can pry my multi-large-monitors desktop out of my dead hands, because I'm not giving that up for serious be-in-the-zone development.

I can only speak for myself, but my development environment being mobile is precisely the point. I can work from home, I can commit changes while on business trips... it's valuable. I also have a large monitor at the office that I plug my laptop into, so I get the best of both worlds.

Uh, my work device is a laptop and I use multi-large-monitors. With external keyboard and mouse.

I can unplug it (single cable) and take to a conference room if the need arises. I can take off early and plug it at home. It has a built-in UPS by the virtue of having a battery. Why would a desktop be better, unless you need crazy hardware?

I do have a desktop at home, but that's for my personal uses.

Yeah, I have no idea who those are for, but the cheaper ones are great for family members who only really need to browse the Web anyway. My tech support burden is way down, and so I love Chromebooks for other people.

Those high-end are amazing business machines for people who work with cloud/web software.

They're very easy to manage from a central location, easy to replace and very hard to break.

I've been recommending cheap Linux laptops like the Pinebook (as well as the Pro) for family members who just use their computers for email and Facebook. Turns out if they spend 99% of their time using Firefox, Linux is a good platform that can suit their needs just fine. Haven't had a tech support call in ages.

The ThinkPad Chromebook 13 is actually pretty nice: https://www.lenovo.com/us/en/laptops/lenovo/student-chromebo...

Of course, Lenovo doesn't make those anymore, but you can find them used.

There used to be cheap linux notebooks, that what I used to buy, then Chromebooks appeared and replaced entirely cheap linux laptops. I stupidly hoped that it would be good enough, or easily convertible to straight Linux; it isn't. In fact the main problem is that Google Docs doesn't properly support OpenDocument format, making it crippled for me even as a basic desktop (it's only useful as a browser and nothing else).

Many of the Chromebooks are pretty simple to convert to a standard OS. I've got two chrome devices flashed with 3rd party firmware and running FreeBSD. It seems more likely if you get devices running Intel Core series chips (cpu model numbers like xxxxU or maybe xxxxY, but not Nxxxx), but it's good to confirm it's supported by your 3rd party firmware of choice before you buy.

I use firmware from https://mrchromebox.tech/ these days.

+1, I'm on the Chromebook Pixel 2 (2015) and have been running Arch Linux for years now. I can't recommend it enough.

Unfortunately, since 2017, Google has been using a new security chip (CR50) that isnt supported in the upstream Linux kernel. So while you can flash 3rd party firmware and install your own OS, the security chip will throw the device into recovery mode if you ever suspend the computer.

AFAIK, there is no workaround for this - if anyone knows otherwise, please share details. I've got a chromebook that I'd very much like to be able to use something besides ChromeOS on.

In fairness, they've enabled Android apps on an awful lot of Chromebooks: https://www.chromium.org/chromium-os/chrome-os-systems-suppo...

All new Chromebooks support Android apps afaik.

What model do you have?

Chrome OS is Linux (it runs the Linux kernel) – you mean your Chromebook doesn't yet support the greater Crostini project? https://chromium.googlesource.com/chromiumos/docs/+/master/c...

What part of the Quickstart: https://chromium.googlesource.com/chromiumos/docs/+/master/c... doesn't work for you? It says “Make sure you're running at least R72 (released Feb 2019).”

Even though there's a lot of abandonware from Google, Android apps and Crostini on Chromebooks are not two of them. In fact, Android apps is one of the migration paths away from Chrome apps, a fact this article fails to mention.

>Android apps and Crostini on Chromebooks are not two of them. In fact, Android apps is one of the migration paths away from Chrome apps, a fact this article fails to mention.

So should someone buy a new Chromebook on the hopes that this new system of apps will continue to be supported down the road? I think once you're burned once, that's hard to buy into.

It's an ASUS C300M, I bought it around 2015. There is no "Linux (beta)" entry in ChromeOS settings, though it's running version R76, the C300 is listed in the "compatible" list, and it's powerful enough to run Linux applications (dual core x86, 4GB of RAM).

I'll have another look but I don't see why it should be any different today than it was last month.

What model of Chromebook do you use? I've 2 chromebooks purchased around $120 each of which is at least 3 years old and I have had android apps since the day I first bought in 2017 and linux support (crostini) for at least a year.

It's an ASUS 300M. A pretty high end model when I bought it 4 or 5 years ago, and as a basic PC, running a standard Linux it would be fine for at least 5 more years. But running Chrome OS, it may become a doorstop any time.

Looks like your model should be compatible with GalliumOS: https://wiki.galliumos.org/Hardware_Compatibility

So you could go that route, if you like. I ended up doing something similar with my original Chromebook Pixel when Google stopped supporting it: I removed the write-protect screw from the motherboard, flashed a normal BIOS, and installed Linux Mint on it.

This [1] is a useful site for replacing the firmware on your Chromebook.

[1] https://mrchromebox.tech/

Never got my keyboard for my google Nexus tablet either, 8 years and counting, actually, reminds me, under UK consumer law I think you can get a refund or part refund if extra functionality was promised and never delivered, will have to look that up :)

you need to recheck as some recent chromebooks are going to get fuchsia which will support the legacy android apps

why do you suppose your Chromebook will be a useless brick ?

which Chrome App(s) do you use that will cease to work in 2021 ?

I have a Chromebook Pixel (2013) and they stopped issuing updates a couple of years ago (os or security).

That’s essentially bricking for me because I don’t want to log into anything secure on a potentially vulnerable device.

So what have you done with the laptop? I love the build quality and lightness of the pixelbooks but spending so much for something to brick in 5 years is an issue.

Parent was making a tangential comment about their Chromebook, not related to chrome apps.

For what it is worth, my current chrome book that was about £150 12 months ago has both linux and app support.

From what I understand about it, it is a linux kernel + driver issue. I do not know if this means it is the manufacturer of the chromebook's "fault", or google's "fault" for not supporting these things. E.g. does the manufacturer need to do a new build for their hardware? I don't know.

Why is the manufacturer even involved in this? I thought all updates come straight from google? I have zero sympathy for Google at this point. It is one thing that the drivers are not all free and open source but Google should at least require that Google has access to the drivers or software it needs to keep devices up to date. If a vendor doesn’t do that, don’t do business with that vendor.

I guess it depends if they've done something funny when producing it? I've had the odd PC motherboard and laptop that won't work right with drivers from the manufacturer (intel/amd/nvidia), they need the specific gigabyte or asus provided driver package. Usually a sound problem. This might explain the wide range of bluetooth complaints.

My dell xps 13" touchpad drivers didn't work from synaptics, it had to be from dell. Something else needed realtek driver direct from their website, windows updates version were broken.

My understanding is Google certifies each chrome book. They should be able to say give me all the source code and the rights to modify them as I see fit or you won’t get certified.

Well I think it is more "this feature requires kernel 4.x, but the drivers for this manufacturer's hardware do not support 4.x, so we cannot offer this feature without losing sound/WiFi/touchpad/etc"

The chromebooks do update themselves all the time but I gues there is some more fundamental stuff that requires more low-level changes.

What I’m saying is Google should require access to the source code and the right to modify it. At that point, it is Google’s responsibility to make any changes to keep it compatible with the latest kernel.

Ok sure - you're welcome to your opinion. I feel like expecting endless maintenance builds from Google for any manufacturer's hardware forever is a bit of an optimistic expectation though, especially for old models. They guarantee at least 5 years support (1) for updates which I don't think is unreasonable considering that they are so cheap. If that is not long enough for you then I guess you are not the target audience (students and enterprise users are the target audience)

1 - https://support.google.com/chrome/a/answer/6220366?hl=en

Well,

Microsoft just ended support for Windows 7 - released July 2009 - yesterday. It’s currently running on my old 2006 era Mac Mini that my mom is using.

I also have a 2009 Core 2 Duo laptop that until early last year was my Plex Server. It is running a still supported Windows 10.

> I guess you are not the target audience

Google always kills their products unexpected for all, don't matter which audience you are.

My Asus laptop bought with Windows 7 on it back in 2009, is running the latest Windows 10 version.

Both Microsoft and Asus keep having updates for it.

Google just needs to actually care.

The driver model for Windows is different than Linux, though. Realistically, most Linux drivers need to be open source and maintained as part of the kernel to continue working, whereas Windows offers a stable driver interface so its not uncommon at all to have drivers that haven't been updated in many years work just fine on current Windows releases.

Google as ChromeOS/Android owner can do whatever they feel like it, regardless of the driver model.

In fact, Android since project Treble has a driver model similar to Windows.

Linux drivers are considered legacy on Android.

Likewise they could have contracts in place that would force their suppliers to provide updates for X years.

> What I’m saying is Google should require access to the source code and the right to modify it.

I guess there would be 0 hardware manufacturers left to choose from then.

> I guess there would be 0 hardware manufacturers left to choose from then.

Maybe I overestimated how much leverage Google has then :/

For Android, they're proposing GKI module shims that a vendor can code against. So a vendor plugs a binary blob into, say a 5.4 GKI module, which is hopefully forward compatible with the interface to a 6.25 GKI module released sometime in 2026.

https://arstechnica.com/gadgets/2019/11/google-outlines-plan...

(GKI = generic kernel image)

They are, but it is only post Android 10, so it will take years before it matters at all.

As an example, Vulkan was added in Android 7 as optional API, the adoption was so low, with so many buggy drivers that as of Android 10 it became a compulsory API.

On Android everything that Google leaves the OEMs free reign hardly gets done.

> I have a Chromebook. It was listed on the models that would support Android app, this support never came (it's still "planned", 2 years later...)

You should be eligible for a refund, file a complaint with consumer protection. In most places it is illegal to mislead customers and market products with planned features that you do not deliver on in a reasonable timeframe.

I’m genuinely curious to hear how this would go.

If this was indeed a fact advertised on the marketing material or worse on the packaging/product page, then he can have a refund or at least a partial one, depending on the country. If it was talked about what could happen, or not mentionned in any official marketing material but in PR pieces, then nope.

See eg the PS3 claiming it could install a normal linux.

I don't know how the exact laws work in other countries, but in France it's also two different things that work differently, one is you sold me a product that isn't working, the other is your marketing lied. Both are illegal but on different terms. Thus if it's not on the box or product page, then it's a marketing material issue.

Wasn't the PS3 refund only like $10-$20? Why even go through that effort for a Chromebook?

It's not about the $10-$20, it's about sending the message.

Costing a many billion dollar multinational corporation $10-20 isn't sending a message at all. It's just a waste of your time.

Not if you expand the $10 into a class action. That however is a lot of work.

It still doesn't add up to a lot of money. The total class of people who were trying to run Linux on the PS3 and then weren't able to just isn't that big.

Once the class action is in motion, what's to stop many more PS3 owners to claim that they were interested in running Linux just to get the free ten bucks?

That winds up being a big payday for a law firm, and free identity monitoring for everybody else...

“file a complaint with consumer protection”

Can you clarify what you mean with this? What consumer protection?

It would be highly jurisdiction-dependent.

In the United States:

For interstate commerce: FTC [1] or state/county AGs/consumer affairs bureaus, e.g. NYS [2], CA [3], WA[4]

[1] https://www.ftc.gov/about-ftc/bureaus-offices/bureau-consume...

[2] https://www.usa.gov/state-consumer/new-york

[3] https://www.usa.gov/state-consumer/california

[4] https://www.usa.gov/state-consumer/washington

There was a guy releasing SeaBIOS ROMs so you can install regular Linux if your device is supported.

The project is dead now, but since your device is older, there's a fair chance it's supported.

I used it for a Chromebox and the process was pretty straightforward.

https://johnlewis.ie/custom-chromebook-firmware/rom-download...

I have a Chromebook C710 and C720 (about 5-6 years old) and I flashed the firmware on both. The C710 now runs Windows 10 and the C720 runs Fedora 31.

I'm running Fedora 31 on my C710 too as my daily driver. It's a surprisingly good experience! You've definitely got to manage your CPU expectations.

Although he's done a lot of great work those aren't maintained now and they're a bit of kludge in my experience. Mr. Chromebox maintains a coreboot firmware[1] that is really good, and the maintainer is active on r/chromeos.

Chromebook hardware is pretty damn good for running linux, I've had zero problems with this firmware and xubuntu. Everything just works now for the most part.

[1] https://mrchromebox.tech/

Everything will still work right?

Which model? You bought a product for a feature that you do didn't exist?

Not sure what model you have but the Chromebook I own supports Android apps.

If you have the time, install a real OS on it.

https://wiki.debian.org/InstallingDebianOn/Samsung/ARMChrome...

"Chromebook" should be renamed to "Chromebrick"

How in the world would it be a brick if you move some proprietary Chrome apps to PWAs?

How difficult is it to make that move?

If it's not trivial I wonder if there are many apps that will never make the move.

And on top of that you only get about 4-5 years of security updates from Google on a Chromebook before they give up on you. [1] That's should be an embarrassingly short amount of time even for a computer product, but alas.... Google doesn't care and most Chromebook users are none the wiser. I wonder how many would choose differently if they knew that Microsoft offers a bare minimum of 10 years of support on their products. How sad is it that "Embrace, extend, extinguish" provides support for their products for twice as long as Google does?

[1] https://support.google.com/chrome/a/answer/6220366?hl=en

> So now I suppose my Chromebook will be soon a useless brick.

ChromeOS devices don't, in general, rely on Chrome applications any more than windows laptops do. The vast majority of the "apps" you might see on your home bar are, in fact, just web links. Is there a specific app you're worried about?

Some of your branding confusion is, in fact, exactly what Google is trying to treat here. "Chrome Apps" were a particular API for persistent web applications. It doesn't mean "apps on Chrome" and it doesn't mean standards-compliant progressive web apps.

I bought a Google branded LG phone about 8 years ago. It was complete garbage and they EOL'd it after 6 months and never delivered anything that they promised. I've never given google hardware a second look after that.

Ah, here's a review: https://www.engadget.com/2011/04/20/t-mobile-g2x-review

Note that Chrome Apps != Progressive Web Apps, which you can still install on desktop Chrome and the new Edge and hopefully will gain more traction.

My understanding is this is the direction they are heading and I don't see this as a bad move on their part for this reason. If they were getting rid of PWAs then I would be upset.

There are some chrome apps that you can't do as a PWA or Chrome extension. For example, there's a local webserver that serves up a folder over http.

https://chrome.google.com/webstore/detail/web-server-for-chr...

A local web server is something that can easily be setup on desktop and mobile outside of Chrome.

How would you do that on a chromebook or without the user installing stuff?

Install an Android app or use Linux in ChromeOS.

It's nice to be able to do it anywhere chrome exists in a consistent, quick, point/click, non-admin user way.

Didn't Microsoft block connections to servers running on localhost through http?

It can also listen on the actual network interface.

I just wish apple would finally support push notifications for PWAs: https://caniuse.com/#search=push%20api

A lot of use-cases like e.g. chat apps, depend on that kind of technology.

This will happen as late as possible as it’d provide an alternative to a large chunk of native apps (so no way to charge the Apple tax).

The Apple tax is what exactly, in this case?

The 30% cut when selling on the Apple App Store.

Don't they take a 30% cut from in-app purchases? I would assume many would try to get around that with a progress web app.

Yes, for instance YT or Spotify subscriptions would cost you 30% more when purchases via IAP.

Spotify is NOT ok with that (I think they disabled IAP). However, Google does allow for that.

Protip: Tapping `<button/>` elements is 23% cheaper than tapping `UIButton`s, which is something I realised looking at my YT bill one day.

Supporting web push would make Apple responsible for sending all kinds of malware and spam down their pipe. Anyone can access the push service, authenticated by Safari. I think it would look bad for their brand to have notification spam under the "Safari" logo.

Please excuse this naive question, but what is the point of ChromeOS/ChromeBook then?

If everything is now WebApp what's the point of an OS that can't run native App vs any other OS that can do both?

I must be missing something???

lightweight/less bloat, lower resource requirements, generally reasonably cheap hardware

I'm sure you mean this unironically, but chrome is by far the heaviest program that the majority of people run.

Well, with a webapp solution the end user is at least likely to only be running chrome once, instead of once per electron app.

Chrome Apps were Google's idea of packaging websites as self-contained apps. They only worked within the Chrome web browser.

Now, we have Progressive Web Apps, and Google have decided to promote using open standards to achieve the same goal, rather than offering a Chrome-only solution.

Google have even released a migration guide (in 2016, when this was first announced): https://developers.chrome.com/apps/migration

Nearly every Chrome app is going to have an Android app counterpart I imagine. Chrome OS and Android are two competing platforms. Android took off and developers made zillions of Android apps, Chrome apps not so much. However Google has never shown any sign of abandoning Chromebooks. If anything, they're getting nicer and nicer and all new models run Android apps thus solving the app shortage problem on Chrome OS.

Check out this review of Samsung's latest Chromebook at CES 2020 by Chrome Unboxed: https://www.youtube.com/watch?v=mD_m6-nLvfs (btw, I don't own a Chromebook, but I am a Linux fan, and so I'm glad Google have put their weight behind Linux with Android and Chrome OS)

Security and simplicity. Such a low-maintenance system. When I sign in to any Chromebook, my wallpaper, shortcuts, and settings are just as I left them. For example, I may never have to find the setting to maps caps lock to escape ever again if I stick to ChromeOS.

To be fair, Windows 10 does this.

Somehow I have troubles finding lightweight win10 laptop with comparable price.

Feature-wise, it doesn't have to do something more than the others, it just has to be able to do everything the user needs.

And for the purchase decision, there might be other factors involved as well: price, security, design, hardware, etc.

> If everything is now WebApp what's the point of an OS that can't run native App vs any other OS that can do both?

The point is Google's effort to acheive capability equivalence between web and native apps, so that there is no consumer-relevant difference to “cant run local apps, only web apps”. Setting a firm.date for Chrome Apps going away can be seen as equivalent to Cortes burning his ships: they are visibly abandoning their alternative to pulling off that effort.

I developed a few Chrome apps while I was in uni and what a pleasure. The APIs were easy to comprehend, the creation of an app running on any desktop OS only took a few minutes, I loved it.

The lack of documentation from Google was already noticeable 4 years ago thus I'm not surprised about that move.

That's too bad. Chrome Apps allows TCP/UDP communication and it could create interesting apps.

I'm still using JSTorrent (https://chrome.google.com/webstore/detail/jstorrent/anhdpjpo...).

It's impossible to implement the BitTorrent protocol using a webapp, which means that once Chrome Apps are gone I'll have to revert to using a native client, which I don't trust one bit.

Why do you think it is impossible. There are plenty of JS BitTorrent implementations.

They however only work with other browser clients via webrtc p2p.

It could be done with a local TCP / UDP <=> WebSocket gateway, especially if it's written in a safe language like Rust.

Why use a webapp AND a local native gateway instead of just a native app? If your goal is to just not open any native win32/cocoa/etc UIs for some reason, there are existing native torrent clients with web interfaces served over http and those have been available for like a decade plus.

Webapp + native gateway also means that the torrent traffic ends up being suspended/throttled if the tab isn't foregrounded or if you close it, something you wouldn't have to deal with if the native gateway was just a native torrent client. Chrome Apps had a background privileged context that could keep running even if no tabs were open (though Google naturally discouraged this unless the app needed it), something you can't really get with a PWA currently (though Service Workers come close if you keep the tab open, I think? Maybe?)

One benefit would be to protect yourself from security issues in the main torrent code. Anyway it was just an idea, not a serious proposal.

Can we run native apps written in Rust/Go in Chromebook ?

> You probably aren't using Chrome apps anyway

My coworker sitting to the left of me still uses the Chrome App version of Postman because it works, and hasn't stopped working.

It has had an orange banner for the last year saying to upgrade but I'm pretty sure he never will, until it actually stops working.

I think in his mind, the Chrome version is fine and ultimately, he'd be going through the install process then potentially moving his saved requests to what is essentially an identical copy of the exact same application

I still use the Chrome app of Postman too.

Reason is that I'm just used to it and it works perfectly fine for every use case I've had. I did install the new version but they changed the UI around a bunch and it kinda ticked me off because I was just trying to do some work, not relearn an app because they wanted it to be shinier looking.

Probably as good a place as any to highlight how broken the Chrome extension approval process has become over the last month or so.

Updates that would take an hour now take up to 10 days, including minor updates like a change to an extension's description.

Having your deployment date be determined by one of Google's algorithms is ulcer-inducing.

And the support is...well, see for yourself: https://groups.google.com/a/chromium.org/forum/#!forum/chrom...

They can fix this. G Suite support is fine. I would pay for this process not to suck.

Does anyone commenting on this thread even use a chrome app that's not just a wrapper on the site or a pwa? Just curious.

I worked in edtech for a couple of years and we had our own Chrome app.

I want to think that Postman was one of these, at least at one point. There may also have been a SQLite browser that I used briefly.

Otherwise... yeah, I can't say that I remember using any of these.

I think I'd rather have all of the Electron apps that I use now share a version of Chrome rather than each package the whole shebang themselves though.

It's a shame, really. I used to play Granblue Fantasy with their Chrome app. It was very convenient and accessible.

I mean, it's right here: http://game.granbluefantasy.jp/

I've bought a few Chrome apps over the years, but I don't like it. They've all disappeared, or no longer work.

Its a worse user experience than plain ol' apps, imho. Definitely not good value for money when you can't even use the apps a year or so later ..

imagine relying on google to maintain a product in 2020

Not really appropriate here. They announced this change to an open standard...in 2016.

Any idea if this means they will also get rid of:

"more tools" > "create shortcut" > "open as window"

I use this exclusively and chromium is the only browser left that supports it.

I learned about the existence of these today, from this thread.

I'll miss GetPocket app, despite what the article said it works offline on Desktop. Anyone known about an alternative?

Tip: theverge (like medium, etc.) is one of those sites that is better without JS.

I believe that Authy has both a Chrome app and an extension.

What does this mean for the Outlook "App"?

I imagine the functionality will be passed back to a browser tab. Most Chrome apps were basically just that.

Outlook is a PWA now I believe

Does this include Apollo client devtools???

What about apollo dev tools?

I have Chromebooks.

I'm now concerned and confused. The implications are not adequately explained.

This really goes to show that if you're buying something for the potential of it, you're not always guaranteed to be an early adopter of a hot new trend or technology. Sometimes you are, but not always. You have to love what you're buying for what it can provide you right now.

This philosophy doesn't exactly work in this context; the devices are losing applications provided right now.

Is this just a marketing move? Why is Google doing this? I also wonder how will this affect Chrome's market share. I think a lot of people appreciated the apps, especially the adblockers.

Google kills Chrome Apps, not Chrome extensions.

You're right, my bad. Sorry for the confusion.

It is such decisions from Google that make me question if they will continue supporting GCP down the line.

How will google pivot away when entire companies are dependent on their service staying up ?

Sometimes I think Google should simply spin off less than super profitable ventures into a small company and host it under alphabet.

As long as the service breaks even with employees working on Google wages, it can stay up. Once it goes into losses, the company can file for bankrupcy or move employees back to google.

You should read the rest of the thread if this move makes you question Google supporting GCP.

Disc: Googler.

Reply on news.ycombinator.com