I'm really glad Chrome forces the update. Because I would be sad supporting Chrome 32, 33, 34, 35, ..., 40, 41.. in my Chrome Extension. I would hate using polyfills everywhere and modernizers for my webapps. All I need to care about is 1 version. The latest.
Sadly this doesn't happen sometimes as people still don't upgrade, but most do which is fine for me. I hope it stays that way.
As for crashes, that's pretty bad to be sure, but I didn't see the OP try filling a bug-report.
I think a lot of people are missing the point of the author. He's not against auto-updating browsers. He's against
1. Violation of stated policies
If you tell me I can disable updating and then update anyway, I have been lied to. A promise was broken, an expectation violated. It is totally reasonable for him to be upset about this.
2. The excessive harassment of a tiny minority of users.
Several people in this thread have lauded the default nature of Chrome updates, and they're right! But it is that very default which makes the harassing banners and warnings all the more pernicious. Who is google bothering? The 99% of users we (as developers) care about and who haven't gone to great lengths to disable auto updating? No, they are only bothering the tech savvy, persistent, already annoyed user.
This is simply not where the burden of support comes from. As an anecdote the company I work for recently received a complaint that our website wasn't rendering properly. On investigation it turned out the user was trying to browse in lynx! While we're obviously disappointed that wordpress isn't lynx friendly we are going to spend 0 minutes trying to fix that.
>If you tell me I can disable updating and then update anyway, I have been lied to. A promise was broken, an expectation violated. It is totally reasonable for him to be upset about this.
I don't think they told him anything -- he just used an unofficial method to stop it that worked before, and then they changed it.
Chrome's updates aren't infallible. A year or so ago, there was a Chrome release which was crashing nearly constantly. Fortunately, switching to the "beta" channel left that bug behind. It got re-introduced in the next update, so I followed the actually stable release back to the "stable" channel. When the following version arrived in "stable", the re-introduced bug hadn't been caught, so the crashes started again. Switched back to "beta" again to dodge the bug, and have stayed there ever since -- it's been quite stable.
Conceptually, that was a helluva pain in the ass compared to simply downgrading to a working release.
That said: would I rather live in a world where everybody is running an up-to-date browser, even if that browser is occasionally buggy? Or would I rather live in a world where everybody is running a panoply of wildly out-of-date browsers, just because they're afraid to click "yes" on an upgrade? (I know plenty of less-tech-savvy people have been trained to never approve a dialogue box which they don't thoroughly grok; for them, browser and system updates, etc., will NEVER be applied.) The answer is very much the former. It's absolutely worth a rare bit of personal pain for the collective benefit of an up-to-date browser ecosystem.
I agree that auto update should be the default. But people who are scared of clicking "yes" on an upgrade are not the ones who will go into the registry or edit Windows group policy settings to prevent updates.
I think what Google tries to do is to prevent sysadmins from doing with Chrome what they did with IE in the past. The average sysadmin loves to freeze things because that reduces the number of variables that matter for troubleshooting.
Or they don't want to risk finding they've installed something with a wonderful new undisclosed Zero Day exploit throughout their shop.
Updating in a corporate environment is very different from what you do at home or a small business. Some sectors require extensive testing before rolling out any change whatsoever.
In that case, they don't want things happening before they have tested any proposed changes.
When a package is blown down to the network with it's configuration locked (and set to not update), they don't expect a package to spontaneously start updating itself.
Google might let their browser ignore "no auto updates," but may find their browser losing market share in the corporate world. If they get booted, chances are 50/50 if the app ever make it back in the mix.
It doesn't take much.
If I wanted a closed ecosystem, I'd have chosen Apple. Google is becoming less and less appealing with every new decision they announce.
> In that case, they don't want things happening before they have tested any proposed changes.
True, but let me continue that sentence: ... and as they have too little time for testing, they tend to resist all change.
There are a few very capable IT organisations that have a systematic approach to change. It would be a sensible thing to give them the necessary tools to manage change. One such tool is to delay browser updates.
But there are many more incompetent IT organisations that are overwhelmed and will keep the same browser version for 10 years if you let them.
I'm just making an observation. I don't know what the best solution is.
The time isn't an issue at that level. They have specific tests for various situations and the time it takes, it takes. I've seen banks spend upwards of 6 months testing before instituting software changes to trader/production systems.
As for capable IT departments, on the large corporate scale, they fairly bulge with competent people. Obviously not everyone's a star, but they don't suffer idiots to live.
I don't upgrade anything until it's been on the street for a time. I let other people play canary.
This is why my company won't allow Chrome on their workstations - it's a security risk. As far as I'm concerned installing Chrome on your system is installing a backdoor for Google. You don't know what that next version is going to bring, what bugs and vulnerabilities it will contain. All installed for you automatically! No thanks.
The author of this article obviously doesn't understand the immense pain of a world where most browsers update manually. Besides the enormous security issues, that also gave us IE6 headaches for 10 years.
I've honestly never heard of anyone having this crashing issue, and it sounds like an edge case. The author is also still on Windows 7, which is a 6-year-old OS and probably not the highest priority for Google's testing of Chrome.
A few isolated edge cases of problems like this are worth an overall safer, FAR easier-to-develop-for web.
Given that Windows 7 is by far the most widely deployed and used version of Windows, I would imagine that testing it would be a pretty high priority for a company developing a desktop web browser.
"The author is also still on Windows 7, which is a 6-year-old OS and probably not the highest priority for Google's testing of Chrome."
What? At the time of the bug 50% of laptops/desktops on the world ran Windows 7. How could Win7 not be the highest priority for the Chrome team?
While I agree that auto-update makes live a lot easier for web developers, i think this strategy of actively preventing users from running an older version requires more responsibility on the developers of Chrome to make sure the browser is stable.
It's not that simple to just upgrade from windows 7. There are other things than browsers that people care about. For example that game collection that was bought 10 years ago. Microsoft has blocked a few copy protections from that time so the games will not run any more. Buying a new computer just to keep your game collection is usually not an option when everything is working just fine as it is.
"the pain" ? Look at Firefox, it checks if their is a new updates and asks you if you want to update. (Well last time I checked anyway). So unless you disable those update checks, yes, then it might be a pain. But you either want that or not.
In the past, when it didn't, people got stuck on old, insecure versions. There was a period of time (pre-Chrome) where people would replace IE with Firefox for their parents, and then their parents would be many versions behind.
Thanks for reminding me of more information that proves my point. If people aren't forced to update, many of them won't, and that makes all of us less safe.
This is actually an identical controversy to vaccinations. Saying that you're against automatic updates because 0.000001% of people have problems with them is the same argument that anti-vaxxers make.
Eh. I'm sorry the author had a bad experience, but ultimately, we're better off in a world with evergreen browsers.
If we let people who think they know what they're doing decide what browser version to run, then people who think they know what they're doing will decide what browsers version to run. It's not an acceptable tradeoff.
I'm glad to hear they are forcing updates now. A family member's computer had some extension/malware that disabled auto update of Chrome so they were several versions behind. It seems like this policy is there to prevent malware from being able to lock down Chrome.
This issue is why I don't use Selenium (in Windows) with anything but Firefox. A forced upgrade can render your boardroom demo non-functional moments before you have to give it.
Use the best tool for the job. I use Google Chrome for Googly things. I use Edge/IE for Microsofty things, and I use FF for the actual web. When javascript/Flash/HTML5 video in FF is too slow, I use Chromium.
I use Chrome for everything, and it all works great. I never have to switch to another browser the way you're describing. That sounds truly exhausting (as well as being inefficient and unnecessary).
I understand why Google is doing this as I have had to trouble shoot a few virus laden pcs where the auto update was blocked by the virus by using the registry workaround written about in the post.
I had to nuke the whole install and start from scratch. Just to be able to install the latest version of chrome
I suppose if a bit of software crashes all the time, dumping it is a fair decision.
But of all the software I use, Chrome crashes the least (single tab page crashes very occasionally). I stopped using Firefox years ago for Chrome because Firefox crashed so much.
I would want to know why your Chrome crashes so much.
I am not reaching conclusions about Firefox today, I don't know where you got that from. I haven't used it because Chrome has been good to me and I haven't had a reason to switch back.
Unlike Chrome, Chromium is open-source, so presumably no one is forbidding people to distribute it in binary form and archive older releases. Worst comes you could always retrieve the source code of older versions and build them yourself: https://dev.chromium.org/developers/how-tos/get-the-code
As for the problem of auto-updates, if you run Chromium on Linux, install it as root and run it as your normal user (as distributions usually do), then presumably it won't be able to update automatically.
So many people here are pretty quick at jumping on the author and defending Google's behaviour but I don't think those people actually understand what the issues are here.
There's nothing wrong with Google's updates except when they very rarely end up breaking something, but that's not the issue. It's that when someone has decided to block auto-updates from Google, it will still sometimes work around that to update itself.
99% of the people will likely want or not care if it auto-updates, but it shouldn't try and fuck over people that don't want that and expect that.
I find the current browsers landscape not great actually. I'm a long time FF user as my main browser but lately I find it anormaly slow and unresponsive on my mac (using stable version of FF). I'm so bothered that for the first time I thought about switching to another browser but as I'm mindful about my privacy I don't want to use Chrome and also I'm not using Safari either because there are UI things I don't like and too few extensions available. So inertia, inertia, I'm keeping my FF but I'm not happy though.
On Windows and Linux, FF is faster for me than it's ever been. I wouldn't be surprised if your issue were caused by extensions or by OS X itself.
The quality control for OS X has been spiraling down the drain for several years. I've had tons of bizarre, work-stopping bugs on OS X lately. I had to stop using it except to build apps.
Of course it's hard to say, what I can tell you is I have not changed my hardware, this behavior is quite recent maybe since 1 or 2 versions back, I tried disabling extensions, I tried keeping a small number of open tabs on my main window (but I also must keep open tab groups with maybe a total of 30 tabs for things I'm working on), all that with no success. When I click on the button to close a tab I observe some latency to react and effectively close the tab. When tabs reorder on the left as tabs are closed I also observe latency and sluggishness and likewise with scroll on some sites.
Try the developer edition of Firefox. I had the same problems with stable Firefox on OS X, it felt slow and unresponsive most of the time, and would regularly choke on js-heavy pages.
Dev-edition feels much more performant to me, plus it gets new features sooner than stable.
I've been using Chrome for six year on the month now. Probably about the same amount of time I used Firefox for before that. Been thinking a bit about switching back lately even though I have no real problem with Chrome, just think it might be healty to change stuff, six years is a long time. Plus, FF seems to be working better on a couple of points on my linux setup.
I know it won't be the same as when we all moved from IE to Firefox or Firefox to Chrome all those years ago, so I'm not in a hurry, and no Netflix in Firefox yet.
This author seems to think that simply writing about it, replacing graphics card, OR disabling updates will simply solve the problem for Chrome, but also not affect browsers (without evidence to the contrary). They did all of this without ever filing a bug report to the Chromium team.
So, what is everyone who isn't using Chrome going to use?
It's been the defacto browser for the modern web for as far back as I can remember without thinking too hard. (Like I know once upon a time I used some web spider search, and Netscape and IE6 forever... but I haven't realistically thought of a main browser other than chrome in forever).
I'm going to call this out as trolling. Chrome was first released in 2008. FF, Mozilla, Opera, etc. were already known and used at the time. Anyone who cared about web dev was not using IE before IE7 was released. (at least not as their primary browser) Just look at some basic stats, there are lots of alternatives.
Chrome _is_ Safari is on your iPhone. It's just a UI layer on top of Safari. Apple doesn't allow third party browser engines. It's why Firefox isn't on iPhone but it is on Android.
But what does any of this have to do with desktop browser choices and auto-updates?
You're right. Google did fork WebKit. But Apple doesn't allow that fork (or any other rendering engine) on iOS. See GP: " Apple doesn't allow third party browser engines."
For a while, before iOS 8, Safari would be the only application to take advantage of the Nitro Javascript Engine. No other web view (Chrome for iOS, Facebook, etc) was able to take advantage of the increased JS performance. I had heard that it was due to the way Nitro accessed memory within iOS that prevented it from being safely used by other developers.
Apple does this when testing a new feature in both iOS and Mac OS. They make it a private frame work and test it in 1 or 2 of their own apps until the feel it is fit to release for everyone.
I'm unsure how you can think this, really: according to Wikipedia[1], Opera was first released in 1995 (whilst Google was founded in '98, Chrome released in '08).
I wondered if maybe you meant that it had forked the Chromium webkit stuff, but it seems that Opera is using proprietary engines.
Opera is not, as an entity or in terms of identity, a fork of Chrome.
However, it is, as rendering and JS goes, functionally a fork of Chrome, since as you'll notice from the source you cited, it uses Chrome's "Blink" rendering engine and Chrome's "V8" Javascript runtime.
Also from the link you cited:
>On 12 February 2013, Opera announced it would drop its own Presto engine in favour of WebKit as implemented by Google's Chrome browser, using code from the Chromium project. Opera Software also planned to contribute code to WebKit.[29] On 3 April 2013, Google announced that it would fork components from WebKit to form a new rendering engine known as Blink; the same day, Opera confirmed that it would follow Google in implementing Blink.[30]
>
>On 28 May 2013, a beta release of Opera 15 was made available,[31] the first version based on the Chromium project.[32][33] Many distinctive features of the previous versions were dropped, and Opera Mail was separated into a standalone application derived from Opera 12.[34]
Did this a few months ago. This is what happened:
* No syncing history with phone
* No chromecast, literally had nothing to watch on TV
* Even slower than Chrome. Thought Firefox was supposed to be fast, but nope.
Firefox does both of the things you say it doesn't do[1][2]. Also, it's strange that you had "literally nothing" to watch on TV because casting your browser couldn't be mirrored onto the TV... what about Netflix, Hulu, etc.?
Assuming I'm reading the parent post correctly, neither of those links actually do what they would've been hoping for when moving browsers, even if they certainly get closer than nothing.
Firefox Sync only syncs with versions of Firefox, and Chrome's sync only syncs with versions of Chrome, so they'd have to switch mobile browsers too just because they switched desktop browser (and unlike on desktop, I've found Firefox for Android to generally be clunkier to use than Chrome for Android).
And that second link is about Chromecasting from Firefox for Android, not the desktop browser. Being able to send stuff from desktop Chrome is definitely a convenience someone could get accustomed to.
I own an iPhone and there's NO firefox on iOS. The benefits of having synced history across every browser I use is pretty useful(Ex: that restaurant I looked up online on my laptop this morning is going to show up pretty quickly on my phone because it's in my Chrome history)
Also, Chromecast is just amazing. It works with Netflix, Hulu, YouTube etc. Browse for videos on the website, because it's usually easier to click than to do weird controller motions => chromecast. Back to coding. DONE.
I have no hard numbers or evidence for performance, but that was just my feeling.
Of course you'd have to switch mobile browsers, since mobile browsers don't currently have the option of installing extensions that would allow cross-browser syncing.
That may be a deal-breaker for you at the moment, but I believe that when Firefox's rendering engine switches to Servo, it's going to blow Chrome away on both desktop and mobile.
Sadly this doesn't happen sometimes as people still don't upgrade, but most do which is fine for me. I hope it stays that way.
As for crashes, that's pretty bad to be sure, but I didn't see the OP try filling a bug-report.