Pretty sick, chrome makes it clear that silent updates can definitely be a good thing if used appropriately. Compare to the sad new version adoption rates for pretty much all microsoft products.
Although I do agree with you that silent updates are a far better choice, its not perfect, sometimes it can introduce unexpected changes.
Let's take for example the latest Chrome update (v32). It replaces the OS based scrollbar with a custom one. It also removes the ^v arrows around it. As someone who regularly uses them, it breaks my workflow :( .Sad but I can live with it. Had it ended there it would have been acceptable but the latest update also introduced a series of bugs, for example on when hovering over a html select box, the last item is never highlighted[1] and there is also an issue with scrolling on a select box [2].
There is a reason why silent updates/auto-updates aren't popular in the server/enterprise world because it could sometimes break things.
It's been a week now and my select boxes don't scroll, breaking a good percentage of websites. And that annoying as hell removal of the scroll widgets.
I shudder to think if you're using Google Chrome in the enterprise.
Not sure why you have been downvoted. Chrome 32 was a very disappointing release and has undermined silent updates for me. Many users are experiencing crashes and difficulties with general scrolling and select boxes. These are faults introduced without any must-have feature. If you do silent updates you have to be incredibly disciplined about quality and scope.
My bug-bear is also select boxes. Why do we now have this ridiculous ease in animation? They have crossed the line adding styling that can't be removed with css. I use a site with a dozen select boxes and the effect of each click causing a zoom-in animation is nausea inducing and unfixable.
You make it sound like that's not fathomable, but an astonishing number of security issues are caused by insecure font and UI rendering.
> How about if they fix series security issues promptly without simultaneously breaking basic UI elements?
While I agree that it's great if things don't break, I'm not sure what is really broken. I'm running Chrome right now and the scrollbar doesn't bother me. (It seems more like somebody not liking a change than it being "broken.")
If you really dislike the change so much, you can just disable the auto-updater--but then you're on your own. Running an old browser is like sailing in a shark-infested ocean in a wheel of swiss cheese.
Either way, you're suggesting they fix a hole in all previous versions of the software, and release new, fixed versions with no other changes. This is not feasible for a browser.
Sadly, Chrome v32 has several bugs (regressions) and the unpleasant Mac-style scrollbar on Windows.
Chrome 32 also introduced a WebGL related bug where it calculates the view matrix wrong in combination with mouse-lock.
They fixed it already in Canary build, but where is the QA department? They should push such bug (regression) fixes back! Or push out a new major version.
Sadly, Chrome v32 has several bugs (regressions) and the unpleasant Mac-style scrollbar on Windows.
I think there are quite a few people in the industry (and in SF especially) who seems to think that a Mac esthetic is universally accepted as a superior thing.
It's not. Not by a mile.
Consistency is superior. Breaking consistency is bad, always. So adapt to the platform. Adding custom code just to break with platform-conventions is just a mind-boggling thing to do.
Maybe someone should start forcing engineers at Google to use Windows so they get back down to earth?
Edit: I'd just like to add I react the same way when I see Android-apps which are forcing iOS-conventions on me. It completely kills my work-flow and those apps tends to get uninstalled pretty quickly.
Chrome 32 really is a terrible update with some major regressions all over the place, all (apparently) for the sake of some custom scrollbars. To add onto the list of issues, popup windows (including things like Chrome dev tools in their own window) constantly pop up a "Page is not responding" error, despite working just fine. Same for any Chrome apps that you might be running in their own windows.
I switched to Canary for the time being because 32 is just awful. Canary has saner scrollbars and also lacks the annoying slow zoom transition when opening dropdown menus (seriously, who the hell though that was a good idea). They really should hotfix the hell out of 32 ASAP.
Might just actually go back to Firefox one of these days, because Chrome is going into some annoying directions in general.
I went back to Firefox 7 months ago. That was when it became better than Chrome: faster, more stable, consumes fewer resources, supports more tabs, better UI.
I still think Chrome peaked around version 20, I have an portable version of it around. With it I can open hundreds of tabs without any problems and it is perfect. Chrome 32 is so bad that I question the QA department/process of Google. Three days ago I Canary crashed on start (Win7) had to wait til next silent update, but now it's a lot better than 32.
Firefox 26 is great, the only missing thing is multi-process support - otherwise a single website can hang the whole browser... not so good if you open dozens of tabs and keep them open for days.
I think it was the Chrome 26/27 era that finally drove me away....
Understood re Firefox 26 multi-process support, but it makes a big difference to resource consumption. I've not had any problems but if necessary you could kill the browser thread and restart excluding the bad tab....
Not everyone wants silent updates. I don't want Google or Microsoft replacing binaries on my PC behind my back.
When you see that Google wants to vacuum up more and more data from your PC, it makes sense that they silently update their software - in that context. Chrome continues to find different areas to mine user data e.g. turning your address bar into a keylogger, or monitoring your browsing activity even if you're not using google services via "phishing and malware protection", etc etc.
If used appropriately indeed. Chrome is a relatively safe thing to auto-update, since they have a very good QC process and most PCs have a fallback browser if things do go bad (which afaik hasn't happened yet).
Compare with for example auto-updating virus scanners, which in more than one occasion marked a critical system file as a virus and made people's computers unable to boot up. Caused a lot of damage.
"biased" how? That simply means that for the more recent data set, the owners bothered to decypher a small set of user agents (those represented in the graph - latest verions of dominant browsers), everything else is marked "other" and it shows that in Jan 2014, ~20% or fewer of clients were connecting from anything but some latest-version browser (and of whatever resources this covers, Chrome is by far the leader). That doesn't strike me as bias - just data simplification.
The only show the current versions of Firefox and Chrome on the Top Versions. The other category drops as soon as Firefox and Chrome gain which means that past versions of all browsers (eg, Chrome 31, 30, etc) are grouped into other. The "Top Families" show all versions together of each browser, and you may notice that the other category on that page is very minimal.
The Firefox graph [1] also looks strikingly similar after introducing Silent Updates since Apr. The trend of 'Move Fast And Break Things' seems to be coming from the web applications world into the world of installed applications.
Users seem to be more adaptable to change that previously thought. This is of course thanks to better testing tools which ensures that essential functionality is not compromised by such changes.
With the introduction of OS-level App stores, expect to see 'silent updates' in a lot more consumer applications.
And releasing more often means that the changes are smaller.
Also, developers have introduced a more clever way of introducing changes to the user. Overlays with explanations of new menus etc. which I've seen first on the web (i.e. gmail) are making their way to more and more "desktop" applications.
Introducing these changes slowly over time seems to avoid the "I will never install this version" effect.
the interesting counter though is that people end up being more likely to be "stuck" with something they really don't like, and get to have it happen in slow motion (see the gripes about gmail).
I wonder if that that makes it easier or harder to convince people to use alternative software
Indeed. In fact, IE10 and IE11 show mild weekend/Sunday dips, but IE7/IE8/IE9 show severe dips, and Chrome and Firefox noticeable Sunday dips, too.
Since these are percentages, someone must be surging on weekends/Sundays... and that does appear to be the case for Safari and Opera. Didn't expect that, myself.
Depressing to see a proprietary browser like Chrome gain such marketshare.
I thought the dark days of "Only works with Internet Explorer" were behind us, but Google is pressing for a closed-web again with some of their features.
33 comments
[ 3.5 ms ] story [ 90.6 ms ] threadLet's take for example the latest Chrome update (v32). It replaces the OS based scrollbar with a custom one. It also removes the ^v arrows around it. As someone who regularly uses them, it breaks my workflow :( .Sad but I can live with it. Had it ended there it would have been acceptable but the latest update also introduced a series of bugs, for example on when hovering over a html select box, the last item is never highlighted[1] and there is also an issue with scrolling on a select box [2].
There is a reason why silent updates/auto-updates aren't popular in the server/enterprise world because it could sometimes break things.
[1] http://productforums.google.com/forum/#!topic/chrome/dxXGUaF...
[2] http://stackoverflow.com/questions/21166379/html-select-scro...
It's been a week now and my select boxes don't scroll, breaking a good percentage of websites. And that annoying as hell removal of the scroll widgets.
I shudder to think if you're using Google Chrome in the enterprise.
My bug-bear is also select boxes. Why do we now have this ridiculous ease in animation? They have crossed the line adding styling that can't be removed with css. I use a site with a dozen select boxes and the effect of each click causing a zoom-in animation is nausea inducing and unfixable.
Thankfully we're seeing a shift where people will accept this if it means serious security issues are fixed promptly.
I frankly couldn't care less if my scrollbar acts a little strange if it means two remote code execution holes were fixed.
You make it sound like that's not fathomable, but an astonishing number of security issues are caused by insecure font and UI rendering.
> How about if they fix series security issues promptly without simultaneously breaking basic UI elements?
While I agree that it's great if things don't break, I'm not sure what is really broken. I'm running Chrome right now and the scrollbar doesn't bother me. (It seems more like somebody not liking a change than it being "broken.")
If you really dislike the change so much, you can just disable the auto-updater--but then you're on your own. Running an old browser is like sailing in a shark-infested ocean in a wheel of swiss cheese.
Either way, you're suggesting they fix a hole in all previous versions of the software, and release new, fixed versions with no other changes. This is not feasible for a browser.
I have switched to Firefox for now, but I do miss Chrome : especially the awesome Stylebot user-styles for Hacker News, Reddit etc.
Chrome 32 also introduced a WebGL related bug where it calculates the view matrix wrong in combination with mouse-lock.
They fixed it already in Canary build, but where is the QA department? They should push such bug (regression) fixes back! Or push out a new major version.
I think there are quite a few people in the industry (and in SF especially) who seems to think that a Mac esthetic is universally accepted as a superior thing.
It's not. Not by a mile.
Consistency is superior. Breaking consistency is bad, always. So adapt to the platform. Adding custom code just to break with platform-conventions is just a mind-boggling thing to do.
Maybe someone should start forcing engineers at Google to use Windows so they get back down to earth?
Edit: I'd just like to add I react the same way when I see Android-apps which are forcing iOS-conventions on me. It completely kills my work-flow and those apps tends to get uninstalled pretty quickly.
I switched to Canary for the time being because 32 is just awful. Canary has saner scrollbars and also lacks the annoying slow zoom transition when opening dropdown menus (seriously, who the hell though that was a good idea). They really should hotfix the hell out of 32 ASAP.
Might just actually go back to Firefox one of these days, because Chrome is going into some annoying directions in general.
Firefox 26 is great, the only missing thing is multi-process support - otherwise a single website can hang the whole browser... not so good if you open dozens of tabs and keep them open for days.
Understood re Firefox 26 multi-process support, but it makes a big difference to resource consumption. I've not had any problems but if necessary you could kill the browser thread and restart excluding the bad tab....
When you see that Google wants to vacuum up more and more data from your PC, it makes sense that they silently update their software - in that context. Chrome continues to find different areas to mine user data e.g. turning your address bar into a keylogger, or monitoring your browsing activity even if you're not using google services via "phishing and malware protection", etc etc.
Compare with for example auto-updating virus scanners, which in more than one occasion marked a critical system file as a virus and made people's computers unable to boot up. Caused a lot of damage.
http://clicky.com/marketshare/global/web-browsers/versions/
http://clicky.com/marketshare/global/web-browsers/
Users seem to be more adaptable to change that previously thought. This is of course thanks to better testing tools which ensures that essential functionality is not compromised by such changes.
With the introduction of OS-level App stores, expect to see 'silent updates' in a lot more consumer applications.
[1]: http://clicky.com/marketshare/global/web-browsers/firefox/
Also, developers have introduced a more clever way of introducing changes to the user. Overlays with explanations of new menus etc. which I've seen first on the web (i.e. gmail) are making their way to more and more "desktop" applications.
Introducing these changes slowly over time seems to avoid the "I will never install this version" effect.
I wonder if that that makes it easier or harder to convince people to use alternative software
Since these are percentages, someone must be surging on weekends/Sundays... and that does appear to be the case for Safari and Opera. Didn't expect that, myself.
http://clicky.com/marketshare/global/web-browsers/
The highs and lows for IE (orange) match the opposite for Chrome (pale blue). You can hover over them to get a specific date and then check the other.
You can see this in other data sets too:
http://gs.statcounter.com/#desktop-browser-ww-daily-20131229...
Compare those with IE.
I thought the dark days of "Only works with Internet Explorer" were behind us, but Google is pressing for a closed-web again with some of their features.