why the fuck would you even want this? this is not a feature. you're just exposing information about the user's desktop layout for no benefit to the user 
it's one thing if some app decides to stop doing work because i tabbed away, but quite another if it decides to fuck off because some window happens to cover it up on the desktop momentarily. software does not work like this on any sensible operating system
choosing to use Chrome is basically letting a parade of clowns stroll in and fuck with your workflow every 6 weeks
Probably. On mobile they use page visibility to block background YouTube playback. The desktop version of YouTube does not currently do this (so 'request desktop site' in FF bypasses it) but clearly they plan to and this makes it harder to bypass.
Another shout out to the Microsoft Edge crew for deciding it's okay to just roll over and let the Chrome team do stuff like this to Microsoft customers
Looking forward to a future Chrome release updating this to use gaze tracking so that advertisers can tell whether you were actually looking at the screen when the ad played
Wow look at that! What a good feature
zdnet.com/article/google…
zdnet.com/article/google…
I wonder if incidents like this are eventually going to cause a backlash against A/B testing on users. At least in scenarios like this where IT is deploying Chrome with group policies it's unclear whether everyone involved wants to be an experimental subject (or should be).
It's already pretty nasty that if you want to roll back Chrome to troubleshoot a bug in the latest release, you can't. They don't offer old binaries and they immediately force-install updates on next restart if you install an old version. Add experiment flags on top and yiiikes.
The most recent incident w/extension signing in Firefox has shown why production configuration systems are useful - you can rapidly fix users' computers or shut off broken features - but using that mechanism recklessly makes the software seem alien, hostile, and unpredictable.
A fun footnote: Granblue Fantasy players were telling each other to disable the 'web-contents-occlusion' Chrome Flag on discords days ago, but because Google's documentation is so poor I couldn't tell that flag caused tab unloads and not just visibility events for ads.
Check out my soundcloud, and also Firefox has an Extended Support version that updates less frequently but still gets important security fixes, unlike Chrome which rolls out a blob of new stuff every 6 weeks whether you like it or not
mozilla.org/en-US/firefox/…
mozilla.org/en-US/firefox/…
Thinking about the malpractice demonstrated in this exchange
"We did a thing to some of our users without telling them, and they didn't complain about the thing they didn't know we did, so we had no idea it was bad"
It's not just the Chrome team getting testing wrong this way
"We did a thing to some of our users without telling them, and they didn't complain about the thing they didn't know we did, so we had no idea it was bad"
It's not just the Chrome team getting testing wrong this way
At this point the software industry is full of people who write code for millions of non-tech-expert users as if they're writing it for a small group of tech experts who understand the way computers work and know how to triage+report issues.
Maybe this is just an education gap, people who go from uni CS student to Chrome developer? My first customers were civil engineers, writers, artists, so I learned quick that my FileMaker or VB6 apps needed to be robust and that users wouldn't contact me unless they were stuck
If the app you wrote crashes every hour and they have to use it for work, they'll just restart it every hour. Maybe every 30 minutes in the hopes that they can avoid the crash. They might avoid doing certain things to avoid the bug they vaguely guessed at instead of reporting it.
The planning document for this Chrome change reveals another frustrating issue: For it to work, they have to hook every process on your machine and do an occlusion check every time any window moves. Their goal is to speed up *Chrome Tabs*, but doing so imposes a system-wide cost.
If I have a tab open playing a looping video that I can't see, this occlusion check will reduce battery usage. On the other hand, if I have a chrome window open sitting on the new tab page, chrome.exe eats CPU any time a window moves for literally no reason.
This is a great example of a blind spot that comes from only focusing on your software and not the bigger picture - for my use this cool occlusion feature is an anti-feature that will *increase* system-wide CPU usage because my tabs are cheap (~0% idle foreground usage).
