Ah, WebGPU, one of the “features” I disable anyway because its a major fingerprinting data point and no website is got any business running code on my GPU
As long as it asks for permission first, I wouldn’t mind having the option. I’d almost always rather run something in my browser than as a separate executable, especially on a mobile device.
Same.
Why not? It’s already executing code on your CPU unless you run noscript or similar…
Are there other features you’ve turned off? WebRTC perhaps? Others?
Why not?
If I don’t know what your code is doing, and I can’t modify it, I don’t want it. Are you rendering a cube? Are you mining crypto on my machine? And why would I need GPU-rendering? Webpages should be text, images and hierarchy.
It’s already executing code on your CPU unless you run noscript or similar…
That’s true! Now guess what extension I’m using
Are there other features you’ve turned off?
You bet. Some I tolerate not disabled, but spoofed or containerised per site.
If I don’t know what your code is doing, and I can’t modify it, I don’t want it.
To be fair, 99% of the world population do not share this viewpoint.
Sure…?
I make games for Ludum Dare fairly frequently, and it’ll be a nice feature, because a lot of people will refuse to download a game and only play web versions.
Other than that case, I agree. What’s the point? Your page probably shouldn’t be doing anything where it needs the GPU. What information is a page trying to present that a GPU is better at rendering than the CPU? Maybe very niche topics, but usually text is ideal.
A 3d map for example?
Sounds like a pretty solid setup, I also like temporary containers solutions as well, less hassle IMO.
Does anyone else think that browsers implementing every single feature of an operating system is a dumb idea?
Not really. Browsers were one of the first pieces of software to do sandboxing, but now virtually everything uses sandboxing for organization and security - Android apps have a permissions manifest so they can be sandboxed. Amazon cloud servers are mostly Kubernetes clusters, which is just sandboxed virtual machines. ChromeOS already is a OS/browser hybrid with native sandboxing (and the short lived Firefox OS. Running a 32 bit app in a 64 bit environment requires a compatibility layer, which is a sandbox. If browser technology has already been pushed through the OS stack, why not complete the loop.
The main use case for hardware acceleration is progressive web apps, which is literally a plan as old as 2006 to make browsers able to securely run signed code natively (as an alternative to using extensions like ActiveX, Java, Shockwave, etc, all of which were notoriously insecure).
So honestly, I don’t think it’s a dumb idea at all. It would honestly be kinda cool if I could go to blizzard.com and just launch a game full screen, securely with a simple approval rather than downloading and running a separate launcher app. (Assuming the implementation was otherwise sane; I know the current environment of enshittification could torpedo the idea entirely)
Yep.
Desktop OSes are not suited to running untrusted code, unfortunately, so you want to run as much as possible of your closed source megaco software in the browser
It can come off as a logical conclusion to most people never using anything but web these days.
Maybe we should not listen to those people
Well, not. Browsers are slowly becoming the cross platform for your apps and it makes a ton of sense from resources (create a single app vs an app per platform) perspective.
TBF, Chrome doesn’t support it on Linux either.
Stupid original title
Firefox Catches Up to Chrome With the Addition of This Feature But Leaves Linux Out (for now)
Itsfoss is bottom tier rehashed junk.
