Apple released Beta 9 of #iOS / #iPadOS / #macOS 26 which is supposed to be the last one before release (at least for iOS/iPadOS).
-
Apple released Beta 9 of #iOS / #iPadOS / #macOS 26 which is supposed to be the last one before release (at least for iOS/iPadOS). I've seen quite a few reports that its new liquid glass is broken in more ways than is comfortable.
I understand that it's an ambitious redesign but I wonder if there was an ulterior motive to start in the first place.
#LiquidGlass introduces a rather complex rendering pipeline. It’s dynamic and uses a lot of complex shading and blending modes. It makes it practically impossible to recreate it in third-party UI toolkits.
Since (even before) #Aqua it was not too difficult to make a decent recreation of the OS UI in third-party toolkits if you could stitch together a bunch of bitmaps. In many cases it wasn't an exact recreation but it often was enough to blend in with the rest of the OS to a passable degree. Animations were probably slightly off and more subtle effects might’ve been missing but we had an acceptable level of replication so that at least in screenshots it looked fine to an average user.
#Java, #Qt, and even #web platforms (e.g. #Electron) had recreated OS UI and at least in principle had facilities to recreate it fully with enough effort.
This is not the case with Liquid Glass. As far as I can tell only Qt #QML might be capable to recreate it as it supports shaders for widget rendering. No other third-party UI toolkits (not Java, nor web, nor anything else really) provide facilities for such a complex rendering. In principle all of them can implement the effects but not within the widgets. In order to achieve even screenshot-passable fidelity you'd have to resort to rendering to a canvas.
If before apps using third-party UI toolkits lagged behind Apple now they might completely forego adopting the new visual language.
But this is an issue not only for third-party UI toolkits. It’s also an issue to many native apps. Specifically the ones that use custom widgets. Previous rendering models were relatively simple. So adapting widgets to blend in with OS was not too hard. It seems to me this time it might be much harder. You can't get away with swapping a few images/gradients any more. I wonder how long will it take for some apps to properly adopt Liquid Glass. Are we going to see a lot of apps partially in Liquid Glass (for the built-in widgets like regular buttons) and partially old-style custom widgets?
So far it looks like it’s going to be the worst design update for Apple. Liquid Glass is pretty broken on its own. That on the backdrop of the usual UI quality degradation in Apple products. And on top of that is third-party adoption of the new design language.
-
Apple released Beta 9 of #iOS / #iPadOS / #macOS 26 which is supposed to be the last one before release (at least for iOS/iPadOS). I've seen quite a few reports that its new liquid glass is broken in more ways than is comfortable.
I understand that it's an ambitious redesign but I wonder if there was an ulterior motive to start in the first place.
#LiquidGlass introduces a rather complex rendering pipeline. It’s dynamic and uses a lot of complex shading and blending modes. It makes it practically impossible to recreate it in third-party UI toolkits.
Since (even before) #Aqua it was not too difficult to make a decent recreation of the OS UI in third-party toolkits if you could stitch together a bunch of bitmaps. In many cases it wasn't an exact recreation but it often was enough to blend in with the rest of the OS to a passable degree. Animations were probably slightly off and more subtle effects might’ve been missing but we had an acceptable level of replication so that at least in screenshots it looked fine to an average user.
#Java, #Qt, and even #web platforms (e.g. #Electron) had recreated OS UI and at least in principle had facilities to recreate it fully with enough effort.
This is not the case with Liquid Glass. As far as I can tell only Qt #QML might be capable to recreate it as it supports shaders for widget rendering. No other third-party UI toolkits (not Java, nor web, nor anything else really) provide facilities for such a complex rendering. In principle all of them can implement the effects but not within the widgets. In order to achieve even screenshot-passable fidelity you'd have to resort to rendering to a canvas.
If before apps using third-party UI toolkits lagged behind Apple now they might completely forego adopting the new visual language.
But this is an issue not only for third-party UI toolkits. It’s also an issue to many native apps. Specifically the ones that use custom widgets. Previous rendering models were relatively simple. So adapting widgets to blend in with OS was not too hard. It seems to me this time it might be much harder. You can't get away with swapping a few images/gradients any more. I wonder how long will it take for some apps to properly adopt Liquid Glass. Are we going to see a lot of apps partially in Liquid Glass (for the built-in widgets like regular buttons) and partially old-style custom widgets?
So far it looks like it’s going to be the worst design update for Apple. Liquid Glass is pretty broken on its own. That on the backdrop of the usual UI quality degradation in Apple products. And on top of that is third-party adoption of the new design language.
Turns out Web can reproduce Liquid Glass effect: https://kube.io/blog/liquid-glass-css-svg/
The feature that powers the effect is backdrop-filter. According to MDN it's entered baseline in 2024 and is supported by all major browsers. Though, there seem to be some issue with SVG filters in backdrop in Firefox.
Electron, however, runs on Chrome so it, apparently, can blend in with new OS style.
-
R ActivityRelay shared this topic
