Could there be additional motivations behind Microsoft's shift away from its original UI frameworks?
Could there be additional motivations behind Microsoft's shift away from its original UI frameworks?
I understand the core solution likely involves financial resources. Recent approaches often rely on simpler yet less optimal structures that are currently popular (like React), but I believe it's not wise to completely discard the authentic desktop feel. For instance, system configurations in MacOS and Linux (especially Linux Mint) appear natural within their operating systems, similar to how Control Panel functions with Windows. In these older applications, control elements remain consistent, and adjustments from Windows Settings or other programs don't disrupt the native experience.
They've been attempting to shift away and update the Windows NT interface since its inception (Windows 2000), but this effort remains only a partial success. The operating system is a tangled mess of confusing symbols and outdated references that weren't properly recorded. The settings area still carries remnants from Windows 8, which introduced significant changes, leaving you with a confusing layout where it's unclear what each element does. Efforts to unify the user interface have been repeated multiple times, resulting in several incomplete specifications at once, causing strange visual glitches and other problems. Typically, these major redesigns were postponed because of budget constraints. Now, everything is shifting toward web-based applications using WebView2 to reduce development expenses, which introduces its own set of challenges.
Gnome Shell also uses Javascript. And I think some DEs are now configurable with CSS rules. So there is a trend towards incorporating more web languages in all software because iteration and configuration are faster than if you have to keep recompiling. Not sure if this is a good idea for something so core like an OS file system application, where you need millisecond-fast responsiveness. But for something like a chat app or image viewer, web languages and tools can be very viable.
Innovative concept: AMD and Intel embed a DOM, JavaScript, and CSS engine right onto a chip, enabling web views to function within native apps with extremely fast speed.
You can still sense it, but it bothers me greatly. The internal Photos feature (which consumes around 200 MB of RAM) runs sluggishly compared to native options like the old Windows Photo Viewer or programs such as Faststone Image Viewer or IrfranView when using much less memory.
Photos was originally an UWP application, likely developed using C#. It has now been transferred to the Windows App SDK and rewritten in C++.
They fell short because the 2000 style was meant to be phased out by vista, but it wasn't limited to just surface-level details. Remember that the modern design languages are also native. As long as it isn't a webview2 app, it runs directly on the system. It's just poorly written with weak performance.
It seems you were trying to say something more natural. The UWP app ran well, though it wasn’t the quickest compared to other options.
Meta adapted WhatsApp for Edge WebView 2. Its performance and memory consumption are much lower than expected. Generally, web-based apps lag behind native compiled programs. Examples include Discord, VS Code, Spotify, YouTube—though the latter lacks a desktop app but remains comparable. I’m eager for Microsoft to fully rewrite their OS using code and implement JavaScript at a low level. That would be beneficial. Linux adoption could really grow.