Hey everyone, I’ll take the liberty of creating this thread, and I’ll respond first below.
Interesting topic that @CubicleNate and @Norbert brought up regarding whether it is actually beneficial for a distro to copy the Windows look and feel. I would argue that there are two very different facets to that issue: 1) the Windows appearance, and 2) certain OS paradigms that Windows has established.
- Distros that superficially copy the Windows look’n’ feel end up being more frustrating to the inexperienced user, because inevitably the facade or veneer of Windows still has many foreign paradigms hiding behind them. Users will see a “Windows File Explorer” that they totally recognize and straight off go looking for “Map network drive…”, only to be met by the reality that the whole concept of “drive letters” doesn’t exist in Linux and try to wrap their head around the Unix concept of a unified filesystem with a single root. Then they’ll go looking for an EXE installer to add proprietary hardware drivers, with no clue about the entirely different paradigm of a monolithic kernel with driver modules compiled for that kernel version. So we’re doing no favors by tricking users into thinking that Linux can work like Windows with a familiar but skin-deep interface layer.
- There are however some very sensible paradigms that Windows has popularized or arguably even invented in some cases. The concept of a “Start menu” is a broad paradigm that really does work well in Linux too, whether the distro tries to make a pixel-perfect clone of the Windows presentation layer or not. The same goes for a system tray for applications that need to keep running in the background but not always be visible, and when Linux desktop environments needlessly remove that sensible paradigm they usually frustrate many users, whereas desktop environments that keep an implementation of a system tray are usually perceived as easier. And I would also argue that the paradigm of downloading a self-contained and archivable/portable executable software installer file with no dependencies is also a very sensible paradigm for many scenarios, which is why things like AppImages have taken off on Linux.
So I am in favor of the implementation of sensible Windows paradigms in Linux distros, without necessarily copying the exact presentation layer.
Also regarding @CubicleNate 's comment about ARM being a heterogeneous mess with no standards, 100% agreed. I remember in an interview where Linus Torvalds mentioned that what motivated him to create the Linux kernel was the sense of enablement when he got a PC with a i286 or i386 clone processor, which enabled him to code for a standard instruction set that would work on anything that supported that same architecture, regardless of brand or manufacturer. That is the total opposite of the ARM infrastructure, and it completely turns me off, because I have to do so much research before buying to see if the hardware will support my preferred distro and learn about all its idiosyncracies. Whereas I can buy pretty much any x86-64 machine and be almost guaranteed to at least be able to boot and install software on it even with the most obscure Linux distro out there.
I think this is quite accurate. To copy the Windows look closely can certainly raise some unsustainable expectations that the experience would be the same and when that .exe doesn’t give you the driver you expect, you make an inaccurate assessment about the quality or completeness of the operating system.
I sort of think the general familiarity of the Plasma desktop COULD lead to this expectation which is why I often think something like Unity or even GNOME is a better starting point for many but that is just an opinion. Plasma is my jam as I think it is the ultimate in desktop experience but that is obviously not an opinion shared by all.
Also, not a popular opinion, but I truly think that ARM is just a toy platform until some real standards emerge. I love my Raspberry Pi computers but I don’t view them as serious computers. Maybe the Apple M1 will change it, but I will stick with x86 for the foreseeable future for serious computing.