[gxonatano]

I know there's a place for distros like these, designed to be familiar to users of Windows or MacOS, but to me it shows Linux at its laziest: where exciting new ideas in system and UI design are skipped over, in favor of bad design ideas from 1995 (looking at you, Start menu). On MacOS and Windows you're stuck with whatever OS UI those respective corporations decide you get—the Apple menu, the Start menu, floating window management, and so on—and there's nothing inherently good about those paradigms; they mostly just exist for legacy reasons. On Linux, you have the freedom to customize everything, and to so it just seems sad that so much good development effort is going into building systems that value familiarity over innovation.

Put differently, I find it sad when user-friendliness is valued over user-centrism. Linux is full of software that is user-centric more than user-friendly: look at Vim, for instance, which is famously difficult to quit, yet is designed to be ergonomic and efficient in a way which puts the user first. The Vim philosophy (modal editing, ergonomic arrow keys, etc.) has even been extended to web browsers (Qutebrowser, for instance), and to window managers (i3, sway, etc.). These types of programs, in my opinion, are where Linux really shines.

Most people commenting here, however, describe this familiar/innovative or friendly/centric dichotomy in terms of user archetypes: "techie" and "normal" people. That feels unnecessarily essentialist, implying that "normal" people aren't curious enough to learn something unfamiliar, like a new style of user interface. But if we always assumed that, we'd never have had any innovative interfaces at all: mouse-driven desktop interfaces, smartphone touch screens, or any of it.

Of course, Linux distros are diverse enough to have something for everyone. I just think that conventional, familiar ones like this represent a missed opportunity.

[lproven]

> in favor of bad design ideas from 1995 (looking at you, Start menu)

For what it's worth, that's the point when your comment jumped the shark. I knew then that this was just a rant.

The Start menu was a _superb_ piece of design, as was Win95 in general. If nothing else, the existence proof of this is the sheer number of other desktops that imitate the design:

KDE; GNOME 1/2; MATE; Xfce; QNX Neutrino Photon; Inferno; OS/2 Warp 4; BeOS Tracker; Enlightenment; Moksha; XPde; Fvwm95; IceWM; JWM; Lumina; LXDE; LXQt; Cinnamon; GNOME Flashback; EDE; Budgie; UKUI; Deepin; Aura; FyneDesk.

I could probably find more, but 24 should do for now. Even combining forks, there are over 20.

You may not like it, and that's a legitimate view I am not arguing with, but billions of people use desktop interfaces modelled upon it, representing the combined work of thousands of developers, reimplementing it in dozens of languages.

[gxonatano]

> the existence proof of this is the sheer number of other desktops that imitate the design

That's where you're wrong. The desktop environments that imitate Win95 elements do it to provide something familiar for their users. The KDE team is not sitting around going, "you know what was designed really well? The Start Menu!" In fact, many of the desktop environments you mention (GNOME Flashback, Cinnamon) were a conservative reaction to the new GNOME 3 design which broke from the Windows aesthetic. The Wikipedia page for Cinnamon, for instance, says it aims to "follow traditional desktop metaphor conventions" and aims for a "gentle learning curve." They're explicitly choosing familiarity over innovation.

> The Start menu was a _superb_ piece of design

Not really. It achieves a reasonably clean look, but at the expense of excessively hierarchicalizing programs and documents. GNOME's Activities panel allows you to click "Activities" then click the program you want to run. Even better, you can just tap the Super key, type a letter or two of the program, and press enter. On Windows 95, I remember trying to launch a calculator, and clicking Start, then clicking Programs, then clicking Utilities, then clicking Calculator. In 1995, lots of people were complaining about the Start Menu, how clunky it was and how it slowed down common tasks. GNOME 3's approach is better, as is MacOS's Launchpad, as well as lots of other desktop launchers.

> billions of people use desktop interfaces modelled upon it, representing the combined work of thousands of developers, reimplementing it in dozens of languages.

The idea that pervasive ideas are somehow good, just because they're popular, is a well-known logical fallacy called Argumentum ad Populum. The Start Menu was never good. It was just popular. One does not follow from the other.

[jraph]

> Even better, you can just tap the Super key, type a letter or two of the program, and press enter.

Most DEs do that today, including KDE, Cinnamon, Xfce (at least in some config), and Windows itself (although that last one does to much including web searches when you do this).

The categories in the start menu have its advantages, that's how I discovered what program exists and does what as a child in a KDE based distro. I fail to see what's wrong with a start menu + this feature you are citing. It does discovery + efficience very well. Gnome 3's Android-like icon grid without categories (from what I recall and from the screenshots I see) seems awful for discoverability seems awful. If you don't know the icon or the name of the program you are looking, the icon grid seems awful (although I recognize keyword search should get you there, but keyword search + the categories that other DE provide seems more useful - and today, it's rare to have deep hierarchies like in win95)

[lproven]

I disagree on all points. I suspect that you lack historical context to this design, and may be too young to thoroughly grasp it. I have read that younger people (millennials and younger) tend not to think in hierarchies and find them complex and difficult.

Tell me, what pre-Windows 95 GUI designs are you familiar with? I don't mean know slightly, I mean know well.

[qn9n]

I agree that it's important for user-centrism to be a focal point of Linux, however I am also happy that distributions like this exist, yes you do lose some of the _magic_ of Linux by replicating a workflow a user is already used to, however this is perfect as non-technical people simply just don't care about that and just want to check their emails, social media, do some shopping, research stuff and watch some content online. They don't need to learn user-centric workflows like Vim to do such a thing.

[frollogaston]

As a longtime Mac user, I kinda wish it had a start menu. So much useful stuff easily accessible. But also, none of those choices affect my productivity all that much.

[qn9n]

As a longtime Mac user, just use Command-Space.

[lproven]

I'm both. (Mac and Windows since 1988, before Linux existed.) The point of the start menu is that you can search; the point of Spotlight is that the computer searches.

With Spotlight, you're telling the computer to run something you know is there, without bothering looking for it. You need to know it's there.

With a dedicated app launcher, such as say the macOS Launchpad, you can explore what apps are available to you. Once you know, you can quickly open it with cmd+space and 2-3 letters.

You can't open things that aren't there. You need to find what's available.

They are different tools for different purposes, which is why Launchpad is also there.

[frollogaston]

Exactly, except the problem with Launchpad is it's cumbersome and takes up the entire screen. I never use it, just have to know exactly what I'm looking for and use cmd-space.

[lproven]

> just have to know exactly what I'm looking for

And how do you acquire that knowledge? Browsing. Looking and reading, and remembering.

Mac OS X had no mechanism for this, but iOS does. iOS's Springboard launcher is lifted directly from the Dashboard in OS X "Tiger". Apple simplified it for the phone to only show apps. Then later they grafted it back in its simplified phone form -- Dashboard having been removed in the meantime.

Before that you had to browse the filesystem. To do that you need to know where to look.

That's how it worked on classic MacOS, and Windows 1 and 2, and DR-GEM, and AmigaOS, and RISC OS, and basically all other 1980s GUIs.

(Proprietary Unix left you with a terminal. Job done.)

The innovation in Windows 3 was having an app launcher program with groups. It was called Program Manager. It had groups, because it's quicker to look in the group related to what you want than in all apps. ProgMan was stolen from OS/2 1.1 by the way.

Win 95 had a further innovation that built on that. It shrank Program Manager down from a full-screen app to a single button, that opened up on a hierarchical list, and that list had icons in it because some people are more visual and recognise icons better than names.

Me, I'm a reader, I want words not pictures. Pictures waste my time and my screen space. That's why it's important to offer a choice. GNOME takes away choice. The GNOME devs have a Vision and you must use it. The KDE devs don't have a vision. They have nearly as many visions as developers, and they try to accommodate all of them.

Not everyone: just the devs. Examples:

* I use widescreens. We all use widescreens now. I want the title bars on the side, like in wm2, not on top. That's not an option.

* I liked BeOS. I think title bars should be tabbed, like in web browsers. That's not an option.

* I hate hamburger menus. I want menu bars. There is no global option for that. You can't have it.

* I hate CSD. I want a title bar I can middle-click to send behind all other windows, like KDE 1, 2, 3 and 4 did, as well as every other non-GNOME desktop. I also liked Windowblinds on classic MacOS: the ability to roll up windows into the title bar. Again, like in some older KDE versions. There's no option for that any more.

There is important choice, accommodating different needs and usage patterns, and there is cosmetic choice, merely affecting how things look but not the underlying mechanisms of how they work.

Supporting diversity of usage is more important than diversity of appearance.

Both the full desktops that natively support Wayland fail to do this.

[qn9n]

Realistically you shouldn't need thousands of applications so remembering the names of the 20 or so apps I use isn't a big deal.

[pjmlp]

The only UNIXes that I consider ever caring for the whole experience as a full stack, for users and application developers alike, were Irix, Sun NeWS, Solaris, NeXTSTEP and its evolution as OS X.

Sure you can argue Linux distributions can also offer something similar, the problem is which flavours and for how long, which brings us to shipping the Linux kernel underneath Java and Web frameworks, as being the most successful approach thus far.

[jraph]

> the problem is which flavours

Whatever distro which is pre-installed on the computer you buy, or whatever your geek acquaintance picks, or whatever you pick as a geek yourself (any mainstream distro will do).

> for how long

The lifetime of the computer in all of the cases, or, in the last case, until you want to try out something else.

Seems decent in any case.

[pjmlp]

Worked very well for netbooks, with OEM specific distros.

It was a vision on what something like Android would be.

[jraph]

Not sure what's your point, netbooks like the eee PC are starting to become ancient history anyway.

When it comes to our current era, Dell, Lenovo and HP computers sold with Linux are fine, and there's KDE-related hardware that seem nice too [1] (at KDE they understood that it's important to be the default OS, so they are pushing towards this). system76 too I've heard. Obviously choice is more limited than for Windows (although macOS is doing well with limited choice too), more biased towards pro, but there are decent options. The installed distros are quite standard too, there's some customization but not more than what we see on Windows computers.

[1] https://kde.org/fr/hardware/