Make. It. Simple. Linux Desktop Usability — Part 5

OMG! Ubuntu! broke the news today that GNOME 3.28 is going to remove icons from the desktop, turning the desktop into a useless mere wallpaper. In my series about #LinuxUsability on the desktop (part 1, part 2 ,part 3, and part 4) I have occasionally voiced the suspicion that there is a secret plan to kill the desktop metaphor, and this is the latest proof. The crippling of the desktop has to stop!

Desktop icons. Source: OMG! Ubuntu!

OMG! Ubuntu! nails it:

The ability to plaster our desktop background with an ungodly amount of app shortcuts, files, folders and other clutter is, for many of us, part and parcel of using a computer. It’s a feature that most major operating systems have in common.

Indeed. Icons are what the desktop was made for and have been an integral part of desktops since the desktop metaphor was invented at Xerox Parc in the late 70s to early 80s:

Xerox Star user interface (1981). Source: GUI Gallery

Developer Ernestas Kulik is cited on OMG! Ubuntu! with the following words:

“[The] desktop is blocking us and giving deep issues to further go forward with no direct benefit. Users have expectations for it to work decently and it’s not the case.”

No direct benefit?

Deep Issues? I have to admit, I had deep issues with GNOME in the past, but desktop icons where never one of them, so far. (Besides them being deactivated by default, that is.)

Dear GNOME, even thinking about removing the desktop icons from the desktop is like thinking about removing the steering wheel from a car. Sure it’s thinkable, but no, it’s not something that can “just” be done from today to tomorrow. It requires something that is better than the thing that is removed. It requires customers to be ready for the change. And, let’s face it, once it will be done the “car” will be a totally different beast from the car as we know it.

Dear GNOME, you need to decide: Do you really want to be a desktop anymore? Or do you want to be something like a “smartphone UX for large screens”?

Dear GNOME, how about you fix the desktop rather than continue to cripple it? GNOME once really had strong points, and Nautilus certainly was one of them.

GNOME and Nautilus when it still was a solid desktop. Notice the icons on the desktop. Source: http://toastytech.com

I mean, look at who originally developed it:

Eazel was a software company operating from 1999 to 2001 and based in Mountain View, California. The company created the Nautilus file manager for the GNOME desktop environment on Linux (…)
Mike Boich was CEO, having been a major figure at Apple; Bud Tribble was VP of Engineering, having been software manager and designer of the original Macintosh project; Andy Hertzfeld was a principal designer, having been a lead software engineer and designer of the original Macintosh project; Darin Adler led development, having been the technical lead for System 7 for the Macintosh; and Susan Kare designed new vector graphics-based iconography, having designed the original Macintosh icons. Other notable staff included programmer Maciej Stachowiak, who was a programmer and board member for GNOME; and board member Michael Homer, formerly of Apple, AOL, and Netscape.

Source: Wikipedia

Google these people and their previous experience. Each of them. Seriously! You will find some names covered earlier in this series (part 1, part 2 ,part 3, and part 4).

Since then, the GNOME desktop has been degraded in fundamental ways. The increasingly polished-looking design (as in: fonts, colors, buttons) can’t hide it. To the point where you, dear GNOME, are now considering to remove icons from the desktop.

Workarounds have been built like extensions.gnome.org which some have praised as a solution to work around the issues GNOME has in its default configuration. But let’s face it, good design can never be replaced by extensions. It needs to be at the core.

Linus Torvalds wrote in 2012:

End result: extensions.gnome.org may be a really cool idea, but it seems to have some serious usability problems in practice. And the whole gnome3 approach of “by default we don’t give you even the most basic tools to fix things, but you can hack around things with unofficial extensions” seems to be a total UX failure.

I think he pretty much nailed it.

And no, it’s not just GNOME. I left KDE after the switch from 3 (which was a great desktop) to 4 (which attempted to be an OS for smart touch tables or whatever) when they introduced all the Plasma stuff where you could place widgets your desktop and could rotate them in all directions, but basic documents and folders were much more difficult to have on the desktop all of a sudden. (Luckily, KDE has reluctantly been reversing this trend lately.)

I have to admit: Looking at the screenshots of classic desktops and comparing to what we have today, I cannot help but feel sad. The old machines were many orders of magnitudes less powerful yet they performed much better in terms of usability, ease of use, consistency, and look-and-feel. So sad.

Modern desktops

  • Have crippled menus (replaced them with harder to use hamburger menus — yes I am looking at you, Firefox, GNOME, elementary OS)
  • Have crippled global menu bars (replaced them with useless empty space instead of menu commands — yes I am looking at you, GNOME)
  • Have crippled scroll bars where you need to scroll in order to find out whether there is something to scroll (yes I am looking at you, Apple)
  • Have crippled desktop icons (yes I am looking at you, GNOME)
  • Have complicated things (software “installation” that used to be a simple drag-and-drop operation now needs App Stores and passwords and logins and Apple IDs and registrations and Terms of Service and certificates — yes I am looking at you, Apple)

The list goes on and on.

Let’s make the desktop great again! Can we get some UX people and developers together and create a simple but powerful desktop for Linux that is not ashamed to be a desktop, where every pixel just feels “right”?

Next: Design is how it works. Continnue to part 6


probono is the founder and lead developer of the AppImage project.

This article is part of a series (part 1, part 2, part 3, part 4, part 5, part 6) about #LinuxUsability.