Examining the Usability of Gnome, Unity and KDE

gbjbaanb writes "TechRadar has gathered a few users and subjected the 3 main Linux desktops to some usability testing for both experienced users and some new to the whole concept." I'm glad to see such ongoing comparisons; they encourage cross-pollination of the best ideas. On the other hand, it's a little bit like trying to determine the "best" dessert; even the most elaborate attempts to find statistical consensus won't answer the question of what's best for any particular user.

Comments

General usability should be one of the choices

[Anonymous Coward]

There's nothing wrong with ignoring the needs of individual users to tailor a generally good experience, _so long as power-users are still given the ability to pick the option best for them as individuals_. That last part is the important part that Apple has forgotten of late.

Re:General usability should be one of the choices

[obarthelemy]

Users can't fork Linux, they need something premade.
Further, users have computers skills by now, and have no desire to re-learn from scratch.

Re:General usability should be one of the choices

[sidthegeek]

Well, because they don't have to pay any money to obtain a copy of Linux.

Re:General usability should be one of the choices

[Coryoth]

There's nothing wrong with ignoring the needs of individual users to tailor a generally good experience, _so long as power-users are still given the ability to pick the option best for them as individuals_.

The standards for people calling themselves "power users" really seems to have dropped. The people I think of as power users would have no problem hacking together a nice custom FVWM2 configuration that integrates all the GNOME3 libraries (including the internal notification and messaging systems -- they do have nice exposed interfaces after all) and applications while giving them exactly the custom experience they desire. I mean GNOME3 is pretty damn modular and broken into a myriad of different libraries and components after all; it's just the shell that they've stuck on top as gloss that lacks some customidability. But no, these days people that call themselves "power users" seem to run scared at the mere mention of hand-writing their own FVWM2 or xmonad configuration from scratch; or indeed, of bothering to actually have to get their hands dirty to create a custom environment at all. Today "power users" need to be able to "customise" their environment via pretty GUIs and checkboxes. Heck, I've heard people calling themselves power users who called GConf complicated.

Look, there's still plenty of extensive customisation and configurability inherent in these systems, they just require you actually be a power user and know what the hell you're doing, and not be scared of getting your hands a little messy and stepping outside pretty candy coated "configuration" utilities.

Re:General usability should be one of the choices

[Nursie]

No.

To customise Gnome Shell you need to write javascript. I do not have the time or the inclination to write code to re-add functionality that was available with a right-click in the last release.

People who write custom FVWM2 configs in the way you talk about it were never power-users, they're obsessives.

Linux Format

[Anonymous Coward]

This is why I don't read Linux Format. The article provides no background on the users involved, and reads like something from a high school paper.

Configurability

[NoobixCube]

You know what's best for everyone? Configurability. When the developers can't decide how something should work, when they have two what seem like equally good or equally bad ideas, why should they force one particular decision on the user? Why not just put an option on a big scary controll panel somewhere made just for that? Of course, for usabilitiy's sake, there'd be the normal slick and easy to read control panel, but Gnome used to have GConf (does Gnome 3 have it? I don't know). You could use GConf to configure ANY aspect of the interface, anything at all. It was a very powerful tool if you knew what you were doing with it. So set the defaults to the lowest common denominator, to the grandma standard, but at least leave the powertools where we can reach them! Put up a warning that it may break the interface, sour the milk or bring the rapture to scare off the grandma users, and only those who really know what they're doing need concern themselves with it.

Re:Configurability

[obarthelemy]

Not really. OOBE is more important for the 95+% of users who are not hackers.

From the website that looks like this

[Superken7]

Techradar wants to talk and judge usability of the all-time favourite linux desktops, and yet their own website looks like THIS: http://i.imgur.com/IOyKu.png

I know other browsers render it centered, but that's not the (only) point, it's that their web looks awful: about 1/4 is margins, which is OK, and of those 3/4 1/4 is the content, which is split into 7 tiny sections (just give me the whole article and don't make me page every 3 paragraphs, it's almost 2012, for christ sakes!), tiny text, tiny images, and 3/4 of crap (related content, ads, menus, more related content, more related content).

It's not like they can't provide a very valid examination of linux desktops, but their site does not inspire very much credibility when they themselves get it so wrong, IMHO.

What "usability testing"?

[Animats]

Where's the "usability testing"? The linked article is just typical blogger blithering, spread over multiple pages for maximum ad insertion. They write "Since usability is a personal experience, we invited a bunch of people, from newbies to power users, to share their experiences with 3.2.". Which probably means "we asked for comments on a blog".

Real usability testing is not market research. It's measuring how well people did on tasks, not what they said they liked.

Re:What "usability testing"?

[Anonymous Coward]

And in the computer industry it's almost always done horribly. I've participated in such studies, and they mistake "ease of learning for a complete newbie" with "usability". They are not the same. You're only a newbie on some application for a few days or weeks, but you might be using it for the next 10 years. What makes a package *usable* is not something you can learn by watching me come up to speed on the damn thing for a few hours. Let me use it day in and day out and talk to me in 6 months. I'll have suggestions about whatever scriptability you have exposed, about keyboard shortcuts, about integration with this or that... none of which I will be able to tell you in your three hour usability focus session.

Re:What "usability testing"?

[swillden]

I've participated in such studies, and they mistake "ease of learning for a complete newbie" with "usability". They are not the same.

But neither are they completely different.

Usability encompasses not just newbie experience and not just expert experience, but the whole range.

Because of that, I found this particular article's conclusions very interesting: KDE has long held the position of most scriptable, most configurable, most integrable, etc. In short, it's been a power-user's desktop (well, out of the major options, anyway). Now it is apparently the most newbie-friendy desktop as well.

New user ?

[dargaud]

Where do you find somebody who's never used a computer nowadays ?!? Either it's a complete idiot who will NEVER be able to use a computer properly and hence we should NOT hear his opinion on the matter. Or it's something as rare as somebody who's never drunk a coke. Now that reminds me of a Coca-Cola ad campaign a decade or so ago... But they had to look.

Unity is one of 3 Main Linux Desktops?

[Glasswire]

KDE and Gnome obviously yes, but Unity is one of the top 3? Just because most recent Ubuntus foist this on users (and most feedback I've seen has been negative) - is there any data to show that Unity has even 10% of Linux desktops? While Ubuntu is popular, that just means it's bigger than any other. My totally unsubstantiated guess would be that Ubuntu is less than 30% of all desktop Linuxes installed and of that, not all are 11.x gen and many of those users have installed another desktop. So...
I would be SHOCKED if Unity is running on 5% of Linux desktops - does anyone have any hard evidence to counter this?
I wouldn't be surprised if Enlightenment or Fluxbox had bigger install base.
(I can't believe no one else has pointed this out)

Re:Determining the best turd

[bmo]

You can talk about many shortcomings of Windows, but they pretty much have usability figured out, right?

No, no they do not. And to talk to any typical Windows user, Windows is confusing to many users.

I can't wait for a fully-enabled Metro desktop to be unleashed by Microsoft. It caters to the absolute lowest common denominator of user. The rage it induces in power users is hilarious.

Why does Linux live in a separate world?

Because imitating the Windows UI is stupid and possibly lawsuit-bait. And I didn't go to Linux to just move to a cheap version of Windows.

--
BMO