Gnome 2.6 Usability Review

TuringTest writes ""The user-centric UI webzine" UserInstinct has published a usability overview of the latest version of the GNOME desktop. While their conclusions and recommendations are not mind-blowing, it includes two interesting appendices with a survey of new users (and their reactions to the system) and a list of common tasks of modern computer users with a commentary on how Gnome performs in each one. Note that usually You Only Need to Test With 5 Users (this report tests 4), you need to test additional users when an interface has several highly distinct groups of users and thus the conclusions in this review should not be taken as definitive."

Comments

Oh my god

[Anonymous Coward]

Why do Linux desktops try to mimic Windows so much?
Programers have no sense of aesthetics.

Re:Oh my god

[frankthechicken]

I feel it is mainly because in an open source world there is less of an ability to strictly test usability and functionality, as this requires a different set of abilities than purely that of programming expertise.

Indeed, when there are already a couple of tried and tested UI's on the market (i.e. Windows and OS-X). And with the money having already been spent on user testing the interfaces, the question is why re-invent the wheel?

Build on what has already been developed.

Once the underlying core has been built then experimentation can begin, with the many eyes approach hopefully leading to an even more intuitive OS and GUI.

Finally, a scientific review

[Iesus_Christus]

It's good to see a fairly unbaised, objective look at an open source product on Slashdot. Many times, we see either a "OMG Open Source = good" or a "I couldn't even get it installed, this sucks!" in place of an actual review. In this instance, it seems as though the reviewers actually tried to make this a fair review. They used users with different experience levels to get an overall picture of the usability of Gnome 2.6. While they could have used more than one user for each stereotype for statistical reasons, it seems at though they have done a decent job in their review. It is reviews like this that show us what to work on.

Re:Finally, a scientific review

[HFShadow]

It seems to me they were reviewing the tools more then gnome itself. If trillian, winamp or windows media player have confusing controls, I don't blame that on Windows, I blame it on the respective products. So why do so many linux review sites do the opposite? Totem being unstable has nothing to do with gnome 2.6. Rythmbox not having easy to find radio stations has nothing to do with gnome 2.6. The crappy file browsing window DOES have to do with gnome 2.6 and they apparently didn't even review deeply enough into it to see it.

Re:Project GoneME

[Cid Highwind]

I plan to create the outstanding *fixes* for correcting the buttonorder in the upcoming days

Since when does being an "expert to Unix" imply that you want your buttons in Microsoft order rather than Apple order?

Why emulate windows and not mac?

[io333]

Why isn't there an open source attempt to model what the folks in Cuppertino are doing? KDE and Gnome are both Windows copies. I think folks would switch in droves if they could get an open-source Mac copy to run on their PC hardware.

I can't think of any incentive for switching from an XP interface to one that is almost as good as XP.

Re:Why emulate windows and not mac?

[aixou]

This reminds me of when a friend proclaimed that he hated Linux (KDE to be specific), because it "feels like a shoddy copy of Windows"... It's funny for me to hear that from a Windows user, because that's how Mac users have felt for the past 10+ years in regard to Windows (shoddy copy).

Seriously though, I really don't believe an OSS project could have the focus or resources to take on the task of keeping up with Apple's design. They might get the "look" right, but the "feel" is something much much harder to grasp. In the end all you'd have would be a dock clone, and a clunky interface. The Mac OS has always thrived on having the best UI consistency and a very intuitive feel, which is something that Linux just can't really compete with.

Unix and Windows are much more similar to eachother than they are to the Mac OS. For one, both Win and *nix hide programs down deep in an arcane directory structure that you aren't expected to learn (you can, but most don't). On a Mac, you are expected to navigate the file system to access what you need (if you want to open an App for example, you open the Finder, go to the Applications folder and then open your app). Win and *nix don't really expect you to have to move to where you want ago, hence the usage of the Start menu and Shell PATHs which give quick access to what you need.

An OSS project that copied the Mac would really be copying the Finder and the directory structure. In order to get the Mac feel down pat, you'd have to make the directory structure much more browse-friendly than it is. You can't expect grandma to navigate to usr, bin, and then select from a long list of programs what she wants. Unless a developer/distro with some major clout (i.e. one that wouldn't be completely shunned by the Unix world) decided to revamp the directory structure (or hide the standard tree in favor of a simpler user oriented one), I would recommend that *nix Desktop Environment developers stick the Windows-esque start bar clones that they already have.

For now, if you want the Mac look and feel in a Unix environment, your only and best option is OS X.

Re:Why emulate windows and not mac?

[jdub!]

GNOME is generally closer to Mac OS than Windows, particularly given recent design decisions. We certainly DO NOT aim to clone either interface - we tend to think that the "not quite Windows" feeling has a strong, negative impact on training, worse than being obviously (and positively) different.

Your distributor may set their GNOME defaults up to be familiar to Windows users, however.

Re:Why emulate windows and not mac?

[poptones]

I find it funny when these "studies" say nonsense like "for open source to compete with windows it must handle these proprietary formats perfectly" - as if windows did that! How many times has YOUR browser been hosed by the bloated and buggy Acrobat? Anyone ever tried opening a powerpoint presentation in office97 that was created in office2k? Or for that matter opening a late model powerpoint presentation in microsoft's own powerpoint "reader" on a machine that doesn't have office?

Open source doesn't have to do anything "perfectly" - and none of the competition does this either. What open source has to be is available - that's all. Because, after all, this is all that made windows "the market leader." The one thing that made windows what it is with a wink and a nod is built into the competition. No wink and nod needed - and as people find themselves increasingly sick of paying 100 bucks to the local geek to "clean" the web residue from their computer every six months, more and more will try it - and many won't turn back.

I'm posting this now from a two hour old Suse kde desktop - just a quick note before I wipe it all out and go back (again) to mandrake. I just wanted to reply before all this gets lost to point out a couple of things.

Having a dock does not make a mac clone to be sure - but neither does having an "explorer" browser and a window frame make a windows clone. And the reason some of us (like myself) don't use the mac (among other reasons) is because it, like windows, is just too limited. If I want a dock I have a dock - but I can make it dance any step I choose and I can make my desktop look and feel any way I choose. This is much harder on the mac and damn near impossible in windows.

I don't care about a finder. I don't need a finder because when I install an app it's almost certain to be included in my path, which means all I really need to do to run it is invoke it by name. If I want a shortcut I make a shortcut and type the name of the app and whatever switches I want - so how is drilling down to the folder and clicking the app more intuitive? Just so I don't have to (shudder) TYPE something one time before using it?

This is coming from someone who has had his own long and reluctant journey. Even now I still have a win2k partition on this machine because I was hesitant to make the complete switch - old habits die hard. But with the arrival of fantastic distros like Mdk10 and Suse9 and FC2 using windows increasingly feels "icky" - like driving around in a 1985 Taurus or something. Yeah, it works most of the time and it's familiar because the seat fits my ass after ten years at that wheel - but I hate being seen in public at that wheel and every time I step out of it I have to change clothes and shower because of the dirt and grease.

Look: Grandma doesn't need a finder because Grandma doesn't have anything TO "find" except more arcade sites to occupy her time. Grandma doesn't install programs, gramma USES programs. Give grannie a desktop shortcut to a web browser that'll let her check her aol mail and play her pogo games and she'll be just fine. Give her all this in a computer that won't be brought to its knees by all the shit floating around those popup ads and grannie can die happy, a cigarette in one hand and her mouse in the other. Your only worry then will be how to clean up the rings left on the desk by her vodka tonics.

Re:Why emulate windows and not mac?

[aixou]

That sounds really unscalable- just as soon as you have more than a handful of "applications".

The "Applications" folder on my Powerbook contains 102 items, a very manageable and navigable amount imo.

A press of the tab button in the terminal reveals that I have 1602 command-line apps.

If I had to wade through a directory with 1704 (the sum of my command-line and /Applications apps) executable programs to find the GUI program I wanted to use, I would go insane.

The way Apple laid it out is very practical, and might be an important pattern to follow as more and more 3rd party developers ship GUI driven apps to the Linux platform.

Re:Am I the only one ...

[LostCluster]

In any menuing system, there is no rule that says that the menu has to be a true tree. There's nothing stopping you from having a main menu of "What are you here to do today?", and having the most popular functions being placed in more than one position on the interface.

Doing it that way might lead to a more confusing set of decisions at design time, but the user will more often than not find themselves one click away from what they want to do next if you do it right. Afterall, it's easier to find any given option if it's "hidden" in three places instead of just one.

Re:Project GoneME

[JanneM]

Of course, Gnome does not have "Cancel" and "OK" - if you find a dialogue that does, it is a bug, and should be reported.

Re:I was thinking of ditching XP...

[JanneM]

Are they both putting the same amount of effort in making their desktops user-friendly?

I would say that Gnome has had it as an explicit focus for a lot longer than KDE, and has been working a lot more on various aspects of usability. One example is the (always ongoing) effort to make the desktop fully accessible to people with disabilities - an effort that pays off for the rest of us as well, in the form of a more consistent desktop and some fun toys (like screen readers) to play with :)

As to which desktop is actually the better one for you - well, that's up to you, really. Try both for a time, and select the one you are more comfortable with. Or don't choose; alternate between both as the mood strikes you. Either desktop's applications work fine under both, after all, and interoperability between them is steadily improving.

What you absolutely should do is to ignore all the flamewars and sniping on places like /. - most people dissing one or the other desktop are pretty clueless fanboys that only embarrass the mature users and developers of their chosen desktop.

Be proud of who you are. Post as yourself.

[Ilan Volow]

Ali,

While I completeley disagree with your feelings about where the GNOME project has taken things (I think they should have gone much further and totally flipped off the unix geeks and shouldn't have blindly copied so many of microsoft's mistakes), I do respect you for your decision to fork, as the GNOME guys have been complete and utter jackasses about many things (such as usability, or lack thereof). I have had the same idea as you, albeit to fork GNOME in a completely opposite direction with the Clarux project and making GNOME far more mac-like. While I totally disagree with what you're doing, I'm glad at least someone had the same idea, even if it does run counter to mine.

One piece of advice to the opposition: the Free Software community says they promote freedom, but often, that's not the case. A while back, I created a fork of KDE that removed some really stupid usability problems the project had refused to deal with for years. I provided all my changes as source code people could download, I complied with the GPL, but Freshmeat refused to post the project because they considered it "only a patch". If you do something considered "significant" like modify someone else's code, it can be considered a distribution. But if you modify something that the Free Software community considers "insignificant", like the user experience, it's only considered "a patch". People in the Free Software development community might tell you "if you think you can do better, make your own version"; the thing is, they don't really mean it. So I'm warning you now, if you are really planning on forking a major desktop environment, you won't be able to rely on traditional community outlets for promoting it.

Last piece of advice--post as yourself. Stop this silly oGaLaxYo/Anonymous Coward crap. Post as Ali Agaa, be proud of your opinion, and be proud of what you're trying to stand up for and accomplish (even if it is rather silly).

Re:Be proud of who you are. Post as yourself.

[7-Vodka]

LOL dude,
Since when does Freshmeat = the open source community?
They are a 'unix program website' with every type of license included like proprietary stuff.
If anything you should have tried sourceforge first right?

Re:Be proud of who you are. Post as yourself.

[Anonymous Coward]

The interesting thing about the Open Source community is that they present themselves as a single entity when they talk about how they're united and taking over the world and how damned superior Open Source is.

Rubbish. However SOME people may present the open source community I have no doubt at all that you KNOW that it's a disparate bunch of incividuals and organisations. Obviously "they" haven't ever presented themselves as a single entity. Some individuals or groups within them may have done. If you want to rail against particular statements by OSI or FSF or Linux Torvalds or Slashdot ACs or whoever then go ahead but pretending that you thought that open source really was some sort of single minded collective is stupid.

Apart from anything else, you've identified yourself as part of this community, making amendments to open source software and distributing it. Did you think that you were subsumed into this single entity? Are you responsible for Freshmeat's policies? Are they responsible for yours? Try to get a grip on realirt.

Re:Be proud of who you are. Post as yourself.

[scrytch]

One piece of advice to the opposition: the Free Software community says they promote freedom, but often, that's not the case

Oh gosh, Freshmeat refused to list my variant, I'm being opressed. Look at poor poor me, oppressed by Freshmeat and the whole OSS community.

When Sourceforge, Savannah, and Berlios all reject your project, you might have a point. All you are doing right now is whining.

Need more than 5 users...

[Anonymous Coward]

Too many of the users were familiar with Microsoft software. Are they testing usability or similarity to their previous experience. If they are testing similarity then by having most users familiar with Microsoft, they will get a pre-ordained answer -- the more it looks like Microsoft the better.

Also, how about Chinese user, Hebrew user, Arabic user, to test language differences in the interface. How about blind user, disabled user? Even illiterate user? I am not joking. Check out the Simputer.

Awesome

[Nailer]

I want the Gnome guys to understand two things:

• It doesn't matter if spatial nautilus is good or bad. Overriding people's existing preferences without asking them is always bad.
  
• 
  If you're being paid to work on Gnome for a business, do more of what that businesses clients want. Not less. Currently, most Red Hat staff/customers (who aren't Gnome developers) don't like spatial nautilus. A challenge for the Gnome devs is to either convince those people otherwise (by making it better or explaining its usefullness, not overriding people preferences). If they can't, it shouldn't be the default for EL4.
  

I choose to value post count less than you

[tepples]

You can take their review as gospel all you want, I'll continue to read reviews from people that have withstood the test of time and had a bit more scrutiny.

You can choose your reviews based on ad hominem criteria involving post counts; I'll choose my reviews based on the abstract and then weigh each conclusion based on the evidence that the review presents.

Look at it THIS way.

[khasim]

The Linux kernel and the GNOME desktop did NOT GET IN THE WAY of the applications for those users.

You're correct about too much time being spent on the applications. But that's how most users operate. They spend the MINIMUM time possible interacting with the desktop and the MAXIMUM time interacting with the applications. (Aside from playing with backgrounds and sounds. I hate webshots.)

Personally, I think that a tiny bit of work on that study and the NEXT study would show Linux being incredibly easy to use even for novices.

#1. Get rid of the unstable apps. Each icon that they click on MUST launch an application and that application MUST be the most stable of the bunch.

#2. Populate the desktop with the apps they'll be trying to find (nothing like making it easy for them). This is what I do at work. And remove any other icons. They can put other ones there when they are more comfortable with the system.

#3. Put the controls for changing the background and the sounds in a very visible location and name them something like "Cool effects". Then give them lots of pictures and sounds to choose from.

So, the desktop would have the "My Computer" (or whatever) icon.

The "My Network Places" (or whatever) icon.

The "Recycle Bin" (or whatever) icon.

The "Work applications" folder/link icon.

The "Cool effects" folder/link icon.

The "Games" folder/link icon.

The "Help" icon (context searchable, etc).

Also, once you've run through with each of the testers the first time, have them form small groups and run through the test again. In the workplace, they will talk to each other and share tips/hints/ways to install spyware crap/etc.

Does the desktop facilitate or hinder that kind of human interaction?

And toss in a screensave as a background option just to give them something that Windows doesn't do. :)

That button order is a Windows artifact

[Anonymous Coward]

Did you really mean Unix experts, or did you mean Windows experts?

Do you have some justification to offer for your changes other than, "Windows does it that way?" And what does the way Windows does thing have to do with so-called Unix experts.

I completely fail to understand what the big deal is with the Gnome button order. I have been a Windows user since Windows 3.1 and a computer user since many years before that.

I have used Gnome. Until someone mentioned it on Slashdot, it never occurred to me that it was somehow difficult to understand the button order. In fact, the "button order" as an object of any thought or criticism had never occurred to me.

Frankly: what is the problem here? Do you operate your computer in a dark, smokey room while wearing sunglasses? Or is your Gnome installation suffering from a misconfiguration issue such that buttonfaces are flat blank and you must guess as to their function?

If you can make Gnome better, then by all means do so. But I'm just having a difficult time conceiving of how it might come to be that this is an issue for someone. Utterly mad.

Re:I was thinking of ditching XP...

[xutopia]

Your question is perfectly valid. I enjoy generalizing things and I hope that this is a good way to express things.

If you are an XP user you'll be accustomed to many Windows placements and ideas that have been carried over to KDE (start menu, placement of things in general). However KDE is cluttered and there is no central document that KDE developers can refer to in order to achieve proper usability. So they usually end up having applications with a lot of clutter. If you are familiar with usability tenets you'll know that giving users too many choices is the same as giving them none at all. Every option you give a user is a choice he has to make. Sometimes he doesn't want to make a choice, he just wants to get stuff done.

Gnome has the Human Interface Guideline'. The document is a comprehensive guide which can be used to do "higification" on an application. I find that most Gnome applications tend to follow the HIG and the result is a more consistent desktop. It takes some getting used to for Windows users though because things aren't "in their place". However once you spent a week using Gnome you find it hard going back to Windows. Many say the same about KDE but my personal opinion is that Gnome is despite being different than XP and will need a bit more getting used is the more usable Linux desktop around.

For anyone moving from Windows to Linux I recommend you arm yourself with patience and a good friend that can coach you through the change.

Re:Will gnome die?

[Anonymous Coward]

What are you talking about? Seriously.

If GNOME 2.6 is slow, take it up with your distribution.

I've been running GNOME since before 1.4, and I find the speed of 2.6 fine on my 850mhz pIII (320mb RAM). Additionally, I find 2.6 the fastest of the 2.x releases (but I eschew gnome-terminal in favor of xterm because it is quite slow due to pango).

As far as shortcuts and menus, I'm not sure what you're talking about. You do know about applications:/// right?

Anyways, enjoy KDE. I've tried it on and off over the years, and I'd rather not use it.

Er, and as far as people funding the project dedicated to other things, are you talking about Novell? They're only a small part of the people currently contributing, and they're more interested in mono. What about Red Hat and Sun who both base their desktop on GNOME? I'd say they certainly have vested interest.

Keywords: Spatial and File Select

[oddbudman]

Its lunch time so i thought I would throw my hat in the ring.

I have been using Gnome 2.6 for i guess a few months now and I have to say it is excellent. I used to use fluxbox (a great WM) but to be honest I haven't really looked back.

I use it nightly as a desktop workstation. I do everything on it from developing firmware for Atmel micros, GUIs in GTK2, web browsing, warez downloading and playing enemy territory.

Gnome 2.6 is faily well intergrated these days. Generally a right click on something will bring up options with what you can do, Left click selects - its nice and predictable. Ctrl-C Ctrl-X and Ctrl-V work as they do in Windows - very cool.

Straight after installation I could do drag and drop burning (good for making an mp3 cds for my car). Another thing that has me impressed are all the cool system tray apps that come with it. They are easy to add and handy too. Right now I have one for net, one for cpu and one for local weather. The Local weather one is awesome as I can always have my finger on the pulse.

I can't say I like spatial browsing though. Not default at least. Personally I found it really frustrating. Its not like i didn't give it a go either - I had it in spatial mode for about a month. It spreads like a cancer across your workspace. Before you know it you have waaaaaay too many windows open. Hiding the address bar is pretty stupid too imho - it makes it really easy to get lost and confused (especially when spatial mode decides you need 3 windows open to traverse 3 directorys). Perhaps if spatial mode didn't open a new window each time (or swapped middle and left click functions), and showed where you were I wouldn't mind it as much. My problem with spatial browsing was solved when i turned it off :)

Another dislike is definitely the file select dialog. Who makes a file select dialog where entering the text yourself is not an option? Would it really have thrown the file select dialog into chaos if it was included? Why make it so it is completely unintuitive for a computer user who has been using Windblows for years? Now a file select dialog with text entry and typeahead search on the files in that directory would be great default behaviour. (please don't tell me about the hotkey either - that is not intuitive)

Generally though I think Gnome 2.6 is pretty awesome. It is the best Linux DE I have ever used and I will continue to use it. It is definitely a step foward for the Linux Desktop.

Re:I was thinking of ditching XP...

[bogie]

" However KDE is cluttered and there is no central document that KDE developers can refer to in order to achieve proper usability. "

Let's be fair here. How many Gnome/GTK apps have their own File dialogs? Regardless or Gnome's HIG's, KDE's apps all tend to look and act like they were designed for one another. This is opposed to many of the marquee what I call "adopted" Gnome apps that don't integrate well with Gnome at all. Compare that with how Konq, Koffice, on Kontact work within KDE. KDE does in fact have more options, or what your calling "clutter" but overall its still more consistant then Gnome although that's finally started to change. Gnome also has the habit of making major interface changes between versions. Dialog boxes have been switched around and now the most fudemental way in which you interact with your OS ie file browsing has been turned on its head with Spatial. No warnings, no this is optional in 2.6 but will be switched on in 2.8 etc. KDE on the other hand has stayed consistant in the basic ways that matter for years now. The HIG's the Gnome devs go by are nice but they aren't some magic beans which have fixed all of Gnome's issues. Users from XP should try both and see what they like more, but I'll also say that IMHO KDE continues to be the more polished better managed project of the two.

HIG certification

[0x0d0a]

Currently, the HIG (see above) is made available to developers for voluntary adherence. If more resources were made availCurrently, the HIG (see above) is made available to developers for voluntary adherence. If more resources were made available, the GNOME project could start a certification program to document compliance with the standard. This would allow users to seek out certified applications and know that these applications would integrate well with their existing desktops.able, the GNOME project could start a certification program to document compliance with the standard. This would allow users to seek out certified applications and know that these applications would integrate well with their existing desktops.

Horrible idea. None of the good desktop interfaces out there have *ever* required certification. We know that it is not necessary to produce an easy-to-use desktop. Further, this will discriminate against those people that do not have money to pay certification fees, slow development of applications (as individual versions would have to each be certified), and slow evolution of the HIG itself. I am opposed, and think that any attempt to formalize a certification process as part of GNOME would simply lead to bad feelings, loss of good will for GNOME, and project fragmentation.

Unfortunately, those who were not were just as quickly lost and confused. To maintain the abstraction, we recommend that it be removed from the view of the new user and kept in the application menu.

There were a number of suggestions like these -- hiding advanced functionality. While this is a reasonable approach -- the terminal is still in the applications menu, and easily available and easily found by non-novice users -- it is also extremely important not to work too hard to hide functionality. One of the largest problems with GNOME 2.x (IMHO, of course) is that significant and valuable functionality has been hidden or deprecated in the name of more basic "easy to use" features. This includes two of my favorite pet peeves:

* Viewport support (someone apparently decided that it was "confusing" to allow the user to have a window partly on one viewport and partly on another, so it was replaced with a number of virtual desktops). As a result of this technical decision, sawfish (which is not the newbie-recomended GNOME WM in any event) underwent significant negative technical change.

* User-rebindable accelerators. In GNOME 1, unlike every other GUI that I know of, accelerator keys attached to menu items can be simply and easily rebound by highlighting a menu item and tapping the desired key combination. This is a phenomenally powerful feature that demonstrated that the OSS world really *does* enjoy new ideas and significantly improved the GNOME user experience. It meant, for the first time, that the user was not bound by the decisions of the application developer. KDE has a similar-but-not-identical feature that allows *some* menu item accelerators to be globally rebound (frankly, I'd like to see the synthesis of these two featurs). Anyway, some usability person decided that this could be confusing to a new user (fine, I'll buy that) and the solution presented was to entirely disable this feature and requires manually adding a line to a text file on a per-user basis, instead of simply providing a toggle button in an "Advanced..." dialog or something similar. As a result, few users know about or take advantage of this functionality.

A remedy is needed for this situation. The answer could be an installation application that can speak to all of the popular distributions. It could be built in such a modular way as to allow new backends and functionality.

This is a good idea, and should have been done a while ago. It's a bit disheartening to think that this will likely have a very limited subset of functionality and be used by most users, though.

A solution to this problem that allows for applications to be downloaded from webpages an

Re:Project GoneME

[zsau]

The Gnome buttonorder isn't broken, so it can't be fixed. It's just different.

BTW, can I chose my own buttonorder, or are you just imposing another decision on me? I might find some other aspects of GnoME interesting, but if the only option is Windows buttonorder or old-style Nautilus, you haven't 'fixed' anything. You've just imposed your own preferences.

Recommendations (and UI abstraction)

[arrianus]

The recommendations don't follow from the user experience. Several don't necessarily follow, but I'll present the dominant problem here:

Maintain the Abstraction from the Underlying System

This is almost universally the wrong approach -- this is what Windows tries to do, and MacOS avoids. The key is to make the underlying system simple, and make the UI reflect that.

The problem with abstracting the two is that it leads to bit rot. At some point, the Windows registry will think a file is in one location, whereas the file is actually in another. Or the UI will misunderstand the way that the 5 options in the configuration file should be presented as 2 options in the UI. Or there will be some underlying binary configuration file, with some option that's not available in the GUI, that somehow gets flipped, breaking the whole system.

You very strongly don't want an abstraction. On the Mac, installing an application (I haven't use OSX, so my knowledge is based on the older versions) is as easy as dragging the folder onto the hard drive. To erase, you wipe it. On Windows, if you wipe an app like that, it'll leave bits and pieces of itself scattered throughout the registry, links in menus, DLLs in system folders, and dozens of other places. Worse, the uninstaller will often no longer work. Most clueless users, if they try to erase an application the wrong way, will end up with a semibroken system, since there are different levels of abstraction that do not maintain consistency.

The whole Windows (and increasingly, GNU/Linux) approach of abstracting out underlying complexity is flawed. The trick is to eliminate the underlying complexity, and have a single set of simple structures that the GUI tools (or the users manually) operate on.

When I first used Red Hat in '96, the types of issues that threw me were: I wanted to change the login text. I grepped for the old login text, found /etc/issues, and I edited it. It worked. I rebooted. It went away. It took me the longest time to figure out that an fucking script in /etc/rc.something/ was overwriting it. The information was stored in two places. I overwrote the wrong one, and boom. I was fucked. I stopped using Red Hat precisely because of the complex configuration scripts, which made the system fragile and ultimately, easy to break and difficult to use.

This shows up a huge number of places -- especially in a heterogenous environment like GNU/Linux, you often have multiple configuration tools. I can download a half dozen Apache configuration tools. Very often, if you run one, then switch to another, the thing no longer works, since they edit different options in different ways.

One way to implement this (presented in an oversimplified fashion) is to first design what you want the UI to look like. Once you know, you design the underlying structure to match. This is the opposite of what most GNU/Linux and Windows developers do, where they try to engineer the most flexible underlying structure possible, and then develop a UI on top of that. This doesn't necessarily lead to less flexible underlying structures -- it's just that to have a good UI, you want to give some thought to the user experience when designing the engine, and especially, the configuration files.

Re:Bah... vote with your feet, then

[Nailer]

most behaviors are easily modified there if you're as hardcore as you proclaim to be.

Where do I claim to be hardcore?

And if you don't like GNOME's design goals, then fine

Its not a matter of my own opinion - its a matter of our clients. I work for Red Hat in Australia and I can attest most of our clients don't like spatial mode. I'd like to have it either explained better (welcome to Gnome 2.6! We've included a new spatial mode! Its better for X reason! If you don't like it though, do Y!) or changed by default.

As one can imagine, my opinions are my own and don't necessarily represent my employers.

Example

[Anonymous Coward]

A good example of this approach is the Linux kernel configuration system. Here, the configuration information is stored in a single, standard format. It describes what the configuration options mean, the dependencies, etc. Then, there can be a number of "dumb" configuration tools (make config/menuconfig/xconfig/etc.) that all work on the same underlying format, entirely consistently, and in addition, users can modify it by hand. This single file is used for everything in the build process. It is impossible for the file to be inconsistent, as it is handled.

A bad way to do this would be to have a "Kernel Build Wizard," that was hand-coded for a dozen kernel options, and had its own expert system to set the remainder automatically, and that had it's own higher-level config file that built into a kernel config file. (The right way to do this would be to push the automatic information down into the kernel configuration language).

Linus is brilliant, and always does things the Right way. Now if only the major distributors were as smart...

Frustration with "usability studies"

[ky11x]

I'm beginning to get very frustrated by these usability studies because they all tend to make the same false assumption that "familiarity for new users" == "usability for all users."

This is simply NOT true. Usability is a complex quality, and it is the result of compromises among often conflicting goals such as discoverability of options, reduction of keystrokes/clicks for common tasks, customizability where common base cannot be established, compatibility with competing interfaces, humaneness of interface after long-usage, accessibility, internationalization, etc. etc. How quickly New Users can discover and perform tasks is only one dimension of the usability scale, and one that's not even all that important except in a setting like public access kiosks or Internet cafes.

Different OSes approach this problem differently, and where as Mac OS X has chosen to compromise all the goals with an emphasis on discoverability and tolerance after long usage, Windows has chosen to place a different emphasis on sacrificing flexibility for complex tasks in favor of making simple, repetitive tasks easy to accomplish. Both approaches have their advantages and disadvantages.

The traditional strength of the various Linux desktop systems has been flexibility and customizability, with less emphasis on the other issues. I'm not suggesting that the needs of new users whose primary OS is not Linux in settings like Kiosks and labs should not be taken into consideration, but it should not be the ONLY consideration.

Usability studies like this one emphasize the needs of new users with Linux as a secondary OS over everything else. Take this as an example:

These extra areas (the desktop-reveal button, the workspace switcher, the file manger icon, the terminal icon, and the running application button in the top right) could be removed by default...

This is the sort of recommendation that makes sense for kiosk machines (simplify the UI as much as possible and go for task-orientation), but it doesn't make sense for long-term usability. Removal of these features means that users will have to discover them and add them back in, and that plays into one of the weaknesses of the current Linux desktops: discoverability is relatively poor. This is a very shortsighted and pointless recommendation for a desktop system that is also meant to be used as a primary desktop system for many home users.

I wish usability studies would really think about what usability is, over all and long-term, rather than just "can new users in a hurry get an email written?"

Other bugs

[Anonymous Coward]

It's not clear that the task-oriented approach is better -- my experience has been that most users don't want to do the recommended tasks, or want 10% functionality that's not in the recommended tasks, and so need to learn more of the system either way. Few people sit down at a computer once, use it for 15 minutes, and never use it again. Most people use a computer at least a few times a week, and it's important that those people can get things done. The addition of a "compose e-mail" task to the desktop buys very little in practice.

Also, the breakdown into "new user" and "experienced user" is a little too broad. As an "experienced user," having to figure out how to reconfigure a dumbed-down GUI usually takes a few hours, and is often not worth it. I also don't automatically know where things in GUIs are located. When I first sat down at this Sun (running some CDE hacked up for ease-of-use), it took me about 15 minutes to find a terminal. It sucked. An intermediate user shouldn't have to spend half an hour figuring out common "intermediate user" tasks to make the system simpler for a new user, nor should they be assumed to be an "advanced user with a surplus of time" who can or wants to take the time to reconfigure the system, because some asshole decided new users would be confused by having the terminal in any ready-accessible menu, and hid it 5 layers down. There really needs to be a smooth gradient that balances the needs of all users, not just new ones. This should include very advanced users who don't have time
to spend reconfiguring every machine they sit at.

One way to solve this would be to have Gnome, upon first startup, be able to select one of several default suites ("Expert" "Intermediate" "Beginner", as well as some level of screen realstate used by Gnome vs. proximity of features).

Re:Why Bother?

[kundor]

Have you ever tried programming KDE?

It's sooo nice. Really. You can whip up a full web browser in about 5 minutes thanks to khtml.

Re:Project GoneME

[sploxx]

Hi!
GOOD idea! Really. I'm using gnome right now, but the "remove-every-option-which-maybe-hard-to-understan d-for-some-people"
attitude which some of the more influental gnome people start to have really pushes me back to KDE.
Of course, this is only my opinion. But apparently I'm not the only one here...

But I think you're doing one thing wrong. Instead
of changing button orders permanently by a patch, better make such things configurable! That is the
thing I'm missing most in the newer versions of gnome, removed flexibility because "it is better for you". Make an option for the button order. Give the user the power back to change window managers. Etc.pp.

Maybe this is the least common denominator for all gnome forks? Other options/other configurations?
That'd be nice because you could cooperate with
the other forks around and, for features you agree with each other, have a productivity boost :)

PS. And mods. The above poster does an advertisement. For a free software project. But that is clear and not hidden. He doesn't bash gnome, he doesn't bash KDE, nor does he belong to the GNAA.
IMHO, the parent is in _no_ way "flamebait"!