Slashdot Mirror
How Configurable Should a Desktop User Interface be?
The Original Yama writes "In the world of user interface design there are two main schools of thought. The former maintains that the environment must be flexible and configurable enough to adjust to a user's needs. The latter takes the opposite perspective, arguing that many of today's user interfaces have become bloated and overloaded with features, and consequently have become difficult to maintain and use. KDE developer Mosfet shows how the KDE Project has managed to bridge the gap between the 'highly configurable' and 'less is more' camps."
If you use Windows and Stardock Windowblinds, you quickly realize that there are very few 'Skins' that allow you to use Windows as efficiently as the default appearance. The same is true of Windows XP themes, Winamp 2 Skins and Winamp 3/Wasabi skins. Even if they are beautiful, they have to have enough visual similarity to the original, or else the user ends up feeling that ha has to re-learn the interface.
The next Slashdot story will be ready soon, but subscribers can beat the rush and slashdot the links early!
Modular is a good idea, but you run the risk of overwhelming the user with too many possibilities, or having to include the modules in an awkward way - which has been a problem with the repacement windows shell LiteStep, and it's text based configuation.
These sorts of things are fine for advanced users, but your average Joe will balk when attempting to edit even the most simple of text configuration files.
Much experimentation will be needed for this.
It lives up to it's name: http://www.sanspoint.com
If the UI is completely configurable, eg, wrt mouse button function, how can I possibly document the functionality of my program? I can't say: Ctrl-Left-Click on this picture for foo-functionality, Right-Click for bar, Middle-Click for baz, because someone may have configured the UI so that the middle button closes the window, right simulates a double-left-click and ctrl-left might prompt them to save a copy of the picture in question.
Although some may find this a convenience I have found that most beginner to intermediate users find this a nuisance.
The problem is that when choosing a command like Tools - Options, the user expects the Options command to appear at the bottom of the Tools menu where it traditionally lies. It is the relative location of the menu entry that is significant, not the actual menu text.
The same applies for dialog boxes. After you have used an interface for a period of time, you eventually get to the point where you can place the mouse cursor at the position on the screen where the control you will click will be, even before the dialog appears.
Lots of irrelevent menu choices is a definitely a bad thing (*cough*kde*cough*) but randomly moving menu entries isn't necessarily a good thing.
If you don't want to repeat the past, stop living in it.
I'm a long-time Gnome user and was quite dubious when I heard that Gnome 2.0 was going to simplify the configuration operations. Having used various 2.x versions for some size months I've realised that there is nothing I've missed. I've come to the conclusion that while there are a lot of useful configuration options you could put in there are also a lot that are useless or could be simplified (but seemed so cool at the time). I'm generally for configurability, but sometimes a good purge helps.
I can't take anything Mosfet says seriously after his inane rant against RedHat's "crippled" version of KDE. The article linked to does not improve my opinion of him. "I think KDE is better than GNOME." Thank you for your unbiased opinion, Mr. KDE Developer.
KDE has lots of configurability and hides some of it to not scare people. Is this really worthy of an article? Does he really think he's found some new principle of design that he needs to share with everyone?
I rate his article:
0, Stupidly Obvious
-Erwos
Plausible conjecture should not be misrepresented as proof positive.
--sex
Very popular slashdot journal for adul
The problem on most open systems is inconsistency, because there are many different interfaces. For example, on my Linux system, I've noticed almost half a dozen different file open dialogs. Any one of them would be OK (sure...some are better than others, but I could get used to any of them), but having all of them is a problem.
This situation arises because the user interface decisions are made by the developers of each individual application (either directly, or indirectly by which X toolkit they choose to use).
Here's what I want. I want to decide which file dialog I like, which scroll bar behaviour I like, what dohickeys are on the titles bars of windows, etc., and I want every program on the system to obey that decision when I use it, unless I specifically ask or give permission for that program to do something else.
Personally, I want to be able to hack the living daylights out of everything in my environment. I dropped Gnome 2 because it started to feel like Windows. It reminds me of the early days of the phone company: you can have any color phone you want, as long as its black.
One of the major reasons the Windows desktop sucks, is that the programmers are forced to pander to some mythical vision of what "users" want that is the direct product of marketing and usability studies. Linux software is created by programmers who actually intend to use the software themselves, and it's better stuff all around. I am mystified by the attempt to adopt a process that has resulted in a car crash of a UI in the name of "making Linux mainstream."
Who needs it.
This sounds like a simple but it's a very important point. Configurability can be powerful but it totally screws up documentation.
It also makes it so that users from one system to another don't know how to use the UI.
Now, you can have high configurability, but in most cases THE DEFAULT configuration should be solid and easily available.
Simple configuration (like desktop colors, fonts, etc) should be easily available, but more advanced features should be kept for power users. And there should always be a quick way to "use the defaults" (like login dialog with a use default window mgr settings option).
- sigs are for wimps.
The UI should be appropriate to the task. Some tasks will benefit greatly from flexibility, some will not.
AutoCAD, for example, provides a pretty busy interface that regularly overwhelms beginners. However, it is also a highly configurable combination of command line, toolbars, menus, and context menus, with both scripting and programming facilities built in. When a moderately advanced user starts tinkering with it, the interface can be customized to provide enormous productivity gains for that user. (On the other hand, woe to the drafter who sits down in front of someone else's custom set-up.) It's very complex, and when I was a drafter I learned to love it.
However, as a drafter I was doing very repetitive tasks... I had small tools that saved me a few keystrokes and big ones that saved me hours. If I hadn't been doing highly repetitive tasks which were subject to some optimization, all that interface customizability would have been for naught.
Does an MP3 player need much beyond simple controls and playlists? I don't think so. Much of the customization in a program that performs a simple task will amount to eye candy... useless, but fun.
Simple tasks generally call for simple interfaces, while complex (and especially repetitive) tasks call for major customizability.
With reasonable men I will reason; with humane men I will plead; but to tyrants I will give no quarter. -- William Lloyd
Personally, I found NeXTSTEP both easier to use and generally more powerful than MacOS X, and that was on a 25Mhz 68040. It really came into its own on a Pentium Pro or Pentium II with a (then) good 2D card like a Matrox Millenium.
"Efficiency" and "speed" are the result of "thinking" about the "system". You can build fast-and-pretty systems just as you can build slow-and-ugly systems.
As mosfet writes:
Check this out to see a developers take on the whole "less-is-more" debate going on about Linux user interfaces
Now, I've no doubt he's a very gifted developer. But more often than not, a developers opinion on UI issues should be disregarded.
It's not because they couldn't potentially be good at it, it's because their brain is occupied with technical issues that have no relevance to an end-user.
Alan Cooper has written some pretty fine books about these issues, which I'm sure any developer related to UI design finds very informative. Some of the anecdotes are hilarious.
Unfortunately, it seems that Human Interaction Design is still not very high on the list when people design a product, resulting in there not being very many people that have specialized themselves in this field. It goes without saying that finding the right people for such a job for open source projects is tricky, to say the least.
This article doesn't show anything of the sort. Having read the whole thing, it's just another article arguing that more configurability in KDE is better. There's nothing wrong with arguing this, but it certainly doesn't (successfully) explain how any gap is being bridged.
If anything, it again fails to solve the main problem that highly configurable interfaces have: That most people don't know what the best interface for them is.
If lots of configuration options are offered, people will choose what they think they like at the time. This doesn't mean they're right. It doesn't mean they'll get things done better or more efficiently. It doesn't mean they'll improve their reaction times with respect to Fitts law or Hicks law. It doesn't mean they'll have a better or more enjoyable experience using the interface. It doesn't mean they'll design a superior interface that'll prevent them from getting RSI or damaging their eyes. It certainly doesn't mean that someone else will be able to quickly and intuitively adapt to their interface, nor that they will be able to quickly and intuitively adapt to someone else's interface.
All that a configurable interface allows is for a user to change it to something that they think might be useful. Frequently, a person won't get around to changing the option, even if they realise they've made a mistake. Unfortunately, users aren't experts.
There certainly are problems that need to be solved, but the linked article doesn't solve anything.
I've come to expect a lot of bitching about gnome(-2.x) on slashdot, simply because the vast majority of users here are people who spend a LOT of time with computers. These are the people who bring up the dreaded workspaces vs desktops debate, bemoan the loss of edge-flipping, and berate the fact that you "can't do that neat thing I've always been able to do in WindowMaker." Others just shrug, smile patronizingly, and say "I don't see why everybody can't just use Emacs."
Ok, so, I have about 100 users in a large department all using linux -- currently both KDE and gnome-1.4 (rhl-7.3). All I can say -- I want an environment that doesn't require a computer degree to configure and operate it. Note -- my users are nearly all PhDs in Physics, or are on their way to obtaining a PhD in Physics, yet still they have trouble figuring out the interface. The notion of setting up our administrative assistants with a gnome-1.4 or a KDE workstation is bordering on silly at the moment.
However, looking at my shiny new gnome-2.2 installation, I must say that perhaps that is slowly changing. This looks MUCH more like an interface for the ordinary people who want to actually USE the applications, not hack them, or learn a separate programming language and a slew of wrist-numbing keybindings just to launch one successfully. Simplicity and responsiveness is the key.
Gnome is a very valuable project for those of us who are looking at maintaining a lot of desktops in a business or educational environment. Currently such setups are frequently limited to Windows or OS X, but both of these platforms come with a huge price-tag both in terms of the OS itself and in terms of admin time spent per each computer -- not only when it comes to the quickness of setup (remote customized pxe kickstarts vs. disk imaging, for example), but also in the area of patching and software maintenance (centralized package updating via RPM, including custom packages, vs... oh, hell, I don't know, everything I've seen on windows/osx is such an horrible hack). However, while administration benefits have been clear, adoption on the desktop has been slow to none, simply because there hasn't been a good, simple, and intuitive WIMP interface available for use on Linux for those who think of their computers as tools to do their day jobs and not in terms of a lifestyle or a political statement.
So, to those working on making GNOME a success on our business desktops I give a resounding cheer. To those who whine about workspaces vs. desktops, edge-flipping, and the fact that there are no longer five different clocks available for their panel -- I'm sure there is a windowing environment that will gladly welcome you. If you want eye-candy, look at Enlightenment. If you want a slew of features -- look at KDE. If you want lean-and-mean, look at WindowMaker and such.
But please don't abuse our cherished gnome developers if your favorite wm feature is not in 2.x, or that you cannot pass the --enable-throbbing-transparencies flags to applications any more. They are out to make a good business end-user desktop, and they seem to actually have a clue as to how to go about doing it. Now we need a gnome2-ified evolution and a decent gnome2-ified browser, and the underlying desktop structure is pretty much complete. Too bad OpenOffice(.org) is using its own widget set. I'm so tired of the "let's make our own widget set" mentality of modern projects...
Anyway -- Gnome2 developers: you are my heroes.
If you open yourself to the foo, You and foo become one.
I've turned full keyboard control on but I'm constantly presented by applications which refuse to focus to check box or radio button widgets with tabs. I can't even imagine what it would be like to use a mac if one were handicapped in any serious way. I have never had any problem using any PC OS without a mouse. On a laptop forcing a user to go to the trackpad to check a box before hitting enter should pretty much relegate doftware to beta status.
It wasn't until 10.2 that Apple finally standardized switching between open windows of an application with Option-~.
I had to install third party software just to get the dock on the right side (where as a poster below pointed out is where the mouse spends a lot of time due to scroll bars and volume icons being there by default).
I even had to create my own black tiff just to make the background black, Apple doesn't even let you select background colors (they use image files in their "solid color" option.
Also the horizontal lines on the default Aqua interface are in my opinion hideous. They make small text hard to read and don't offer any real advantage. Again it required third party software to install a theme that lacked the lines.
OS X is nice, but it is far from being the pinnacle of usability. If anything it is a clear example of an interface which has stressed learnability over usability far too much.
I'm not saying that GNOME hasn't had a lot of rough edges, but in it's current state of 2.2 it seems to have created a simple ui that is as learnable for novices as OS X. It seems to have done more work on acessiblity than any other desktop out there. But I can still define my own keybindings for sawfish if I so desire. It's just a pleasure to use if you're the sort that uses your computer daily.
Metamuscle.com - News in the Iro
Then
For sure!
Profit for...
It's win-win for everyone except Microsoft.
Fire and Meat. Yummy.
Here's a prediction: the moment you succeed in making the distinction between "user" and "programmer" disappear, most of your potential users will disappear with it. Almost all of them who are not also programmers (in the old, pre-disappearance-of-the-distinction-between-users -and-programmers, sense), in fact.
Configurable desktops give users choices, and users like choices provided they're easy to make and easy to undo. Programming is about making decisions which are often hard to make and hard to undo. Unlike selecting a theme for a desktop, which is a matter of taste and the whim of the moment, programmers have to try to get lots of complicatedly interdependent things right in ways that matter.
I look forward to the first release of Tunes - hopefully some time before my retirement - with great anticipation, but at best such "entirely new software architectures" will empower some users do some things that they used to need to be programmers to do.
Experience is a hard school, but fools will learn no other.
Consider this:
- How much time do you actually spend setting up your desktop/preferences/themes/etc... If you're like me, once in the beginning, and very occasionaly a little tweak.
- If you take the previous point into account, don't you agree that these settings should not clutter up your UI, and be stashed away somewhere?
- The more choices are on the screen at any given time, the more time you loose making up your mind. You don't need the distraction of having to deal with prefs when working on something.
So, I personally generally like them "invisible" - give me good defaults, and no boatloads of pref applets. Gnome2.2 does this pretty well imho.
Now when I DO want to change something, I would be very annoyed if I wasn't able to. However, I think most people that want to change their settings are a bit more advanced (caveat: for this premise to be true you better have to have good defaults). I very much doubt that these advanced users (what mosfet refers to as the userbase that want their prefs) can't handle changing the settings in a gconf-like registry. After all, you don't need to do it that often.
If changing a certain setting in Gnome becomes very popular, someone will write a little frontend for it, which over time gets integrated in Gnome proper I suppose - that's what happens if there's enough demand.
Bottom line: make customizing possible, but resist exposing the prefs in the UI when it doesn't make sense to. The newbies and experienced users get something with good defaults that they can use right away, the experienced users additionally have the option of diving into gconf-editor when there's no pref applet for what they want. Since they do this really rarely it shouldn't be a problem, and it's a bunch of clutter, implementation work and testing that's avoided.
That way everyone's happy. Newbies, experienced users, and the developers/designers that actually have to implement and test the little UI's that end up everywhere and untested because it's such an obscure pref very few people use it.
Of course this only works if Gnome does actually expose the needed prefs through gconf - but since I've been using Gnome2(.2) I haven't found one I needed or wanted that wasn't there yet, so I suppose for this user they're doing a good job.
Configurability YES. Clutter NO.
Most non-technical users don't configure anything. They just don't see the need. If it works, why change it? Witness the enormous number of Windows users who have MSN as their IE homepage, who have Tahoma (or Arial for older Windows versions) as font, and who have the standard icky-turquoise, blue, or Teletubbyland backgrounds.
Hence, as long as the defaults are right, there's no great need to spend months umming and ahhing over just how configurable the interface should be, because most of your users won't even bother looking for the controls, they'll just take what they're given. Let's call these people the appliance users, because they tend to treat a computer as a simple appliance to get a few simple tasks done, and outside of doing those tasks they tend not to play much (apart from Solitaire and Minesweeper!). Some of these people are probably better off with a simple appliance-like interface like OEone, rather than a classical WIMP interface. Altogether, the appliance users are maybe 85% of all computer users, possibly more.
On the other hand, it seems as though technically-minded users prefer greater configurability rather than less. They do not mind spending half a day setting up their work environment, because they feel it gives them extra comfort and productivity in the long run, or simply because 'it's cool to make things work the way I want them to'. Let's call them power users, because frankly, that's what they are. This group is somewhere between 5-10% of all computer users, but are by far the most influential on the purchasing decisions of others because of their knowledge about computers.
Then there is a third large group, who sit somewhere in the middle. This group looks for a modicum of flexibility but also looks for a sane set of defaults, enough customizability to make a GUI their own but also to have an environment which 'just works' to get on with the things that they need to do. Let's call these people the 'happy medium' users. These users probably make up somewhere between 5-10% of all computer users, a similar figure to the power users. Apple specializes in catering to this group of users, and probably has a hold on half the users that belong to this group. With MacOS X, Apple have begun to branch out into catering for power users too.
This is what I don't understand about Gnome 2's change in direction: all the major GUIs make an effort to satisfy the applicance users simply by having an initial interface that's not too overwhelming and by having a sound set of defaults. Linux GUIs have made great strides towards satisfying this group in the last year or two - I think there's not much to choose between KDE and Gnome for this group, to be honest. Linux GUIs have, previously, always been good at satisfying the power users. However, Gnome 2 appears to have all but dropped support for power users in favour of catering for 'happy medium' users.
This doesn't make good strategic sense. They have dropped support for one group (power users) in favour of another group, less influential on computing purchasing decisions ('happy medium' users) of approximately the same size. They haven't gained any numbers in doing so, but instead have upset and annoyed a lot of their existing users, most of whom were influential 'power users'.
And why have Gnome done this? Sadly I think it is because their usability team is in awe (for evidence, see for example the essentially pointless button order switch to copy Apple's way of doing it) of what Apple has done with their GUI but don't understand that Apple is a niche player with a GUI that is heavily adapted towards their target market - the 'happy medium' user.
This is why I cringe whenever I see usability engineers talking about Apple all the time. Sure, MacOS has a nice GUI, but without understanding the reasons why Apple took the decisions which make it a nice GUI and the context in which those decisions were taken, you have learnt very little about it at all.
No, they are guaranteed to see downtime for worms and bugs. Linux has more than it's fair share of silly errors and stack buffer overruns abound. Don't believe me? Check out Debian's security announcements for this year alone. And the worms will follow. They were born on Unix. It doesn't matter how configurable or how securable a system can be.
Indeed; no software will ever be perfect. But I think that Unix (Linux being a derivative thereof) has a better ground-up concept for scalability and securability.
I think open source is a double-edged sword here. Open source lets the peer-review process close holes - but at the same time, digging out the source code that a given webserver is running would allow one to look for possible weaknesses.
I personally think Windows is damned securable as long as the person behind the keyboard puts in some effort.No one knows more about Windows security than Microsoft, and even so, they keep on getting hit by worms and attacks.
Windows is obviously quite securable - I'm sure that www.microsoft.com must be a prime target for crackers - but it feels like security has been tacked on after the fact. Look at Windows' ancestry - single user, single computer, single tasking operating systems (CP/M begat DOS which begat Windows which begat Windows NT). Windows itself grew out of that as a series of additions and major redesigns requiring backward compatibility. By contrast, Unix (and its first-order derivative, Linux) grew out of an environment where one computer cost $12 million and had to be used by 1,800 people simultaneously - security had to be inherent to the earliest designs.
Microsoft has done impressive things within their design constraints, but I still don't see it being much more than a sun-room tacked onto the side of a mobile home. There's still no foundation.
Fire and Meat. Yummy.