In this article Joel talks about the number of choices in applications. “This highlights a style of software design shared by Microsoft and the open source movement, in both cases driven by a desire for consensus and for “Making Everybody Happy,” but it’s based on the misconceived notion that lots of choices make people happy, which we really need to rethink.”
UI: Choices = Headaches
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSAlert.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I read the article, which narrowly focused solely on the “shutdown” thing in Vista. He apparently is greatly annoyed by this.
In the end, choice isn’t bad. Differences aren’t bad. End users will see a few ways to shut down (he mentioned keyboard functions on laptops, in addition to the icon and the menu selection). Users will see that this is a little different than XP, but they will choose the one that suits them best.
I am sorry that he is annoyed by this, but please let us not have a society (or an OS) that force-feeds us the one and only way to perform a task. Let people choose for themselves, but provide sensible default methods (and presentation of those methods).
And that is where I think his real problem was: “default presentation”, not the fact there was a “choice”.
EDIT: even using Firefox’s built-in spell checker, I managed to ignore the red underlined misspelled words.
Edited 2006-11-21 18:34
“In the end, choice isn’t bad”
Read the article again. Joel quotes research saying that too much choice IS bad. I have several analogies explaining this:
People have to constrict their options in order to live happily in a big city. Everyone does this. If there are thousands of pubs/restaurants/cafes to go to, most people constrict their choices artificially to the few they get used to quickly, adding in and removing a few choices now and then.
There are currently hundreds of graphic card options from nVidia and ATI. The choice is confusing and sometimes downright infuriating. It makes me unhappy to have to choose between all of these choices. A good shop would do things to help me restrict the choice. That too many choices makes people unhappy has been backed by scientific research.
The exception to this rules are people that care SO MUCH about that particular niche that they love to customise something exactly to their needs. These people will find the iPod infuriatingly simple and will hate “dumbing down” of interfaces. These users will also have hugely different wants and needs and catering to them will not just slightly increase complexity, it will multiply complexity many times over.
They are, however, the extreme minority and if you are making mainstream software it does not make sense making the average user very unhappy by the sheer complexity just so you can cater for these users.
That is why you need to CHOOSE which users you cater for. Either making 70-80% happy or 90% unhappy. Microsoft seems to be choosing the latter, despite being the most mainstream software maker there is.
There are currently hundreds of graphic card options from nVidia and ATI. The choice is confusing and sometimes downright infuriating. It makes me unhappy to have to choose between all of these choices. A good shop would do things to help me restrict the choice. That too many choices makes people unhappy has been backed by scientific research.
Most people I see just go into the store and buy one in their price range. They don’t go into which video card is installed.
You talk about iPods not meeting some people’s needs. This shows that choice is good, too. Not necessarily that every iPod should have every choice option, but that the MP3 player market needs choices. Those that are “infuriated” at the iPod need to have choices and options available. I am equating the MP3 player market here to the UI in discussion. Not just single products. Looking at it in this larger sense, you see that choices are necessary.
I am not advocating that every device should have every choice, just that there should be options and choices available to the consumer.
If there are thousands of pubs/restaurants/cafes to go to, most people constrict their choices artificially to the few they get used to quickly, adding in and removing a few choices now and then.
Yes, but _I_ want to choose _my_ $FEW restaurants and _you_ want choose _your_ $FEW places. My “few” and your “few” don’t have to overlap. That makes cca. 1.5*$FEW (average between no overlap and total overlap) just for the two of us. Taking every citizen into account, you get hundreds of cafes and restaurants in a city. I also want to choose “my few” _myself_. And I also want the rest of the places be available, should I change my mind later, or should the quality of my preferred places drop. So choice, in restaurants, is good.
Actually, the article blew up the number of choices. For one, the icons provided the same choice as the menu text entries. Also, the Fn- shortcuts and display closing provide the same results, just in a different way. So there are 2 layers of choice – what to do, how to do it. And each combination makes sense for someone, so it’s good it’s there.
People complaining about too much choice should live, for a while, in a place with limited choice. Figuratively speaking: one type of pants, one type of shoes, two types of beer, two types of bread, one type of bread-rolls, one type of butter… I lived in such a country, until the Velvet revolution happened. Compared to now, it sucked a lot… Value the freedom to choose. And if you don’t like Vista, choose something else
.
People complaining about too much choice should live, for a while, in a place with limited choice. Figuratively speaking: one type of pants, one type of shoes, two types of beer, two types of bread, one type of bread-rolls, one type of butter… I lived in such a country, until the Velvet revolution happened. Compared to now, it sucked a lot… Value the freedom to choose. And if you don’t like Vista, choose something else
.
Your talking about market choice, this is something completely different. The freedom to choose between Operating Systems (or places to eat or whatever) is a VERY good thing. I dont think anyone is saying otherwise. But the point is that as a guiding principle of good engineering, freedom of choice shouldnt not filter down to the level of variation. It is always better to have one and only one way to accomplish a particular task unless you have a damn good reason to include another one (and variations between user preferences is NOT a damn good reason, you have to draw lines somewhere for the greater good).
For example, I don’t think you are honestly suggesting that having a particular eating establishment serve 50 different kinds of turkey sandwiches would be a good thing… There may easily be at least one or two people in the area of the place that like one of those particular sandwiches, but the fact is that the place would probably do VERY badly in the market. When confronted with that many choices people will (for the most part) simply give up and go somewhere else next time.
Edited 2006-11-22 16:57
For example, I don’t think you are honestly suggesting that having a particular eating establishment serve 50 different kinds of turkey sandwiches would be a good thing…
Well i know some pubs who have more than 400+ different kind of beers.It’s something different.The place is allways crowded in the weekend.I don’t know wether that’s due to the large variety of different beers but still it doesn’t shock them either.
Well i know some pubs who have more than 400+ different kind of beers.It’s something different.The place is allways crowded in the weekend.I don’t know wether that’s due to the large variety of different beers but still it doesn’t shock them either.
There is a difference between the beer example you give and the one about the turkey sandwich. Restuarants _make_ turkey sandwiches, pubs dont make beer – they buy it from companies that do. Now name a company that actually makes 400 different kinds of beer… Pubs are retailers, not producers.
Also, pubs generally work on the principle that they have “anything” you might want. If you think about from the point of view of the customer the pub is in fact not only limiting the burdon of choice, they are essentially eliminating it. The customer knows if they ask for _any_ kind of beer, the pub will have it. There is no complicated interface, and there is probably only like 3 price categories (domestic, import, really explensive) so there is pretty much no reason to even display a beer list.
What if when you ordered your beer the bartender said “sure, you can order, but first you MUST physically point at your beer from our list – even if you already know what you want.” That wouldnt fly would it? That is what UI designers have to tangle with.
Unfortunately the “we have anything you want” approach doesnt work in the software world yet…
Edited 2006-11-22 20:12
Sure it could, it’s all in the packaging. Obviously there is some truths in what the article says, some level of his proposed simplications would be great.
The Linux community has this figured out in principle. Most distro’s default to Gnome or KDE, but for those of use who prefer otherwise there’s a ton of WM’s from ratposion to enlightment, all kinds of applets & whatnot, etc. – the choices don’t necissarily have to be appearant to the [new] user, so long as their easy enough to do & the features still exist.
One of the nice things of a CLI, you can add as many programs as you want, it hardly clutters the environment nor interferes with it’s function.
Um. Correction. Joel doesn’t quote “research” as in peer-reviewed scholarly articles. He quotes a reviewer summarizing a book that goes on the same shelf as “The Tipping Point” etc. There has been actual research along these lines, but the article I read (admittedly, a “layman” re-write of actual research in SciAm) weren’t investigating things like gui buttons. More along the lines of important life choices… It’s a long stretch to apply it to windows buttons, and he certainly doesn’t make the extension in any rigorous way. More like throws it in there for the appearance of credibility… I thought I’d give Joel another try after the admins here assured me he was “great”… oh well..
“Joel quotes research saying that too much choice IS bad.”
I bet we can find research that comes to exactly the opposite conclusion.
“People have to constrict their options in order to live happily in a big city. Everyone does this. If there are thousands of pubs/restaurants/cafes to go to, most people constrict their choices artificially to the few they get used to quickly, adding in and removing a few choices now and then. ”
You realize of course that this is exactly the opposite of restricting choice. The article would want you to only have a few select pubs/restaurants/cafes to go to, not thousands.
Edited 2006-11-22 04:33
That is why you need to CHOOSE which users you cater for. Either making 70-80% happy or 90% unhappy. Microsoft seems to be choosing the latter, despite being the most mainstream software maker there is.
When you are granted a monoply (copytight) on production of a good that costs close to zero to produce. It seems like a good strategy to base your income on selling as many identicalt copies as possible.
Add to this the fact that the more features you add to a software the less common denominators among the users to design for will be found.
Hence copyright sort of implies that the best software is the software that is designed to make 90% of the users unhappy (using your numbers).
If we now add to this that only one third of the industry chooses this approach, even if it’s clearly the most profitable. I think we can safly conclude that the abolishment of copyright would result less software designed to make make 90% of the users unhappy, and more software designed to make 70-80% happy (actually I think this number would be much smaller).
I read the article, which narrowly focused solely on the “shutdown” thing in Vista. He apparently is greatly annoyed by this.
It was more likely just a convenient and highly appropriate target for his rant.
In the end, choice isn’t bad. Differences aren’t bad. End users will see a few ways to shut down (he mentioned keyboard functions on laptops, in addition to the icon and the menu selection). Users will see that this is a little different than XP, but they will choose the one that suits them best.
I only partly agree. SOME types of choices are good, but in this case I think we are talking about one that definitely is not. All of these methods of achieving what is essentially (read that again… essentially) the same thing (turn computer off) are nothing more than abstractions designed to convey meaning to the user. The problem is there are simply too damn many of them. As he pointed out in his rant, hardly anyone knows or cares about the difference between sleep, hibernate, lock, poweroff, etc. IMO their shouldn’t even be a distinction, that was the mistake made in the first place. If MS (or any other OS maker) had simply thought about it a bit they could have come up with a sensible way to integrate all this functionality in a way that was not exposed to the user at all. There simply is no need for it.
To use the popular example that is always brought up when speaking of good UI design, think of how a car stereo works. There are exactly 2 ways to make the music stop:
1. Press the on/off button.
2. Turn the volume all the way down.
Both these methods have a concrete purpose, the most obvious being the on/off button (turns it on and off – music/no music). The volume knob, as well as allowing volume adjustment, can also be used as a mute, i.e. “I want to make it quiet for a minute or two so I will turn it down”. There are VERY few car stereos with mute buttons (although there are quite a few that integrate mute with a connected cell phone being used – that is a good thing). There is essentially no reason for a mute button, it would just muddy up the interface and consume valuable real estate.
If a car stereo was designed by a typical clueless-about-interface-design software engineer, it would almost certainly have a mute button, because that would seem to the engineer an obviously necessary function (“Why would you want to have to turn a knob when you can just hit a button – its less work?”).
That is the root of the problem I think. To much effort is spent on getting a specific result with the least amount of work. Not enough time is spent weighing that against the cost in complexity for the user.
I am sorry that he is annoyed by this, but please let us not have a society (or an OS) that force-feeds us the one and only way to perform a task. Let people choose for themselves, but provide sensible default methods (and presentation of those methods).
Force-feeding users on methods of performing a task is essentially the FUNCTION of any modern OS… If you think UIs are about choices you are completely and utterly missing the point. They are about _reducing_ choices to a manageable level.
The time when UIs were about choice happily ended a decade or so ago with DOS and its hodge-podge of interface paradigms.
And that is where I think his real problem was: “default presentation”, not the fact there was a “choice”.
Ill concede that default presentation is often a good way to address this kind of issue. But it should be weighed against the costs in terms of implementation complexity and most importantly need. For example, while sleep and hibernate are two distinctly different functions, I don’t think they _need_ to be distinguishable to the user. There is very little benefit to distinguishing between the two. There is a cost of making that distinction available and IMO it simply isn’t worth paying.
and what if the volume knob have a very long travel distance? as in i have to turn it several revolutions to have it go from high volume to quiet/mute?
then a mute button is godsend if one wants to have the music stop NOW!
btw, some car stereos have the on/off and the volume integrated into one. turn the knob past 0 and you hit a notch that turn the stereo off fully (some people have a habit of turning it down low but not over said notch, leaving the stereo on but in a “muted” state”).
now if one was to integrate the on/off button and the volume into one so that if you press said knob you go on/off and turn it to adjust volume (just make sure it does not do so while off) then you don’t need a dedicated mute. just hit the on/off to “mute”.
still, the best compromise i can see is to have a configureable UI. supply it with maybe two defaults:
– simple, that integrates everything into as few options as possible via timers and scripts.
– advanced, that allow the user to select exactly the actions wanted and nothing else.
and if the user feels like doing so, can go fully alien by writing/defining his or her own UI.
thats what we should never forget about software UI’s, unlike hardware UI’s, software UI’s are infinitely flexible.
it would be oh so good if every programs UI was a XML or similar ascii file that defined the actions and options of each UI item, and what signals should be sent to a background DLL or similar when performed.
this way one could build custom UI’s for often performed tasks. or even integrate many smaller programs into a single UI.
hmm, kparts anyone? or even unix style console toolboxes (optionally with a TCL/TK frontend)?
or even litestep on windows?
and what if the volume knob have a very long travel distance? as in i have to turn it several revolutions to have it go from high volume to quiet/mute?
then a mute button is godsend if one wants to have the music stop NOW!
No. If that were the case your problem is that you have a very badly designed volume knob… You solution is like applying glue to hold together ducktape.
now if one was to integrate the on/off button and the volume into one so that if you press said knob you go on/off and turn it to adjust volume (just make sure it does not do so while off) then you don’t need a dedicated mute. just hit the on/off to “mute”.
Actually, that is EXACTLY how my car stereo works. But the point is that “mute” is not really a separate function. Mute is just a pseudonym for on/off. So as I said originally, the presence of a mute function is pointless and can easily be “designed away”.
it’s like when you go to buy your colgate tooth paste. You have so many options, and all seem to do the same. It really annoys me.
I don’t know about you, but i like the “physically off” option very much and would miss it. How it is invoked i don’t care, but i want to be able to do it in a distinguishable process.
So, i want two options: “switch user/log off” and “power off”. I don’t know why i have to look at a “hibernate” button when logging out off Gnome or KDE.
But I want “Reboot” and “Hibernate”. “Physically off” makes no sense to me because I don’t want to restart all the apps every time I start the computer. “Switch user” also makes no sense because only I use the computer. “Log off” makes no sense either because all you can do is log back in.
I think it’s you who are in the minority. But this maybe highlights that after all two options are not enough.
Edited 2006-11-21 19:58
Switch user would make sense to only appear on a machine with more than 1 user (other than System).
Log off always makes sense. Suspend and hibernate only make sense when they’re supported. But I think they’re distinct. Hibernate saves where you are and you can store your machine without power for ages, where suspend lets you resume very quickly but 3 days later it shuts off (or more likely it hibernates after 2 hours).
But I do think getting rid of hibernate isn’t a bad idea. Just give a setting for how long to stay suspended before you hibernate. Or, like in XP, have a shit option to hibernate instead of suspend (this hides complication from users who don’t know the terms and leaves support for advanced users).
Again for me that won’t work because my machine doesn’t support “Suspend” (to RAM). But every machine supports “Hiberation”, and I’m using it, and if it was removed I would change to an OS that supports it, because it really saves me from a lot of time-wasting.
> “Physically off” makes no sense to me
Because you never move your computer?
?
?
Hybernate comes in different flavors. One of the tastiest flavor of hybernate is “suspend to disk”.
On “supend to disk” the state does mot get lost on physical power off, because it was written on HD.
But actually i do prefer physical off, so choices seem to matter.
I don’t see anything wrong with either Gnomes or Vistas shutdown options. Actually they seem to be braindead obvious to me. Without ever having used Vista i can tell you that the most left button corresponds to shutdown, the middle one to log off and special laptop ways of shutting down are useful because a few of us actually sometimes use their laptops in a mobile way.
Closing the lid usually corresponds to a hibernate mode unless configured otherwise by people who know what they are doing.
I mean, comeon… do we realy have to remove crucial functionality whenenver there are few people who aren’t smart enough to use them? Then we would all end up with crap.
I mean, comeon… do we realy have to remove crucial functionality whenenver there are few people who aren’t smart enough to use them? Then we would all end up with crap.
You have that backwards… You (and me and most of us here) are part of the few. We are the minority… Most computer users (probably 80% but that is an educated guess) really have no idea AT ALL about what we are discussing right now. The ideal solution for them is simply a button. ONE button – “make computer go bye bye”. They dont care about methodology – just result.
Anyway, you dont HAVE to remove functionality, you just have to design it in a way that makes it invisible to the average user. Firefox is an excellent example of how to make software for the masses. Its core feature set is simple and effective – does what the majority of people need and nothing else. Simple interface, etc. Plugins handle the rest. Best of both worlds. That is the best approach imo.
Here is another solution: a big button that says ‘END’ which, when pressed, does the following:
1) logs out the current user.
2) goes into sleep mode.
3) after an hour of sleep mode, it shuts down.
In this way, user switching – sleeping and shutdown becomes one. If, the user touches the keyboard or mouse while in sleep mode, the login screen comes up very fast. If more hours has passed, then the computer should be switched on.
Too many ways of achieving the same thing is generally a bad way of designing a user interface. The geek in me would probably defend the design by calling it flexible. However, a better word for it is complicated.
Who’s to say it’s a bad thing to achieve things multiple ways? That’s what makes software distinguishable from (and more flexible than) hardware, and what makes the web such a useful tool–LINKS. What if that one single way is hidden 5 dialog boxes/menus deep (or for the web equivalent, you have to click through ten different pages instead of just Googling it)? Sometimes having multiple ways of getting at the same information is, in fact, a great option. One example is that MS successfully manages to cater to experienced Windows power-users, who get at settings via the tried-and-true route of the control panel, and others who don’t touch anything that’s not top-level visible. The newbies can just see the most relevant options already presented to them in the Start menu without wading through the control panel jungle, while the control freaks can still have their fun. Admittedly, none of this would be necessary if MS cleaned up and cut down on the control panel in general, but to be honest I’d rather have hard-to-find GUI customizability than impossible-to-customize software–at least it attempts to strike a balance.
Edited 2006-11-21 23:49
Personally, this is how I do things on my MacBook:
1. Shutdown – menu option
2. Restart – menu option, although I can’t remember the last time I used it
3. Sleep/Hibernate – I close the lid. OS X used to sleep (keep RAM “on”), now by default it hibernates (write RAM to disk). This is configurable, so if you don’t like the default, you make your choice *once* in System Preferences and are not faced with an extra choice each time.
4. Lock screen – Move the mouse to the top-right corner to activate the screensaver, which is password protected.
5. Switch users – There is a list of users on the right of the menu bar, completely apart from the Shutdown/Sleep/Restart items, which are on the left side of the screen. It can be enabled/disabled in System Preferences.
Windows Vista broke the shutdown / log off shortcuts (which annoys me to no end). But in Mac OS, you can use Ctrl+Option+Cmd+Eject to shutdown with no prompt, or Ctrl+Cmd+Eject to restart with no prompt, Cmd+Option+Shift+Q to log out with no prompt; however there appears to be no shortcut in OS X to switch users or go back to the login screen??
and set up through power management.
Sleep after XX minutes and hibernate after YY minutes where YY > XX.
Use/Case thinking is essential here. What does the user really want to do with the computer when leaving it?
For home computers I think Joel is almost right, but there is three use/cases:
1. I will be leaving it for quite some time and want to save electricity.
2. I may be returning soon and don’t want to spend the necessary time to start the computer again. And:
a. there are others that may try to use the computer in the mean time
b. there are noone else to use the computer.
2a may require a lock-button.
For workstations there is also:
3. I will be leaving the computer and may not return to this particular computer, but the computer needs to be accessible for others.
For 3. A log out button will be required.
Some homes computers work a little like workstations and three buttons may thus be necessary: Switch off, Log off and Lock.
Switch user should definitely be combined with Lock.
“Why do you want the power off? If you’re concerned about power usage, let the power management software worry about that. It’s smarter than you are.”
Sure. I assume it automagically knows what I’ll do next better than I myself do?
Just like the god damn paperclip/lightbulb/dog or whatever incarnation
of help function always knows what I need to do, right?
Die, lightbulb, die!
Every OS should – at least during installation – offer a choice between a dumbed down and a sophisticated interface.
I _really_ hate it when people tell me my computer is somehow smarter than me.
I wish it was but it isn’t.
The computer should not try to outsmart me, it should just do what i tell it to do. Period.
What’s a blessing for the computer illiterate is a curse for everyone else.
The computer should not try to outsmart me, it should just do what i tell it to do. Period.
What’s a blessing for the computer illiterate is a curse for everyone else.
I’m with you on the burning hatred of self-invoked automation. I’d add wizards to the list. All of them. They’re no good. But I completely disagree that simple/advanced switches are any kind of solution, at least in the implementations I’ve seen, where true and unique functionality is hidden behind a curtain, adding an extra step for those who need the features. They also tend to end up looking like an excuse for programmers not to bother making anything organized or intuitive. “They’re ‘advanced.’ They’ll figure it out, or they’ll live without it.”
I know the villagers will grab their torches for this, but I’ll go ahead and mention how Macs do it: the simple, all-the-time operations are in buttons on toolbars and in corners, and more advanced, less frequently used settings and functions are up in the menu bar. The really advanced, not-for-all-eyes settings are in xml .plist and ascii .conf files where only advanced users and neat hacking tools (like OnyX) will dare tread. No walls, no curtains, no warnings, users just find their natural level of expertise and live with it. Stratification is the answer, not options about options (but it takes a lot more imagination and testing than just “toss them an option.”)
Thanks for your reply.
I think you might be right about a switch not being the best idea.
However if I had to choose between a dumbed down interface with an option to switch and one without I’d still choose the first – not because it’s all so awesome but because it simply is the lesser evil.
Oddly enough, I have never met a single person who liked wizards
But I got an idea why many OS seem to be catering for geeks and why that is a smart thing(tm):
Programs are (besides other factors) what makes an OS great.
The more geeks an OS attracts, the more programs will be written for it.
I know I’m simplifying here but I think the point still holds true.
So getting programmers to like your OS is the first step in making it popular.
I think Linux is an excellent example.
On the other hand you try to attract noobs as well so you must make a compromise.
This is probably the reason why every OS sucks for everybody (though not equally bad).
As you said, hiding the more advanced stuff deeper inside the menu is a pretty decent idea.
I’ve never used a Mac but it sounds interesting.
Another alternative should be using a right click menu.
More options definately raises my stress level. Just going out to buy a pair of shoes is one way to set me off. Too many choices. So I agree with the premise that less is more.
However, Vista tries to be a general tool, all things to all user types. In so doing, it can never be exactly what each class of user deems optimal. And it would be difficult or even more confusing to make the UI modifiable or adaptable to user patterns. Office tried this with the expanding menus that tried to limit user choice and confusion. Interestingly, Office 2007 takes the other end of the extreme with the Ribbon interface – presenting as much options and features as possible. So it is clear and valid to say that MS struggles with this as well.
I agree with whoever above said that most folks will find a way to do something and stick with it. The myriad other options are then ignored. I don’t think there is a single right answer on this issue. Grayscale is not what binary coders want to hear I realize!
…it’s decision-making that’s bad.
Let’s say you go into your average office supply store to look at chairs. Wow! There are a lot to choose from! Different styles, colors, fabrics, heights, etc. How do you pick? You could spend three days in there trying to narrow it down, and be in tears of frustration when choosing between your last two choices, because, well, they just both have their good points and bad points.
It’s a difficult three days. No doubt about it.
But, each day when you sit in that chair for the next couple of years or more, you’ll probably be pretty glad that we got past the one-size-fits-all wooden chair days.
—
In the real world, we use analysts to help narrow down choices for us. But, (ideally) that should be what they do: advise us with a course of action, not take that action for us.
You don’t want to eliminate your 15 ways of shutting down the computer. You might even want to add more. You really (if you’re the uncle confused by the options) just want an analyst in there to help you along. Something that will suggest locking at lunchtime, logging off at night, shutting down on the weekend, but always giving you the other choices if you want them.
Because heaven forbid you Sleep when you mean to Hibernate.
In the real world, we use analysts to help narrow down choices for us. But, (ideally) that should be what they do: advise us with a course of action, not take that action for us.
You don’t want to eliminate your 15 ways of shutting down the computer. You might even want to add more. You really (if you’re the uncle confused by the options) just want an analyst in there to help you along. Something that will suggest locking at lunchtime, logging off at night, shutting down on the weekend, but always giving you the other choices if you want them.
Please tell you didn’t mean something like Clippy!
*shudders*
Edited 2006-11-21 19:49
From the article:
“Inevitably, you are going to think of a long list of intelligent, defensible reasons why each of these options is absolutely, positively essential. Don’t bother. I know. Each additional choice makes complete sense until you find yourself explaining to your uncle that he has to choose between 15 different ways to turn off a laptop.”
When I read such bull**** I’d love to go to his uncle home and tell him “Hi, your nephew thinks you’re a complete, utter, hopeless moron and moreover explains it on his public blog. Oh, and this is a baseball bat, just in case.”
People must stop to dumb down computer because, in order to use a computer, you must learn how to. There is no other way. Many people praise the Mac as the easiest to use computer, but I’ve seen people completely unable to understand even it, because they don’t know how to use a computer, period.
No matter how much you dumb down the interface, when they’ll run out of disk space or they’ll have to install a printer they’ll just shout “WTF?”. A friend of mine complained that “the computer is getting slow because the memory is full of films”, unable to understand the difference between a RAM and an hard disk.
The problem is simple. Computers are complex tools. They are the most complex tools ever, because they can do almost everything. Complex tools require complex knowledge, there is no other way. You can achieve some very basic task without this knowledge, ok, but this is not using a computer.
The real problem is that we’re giving computers to people without having them learn what a computer is, like if they are simple, dumb objects. Even worse, most people do not even feel the need for this, and even geeks are caught by this trap, feeling that “computers must be easy for everyone”.
You see the results of this philosophy. People caught in ridicolous phishing traps/virus/spyware, crappy MySpace pages, subpar software that dominates just because of inertia, and people wasting time in dumbing down interfaces for people that will anyway have problems with them instead of help people (and computers) doing complex, useful tasks.
What we really need is alphabetization. In schools, for example. Really. Badly need. Think how more productive would everyone be with a working, inside knowledge of how the most useful, general purpose tool ever built by mankind works. Think about how security would be a much different problem. People must learn. There is no other solution about this.
Interesting rant. Completely off-topic, but interesting.
Oh and btw:
“the computer is getting slow because the memory is full of films”, unable to understand the difference between a RAM and an hard disk.
If he is using FAT32, or pagefile-ing off the same disk, then it’s quite possible that he was right. Computer performance changes are more likely to be based on HDD factors than RAM factors.
She was just full of spyware and unncessary programs starting. The bloat was on RAM and CPU.
And the rant is not off topic, because complex tools require complex interfaces. I’m all for usability and making the interface as easy as possible, provided it allows me to do everything I want to do and it leaves all ultimate choice to the user.
And the rant is not off topic,
I didn’t explain. The reason I said this is because the article is not about ‘dumbing down’ software for people who can’t understand computers. It’s about making software simpler to use. If I’m faced with 1, 2 or even 3 choices when I want to do an action, I can choose the best one without slowing down. If there are 7 or 9 choices, I have to stop, and consider each option before choosing the correct one, every time. If I’m in a hurry, I’m more likely to choose the wrong option and waste power / etc..
I’m not computer illiterate, I just like good design, and don’t want to have to program my own version of a shutdown dialog.
because complex tools require complex interfaces
On what authority do you claim this? The complexity of an interface should always be as simple as possible. A Times crossword setter I once knew believed that coputer UIs should only have one butotn labelled ‘do it’ and that software should be able to contextually work out what to do every time that button was pressed. The reason why we’re faced with 1000 buttons instead of 1 is because programmers aren’t good enough to write interfaces that simplify things.
Again, y
If I’m faced with 1, 2 or even 3 choices when I want to do an action, I can choose the best one without slowing down. If there are 7 or 9 choices, I have to stop, and consider each option before choosing the correct one, every time.
I emphasize “every time”, because it is what is wrong.
If you have even 20 choices, you quickly learn what are the 2-3 ones you need more often. You have a lot of drawers in your home, I guess, but you don’t have always to “stop and consider” where are your pants. Am I right?
The complexity of an interface should always be as simple as possible.
Right. But not anymore simple.
The reason why we’re faced with 1000 buttons instead of 1 is because programmers aren’t good enough to write interfaces that simplify things.
Programmers often spit out clunky interfaces (but remember that good GUI programming is damn hard), but this has nothing to do with the number of buttons: more with their positioning, and meaning.
When I read such bull**** I’d love to go to his uncle home and tell him “Hi, your nephew thinks you’re a complete, utter, hopeless moron…”
What if his uncle is already tearing his hair out in frustration, as my dad frequently wants to do when he has to install some stupid service pack?
Software does not have to be difficult to use in order to be useful. Sure, if you want to do something complicated with your computer, expect to use a program that is fairly complex, e.g., Photoshop. But if you’re just wanting to shut the damn thing off? Give me a break.
A friend of mine complained that “the computer is getting slow because the memory is full of films”, unable to understand the difference between a RAM and an hard disk.
What you’re suggesting here is equivalent to telling someone that he/she has to be able to know how an engine works before he/she can drive a car. Driving and fixing cars aren’t really related, however, other than the object in question is the same.
If you create a simple, intuitive interface (e.g., a steering wheel, gas pedal, and brake pedal) most people are going to be able to learn how to use the tool just fine. They may not understand how to fix a computer when it fails, but just like with cars, that is not a requirement to be able to use it, nor should it be if a software company wants to successfully reach the largest market possible.
Not everyone has time to learn how a computer works. If that’s a requirement, many of them will simply find another way to accomplish their task, or just not do the task at all. Look at it this way: do you want your doctor spending his time learning about hard drives and stuff so that he can use his hospital’s patient database, or do you want him spending his time figuring out how to help a cancer patient?
Would be very desirable if the same convention would be used in similar design cases.Will probably give access to an even larger group of users.But there will always be a learning component no matter what you do.
Consistent design is what matters and not per se the amount of applications.Choice isn’t bad.
Edited 2006-11-21 22:11
Driving and fixing cars aren’t really related, however, other than the object in question is the same.
Using and fixing computers is not related either. Your dad would be less frustrated, if he let a trained person to apply the service pack… (by the same analogy).
Where did he ever say that people needed to learn about the inner workings? All he said is that people need to learn some basic skills. And no, people can’t just get in a car and start driving. That’s what driving lessons are for, remember?
Okay, he actually did use the words “inside knowledge” didn’t he? And for that I’ll have to mildly disagree. But I still think the car analogy is inept.
The car analogy is crap, as cars are a whole lot simpler thean computers. Perhaps an aircraft would be a beter ananlogy. You wouldn’t expect a pilot to fly you from one place to the next without having knowledge of how an aircraft flies.
If you create a simple, intuitive interface (e.g., a steering wheel, gas pedal, and brake pedal) most people are going to be able to learn how to use the tool just fine. They may not understand how to fix a computer when it fails, but just like with cars, that is not a requirement to be able to use it, nor should it be if a software company wants to successfully reach the largest market possible.
Your analogy is flawed.
The OP that you’re replying to is stating that computers are complex, thus there WILL be some learning curve to use them effectively. Yes, cars do have a simple user interface, but we still require a minimal level of competence to operate them on the public roadways. We do not allow young children, the blind, nor the irresponsible operate them. State governments require potential drivers to take driving courses and pass a test prior to being given permission to operate a vehicle. One can very easily ‘operate’ a vehicle with no training (when operating it is defined as making it move in one general direction or another), but that doesn’t make them a driver. Even many who have been driving for years are not really all that good at driving and do not know how to react properly in emergency situations.
If cars are so intuitive and easy, why does it take time to learn these things? Why is the testing necessary?
The OP that you’re replying to is stating that computers are complex, thus there WILL be some learning curve to use them effectively.
Are you suggesting that a car is not a complex machine? Did you miss the part where I stated this: “If you create a simple, intuitive interface (e.g., a steering wheel, gas pedal, and brake pedal) most people are going to be able to learn how to use the tool just fine.”?
Obviously, with computer software (as with a car), you have to learn the things necessary to complete the task. What a good interface should do is make handling those necessary things as intuitive and simple as possible. It should not also bog a user down with unnecessary details that get in the way of completing that particular task.
The car interface analogy is most definitely applicable. If all Fords had 5 steering wheels, each one being used for a different degree of turning, or perhaps a different direction, the interface to these cars would be unnecessarily complicated, and would likely prevent many people from using it who otherwise would like to. Those people are probably going to find another make of car that has an interface that is easier to learn and will accomplish the same task.
The article focuses on the abundance of options presented to the user when the user goes to logoff/shutdown in Vista. Joel obviously believes the number of options in this case is overkill and will get in the way for most people, and I definitely agree.
Look at it this way: do you want your doctor spending his time learning about hard drives and stuff so that he can use his hospital’s patient database, or do you want him spending his time figuring out how to help a cancer patient?
Thank you for your emotionally loaded but flawed argument.
I want my doctor to be able to do both. More: it’s his/her duty to be able to do both. In fact, using the hospital’s patient database is one of the skills he must have to help a cancer patient, just like reading x-rays.
I’m a molecular biologist working in a biophysics lab. My work revolves around protein structure. Nonetheless to analyze, use proficiently and share my data I badly need, in the end, to know how a computer works, and I even had to code a data analysis program (thank god I knew some coding by myself, because the Italian university programs still haven’t learned that coding is a necessary ability for most scientists).
In every workplace that relies on computers you can see how the productivity is worsened because of poor computer knowledge. I am more computer savy than the average, but I’m surely very much more ignorant than I should be, and I’m damn sure this is costing me something.
What you’re suggesting here is equivalent to telling someone that he/she has to be able to know how an engine works before he/she can drive a car. Driving and fixing cars aren’t really related, however, other than the object in question is the same.
Oh, and by the way, all driving courses here in Italy teach you how basically an engine works, before they allow you to drive a car, AFAIK.
The same here in the Netherlands.I think rightfully so.
There could be way less more then 15 options to turn off the damn thing, but if you know many of them you can tell your uncle which one is best for you and I’m sure he’ll learn it and use, so why all this fuss?
I prefer Microsoft because the default installation is perfect. No need to choose between KDE vs. Gnome or Firefox vs. Opera.
Yup, because Wordpad and Paint contains _all_ the functionality we need
EDIT: This is irony. In case somebody should be in doubt.
Edited 2006-11-21 19:57
Maybe if you’re using SP2 with all updates and Virus-scanner preinstalled, then at least you have a system that won’t install spyware just by visiting the wrong website. Any other scenario, if you don’t know what you’re doing, you’re screwed. And then there’s that point about WordPad and Paint… quite insightful… just what are you planning to do about that? (Oh hold on let me guess–just do it the Microsoft way, however crappy it may be in comparison to other products! At least no choice is involved.)
In the end, I think a lot of this comes down to trying to provide as efficient an interface as possible for the supplied functionality.
My concern about these types of conversations is that they often turn towards removing features. I am all for my interface being made more efficient. I’m all for sensible defaults that can be changed by users who have an interest in doing so. I really hate it when people start removing functionality and then tell me I’m better off for not being able to do what I want to do.
Here, here!
hmm… I got modded down for trying to get another person’s comment modded up… interesting.
The Design Of Everyday Things covered this and a whole lot more over twenty years ago now. More developers should read it.
While I’m a proponent of giving choice and variety in the software world (for me it’s one of the biggest appeals linux and free software in general), there are times when such abundance of “choice” can be highly distracting. I’m thinking of DVD players in particular as one example (that is, players you use on your computer).
Now, take a real world, $50 dollar DVD player. No blue ray, uber fancy, ultra expensive high end model, just your basic pop in the dvd, sit back and happily watch a movie type deal. For me, and I suspect most movie watching folks, that model works quite well.
Now take a DVD player on a computer. Here I’m going to have to pick on my favorite platform, free software, but sadly it’s not as always as straightforward and simple as the above real world model. (In the proprietary software world it’s admittedly somewhat better in this regard for this particular type of app.) On a linux machine, we have a nice choice of various video players, capable (provided your codecs, etc., are present) of playing DVDs. In fact, they can work quite well. Problem for me, is that they seem designed by and large for professional or hard core videophiles. Even on the “beginner” settings on something like xine, I find myself rather confused on what all those settings do and signify. Really, all I (and again, I suspect most folk) want is to just get the freaking player to play a movie without looking awful and requiring a fine amount of post-tweaking. Take the case of de-interlacing, there’s one player, (it may be vlc, but I’m not sure) where I seem to recall there being something like 5 different ways you can specify how you want the rendering to occur. Sadly, there’s no “look good, not sucky” option to be found…
I imagine this works good for a certain segment of the population, but again, I suspect that’s a minority. Most folks just want the app, any app really, to do the job intended, and not get in your way.
Here again, though, it’s a matter of sensible defaults. While I agree that VLC’s interface could use a lot of help in the ease-of-use department, in principle there’s nothing wrong with providing all those video settings, as long as they’re set to the best average defaults and the interface to change them doesn’t interfere with the normal course of use. Believe it or not, I actually find playing DVDs with PowerDVD on Windows to be as easy or easier than with a standalone player, since I get to use my mouse and there’s no remote control that I need to acclimate myself to.
That said, I think programs like Firefox have shown that even the setting of options itself can be made much more intuitive, and in this sense there are many Linux programs (like VLC and many KDE apps) that could be greatly improved.
I just don’t get his problem with the off button. You have a normal off button and a lock button readily available. If you want more options, you have to dig for them. The first time I wanted to shut my computer off in Vista I brought up the Start Menu and hit the button that looked like an Off button. Lo and behold, the computer shut off.
Fair enough. I don’t imagine most people will think it through any more than that.
It’s ridiculous that we have that many choices for turning off a laptop. Now, what’s worse, these choice doesn’t satisfy my needs (which are rather basic)! (I want to hibernate and reboot into Linux.)
I think there should be two buttons: Turn off and switch user. The turn off button would hibernate the computer. The switch user button would log the user out.
Additional options (especially “restart”) could be available via a right-click menu or some other clever “intuitive” mechanism normal users will never find.
People want choice. They just don’t want to be PRESENTED with choices they don’t understand or care about.
The Microsoft model is that, in general, they make the choices for you. That way everyone does everything the same way. Uniformity breeds familiarity and familiarity will overcome the inherent inflexibility. If a process is rendered so obtuse by lack of choice, it’s a new market opportunity for a higher-priced niche application.
The GNOME model is that you have choice, making the choice, or sometimes even knowing that you have the choice at all, requires some research and effort on your part.
This is close to the Apple model, where you have reduced choice, but they try to provide a button/link/tab with additional choices that would be expected to be made by a small fraction of the general users. Some choices are hidden or difficult to find, but they aren’t absent.
The KDE model is every choice is presented up front. Care is taken to attempt to pick reasonable defaults, and there’s even customization that can be performed to mask the existence of menus/buttons/functions, but basically it’s very nuts-and-bolts — if there’s a feature, you can get at it. However, there’s been at least a few suggestions that in KDE4 that the notion of expertise levels be associated with things so the amount of “choice” presented to a user is dependent on their comfort level (perhaps like what Xine does).
Of course, now users have a choice which dogma of choices they’d like to be presented with.
“The Microsoft model is that, in general, they make the choices for you.”
The article is specifically discussing the startup choiced in Vista, a MS OS, so your statement is kinda offbase. MS makes a lot of choices for you, but they also always present several different choices to do a particular task. User management on the local computer, for example, can be reached by these different methods:
1. right-clicking “My computer” and choosing manage.
2. going to Start->programs (or all programs, if you like the start panel) and go to Administrative Tools (which may be not available, it’s your choice), and choose “manage computer”
3. goto control panel and choose users.
4. goto control panel, go into Administrative Tools, and choose “manage computer”
and this is one of the many examples of how MS provides choice in it’s OS. it may make some of decisions for you in terms of operating parameters, but you can pretty well navigate the OS the way you want, and that is what is most important. Plus, if you care enough to change the parameters of the OS itself, there is the registry, and countless free utilities to let you tweak to your hearts content.
MS doesn’t make all the choices for you, just the ones that make sense for normal users, at least IMO. Compared to Gnome, which is not nearly as configurable as Windows, or KDE (my favorite) which is infinitely configurable, it sits comfortably in the middle.
Joel makes a good point. Professional trendwatcher Herman Konings ( http://www.nxt.be/index.htm ) calls this phenomenon of simplification consumer.0123 (in dutch consu-mens.0123). O is for the number of manuals, 1 for the number of access buttons (on-off), 2 for the number of function keys and 3 for the number of seconds in which I need my answer. For an example look at 2 of the most recently popular appliances : Senseo and the iPod.
Oh and simplification != dumbing down unless you have a bad designer. Simplification can make an interface elegant and intuitive.
Yes, confusing and confounding sometimes, but it needs to be there. Even the hundreds of versions of Linux have their uses….although I’m still not exactly sure what they are.
As far as mac’s being easy to use….I gave my sister my old bondi imac running OS X 10.1. I also bought her a 2 button mouse so it would feel more like what she’s used to. She couldn’t figure out how to get the pics people attach to emails to save on her hard drive….and even if she did manage it, she would lose where they were. Pretty soon she had that machine as hosed as a lot of the Windows PCs I’ve come across. Everything has a learning curve, especially OSs. Mac OS is no easier/harder to use than Windows. If you don’t believe that, then ask a Windows user to eject a CD on a MAC, for example.
As far as tutoring people who own a PC, you have to teach them one way, and one way only….especially to those who refer to the tower as a “hard drive”. Get them in the OS, get them out of it, and get them done as quickly as possible.
Just a guess here, but I believe that, like memory for numbers, people have a certain number of choices that they’re comfortable with. Too few choices, and they feel trapped. Too many choices, and they feel overwhelmed. The trick is to find that optimal number of choices, and make that what is presented visually to the user for any single operation.
What you would have is a logical hierarchy, so that the optimal number of most popular choices is presented up front, including a “More” option that links to other, possibly more obscure choices. Since those are hidden, they aren’t annoying. But they’re still there for those that really want them.
Gnome and KDE both do this to a certain extent, but they aren’t really consistent about it, and there’s no assurance that they’ve found the “sweet spot” for what they make visible.
>The problem is simple. Computers are complex tools. They are the most complex tools ever, because they can do almost everything. Complex tools require complex knowledge, there is no other way. You can achieve some very basic task without this knowledge, ok, but this is not using a computer.<
ahh, in a perfect world……..
If this were the case there would be a lot of IT and helpdesk workers out of jobs lol
People like the option of choice but don’t necessarily like to be forced to make choices. Apple had a good idea but a bad implementation with the simple/expert variants of the Finder but even that was a choice users had to make, a choice that led to still more choices. How to solve this issue remains a question but the question itself is nothing new.
Apple’s technique of hiding some functionality with an option key is a good approach. More experienced users can still use more complex functionality, and beginners aren’t confused by the extra options.
That said, in terms of programming, it’s always good to eliminate superflous options to reduce a program’s complexity, thus eliminate chances for bugs and user confusion.
Recently I tried to use a printer (Epson RX520 in case you’re wondering). It wasn’t supported by default in OS X so I had to download the driver.
Epsons driver interfaces are usually pretty good, the interfaces are fairly simple and more complex options are there if you need them.
If you want to print a photo, you drop in some photo paper, select “photo paper” then press print, it will usually give you a very respectable result.
This driver on the other hand looks like a quick hack, it gives you a whole pile of random options and you’re left scratching your head. Any attempts at printing so far have failed utterly.
The interface really is that bad it makes the printer completely useless for printing photos. Thankfully the driver on the PC is rather better and it produces fine pics, I’ve given up printing from OS X though.
I may be able to get the perfect image out of this driver but I really don’t have the inclination (or paper or ink for that matter) to sit around for hours working out how to get it.
That’s the impact too much complexity in an interface can have.
First off before we can address the issue we should correct his count. Joel states:
“you have to choose between nine, count them, nine options: two icons and seven menu”
“there are also four FN+Key”
“there’s an on-off button, 14, and you can close the lid, 15. A total of fifteen different ways to shut down a laptop”
No, there are 7 options (choices) and 15 ways to activate those choices. All the ways stated accomplish 7 things.
Having hot-key and mouse methods to activate common options is not unreasonable (try shutting down your machine if the mouse is not working without hot-keys).
The two questions that really need to be answered are:
1. Are all these options necessary?
2. Is the UI effective in its presentation of the options (are the number of options available in each input method reasonable)?
My feeling on number one is that each of the options is needed. Therefore, I do not consider there to be too many options (in this case).
I have not used Vista so I will only comment on the screenshot he shows in the article. It appears that the menu he shows is apparently visible(expanded) if the 2 “quick options” are not adequate. If you want to shutdown just click the shutdown icon (circle with a line through it) and don’t expand the secondary menu. The menu is only there if you need more options.
It seems he is going out of his way to HAVE MORE OPTIONS and then complaining about too many options.
leads to a confused “debate”.
First, the article is never clear about what sort of “choice” too much of is bad. It swings back and forth between choice of effectors and choice of actions and even goes so far as to intermix the two kinds of choice as if they were the same choice.
Second, it violates the principle of least astonishment, which it probably should have discussed in this context, because legacy plus least astonishment is the real culprit behind the proliferation of effectors to implement the same choice.
Finally, it doesn’t talk at all about workflow and how that reflects on the plethora of actions or on the choice of appropriate effectors.
When I’m using a stylus, I don’t want to disturb my workflow to go type a hot key. I want a gesture effector to implement my choice. On the other hand, when I’m typing away on a manuscript, I don’t want to disturb my workflow by groping around on my (real) desktop to find the mouse. I want a key sequence effector to implement my choice.
Those workflow related differences have nothing to do with ‘excess choice’ or ‘naive user versus power user’ and everything to do with different workflows requiring different effectors for efficiency.
There’s no single ‘pedals and steering wheel’ input model that will work for computing because there’s no single workflow in using computers.
The best interfaces do not violate the principle of least astonishment. If ‘point and click’ makes sense as a way of effecting a choice, then there should be a point and click way — that’s very similar to that used for similar choices. Likewise a keystroke interface, and if I’m lucky enough to have an input device capable of gesture recognition, a gesture interface.
As far as the example of turning off a computer, it hasn’t been an on/off decision in the technology in a couple of decades, what with monitor and drive power down being separate from processor and memory power down for power saving. The bigger the budget, or the smaller the battery, the more devices can be independently powered, and you do not want to be constantly deciding on your own what part of the system to have powered up or down at any one time.
Aside: the smartphone I’m currently working on has over 150 different components that can be powered up somewhat independently. Do you really want to decide which of 75 of those 150 power buttons to push in order to use the bluetooth to make a handsfree call? Or would you rather have all-on/all-off and a battery life of about 1% of what it is now?
I get and agree with everything you say except for your last comments about turning computers/cell phones on/off. It sounds like you are arguing for exactly the same thing the author is arguing for–more sophisticated automatic power management, so the user doesn’t have to think as much about it.
It’s an “aside” becuase it is only indirectly related to the topic, but you got it just fine.
I should have either left it out or finished the thought.
The underlying point is that there are circumstances where letting the computer do the book keeping make sense, so let the computer do that when it can and that power management isn’t as simple as on/off.
Take my bluetooth example. If you have bluetooth enabled cellphone, then you should have a way to tell the software “I am not ever using bluetooth, disable it.” But you shouldn’t have to figure out that ‘disable’ means stop this demon, turn off that radio and the uart that is used to talk to it, and so forth.
Further, if you do have bluetooth enabled and are paired with a headset, you shouldn’t have to do something to explicitly wake up the bluetooth radio when you get an incoming phone call. The software should do the book keeping to figure out that it has to wake up the receiver part of the bluetooth radio long enough to see if you push the button on your bluetooth device.
Well, I guess he’s forgetting other ways to shut the machine off, like using a jackhammer, shotgun, etc.
Now seriuosly, does he honestly think people will go out and buy expensive SSD-/Flash-/nvRAM-/whatever-term-they-come-up-with-next-for-the-same-t hing-based HDDs just to make the shutdown easier? Jeez, some people are happy when the damn OS actually boots (and in Windows’ case, amazed ). Deal with it.
Solution: a button which says ‘end’ (not found where the ‘start’ button is, because otherwise newbies are confused) which brings up a menu of the following options:
1) turn computer off
2) log in as another person
‘restart’ is not needed because anyone can turn the computer on when it is off.
Another button labelled ‘pause’ does the hibernation.
Coming back to the iPod comparison from the article (not that i think computers and iPods are the same complexity class of hardware) i actually tried to explain the usage of an iPod to my grandma so she could listen to her operas without torturing her neighbors and, guess what, she did not understand it. She did not understand how such an click wheel could do anything if it doesnt has physical buttons or physical turning. She did not understand how to navigate graphical menues and overall had much trouble with that thing.
So obvious iPods are to complex yet… we need an player with exactly three big buttons “play something (i don’t care what)”, “stop play something” and “who the hell am i and how did i get here”.
Analog to that we need Computers with three buttons “start and give me that MS Word thing”, “stop that word thing and shut down” and “who and where the hell am i and what is that strange thing on my desk”.
———————–
So obvious iPods are to complex yet… we need an player with exactly three big buttons “play something (i don’t care what)”, “stop play something” and “who the hell am i and how did i get here”.
Analog to that we need Computers with three buttons “start and give me that MS Word thing”, “stop that word thing and shut down” and “who and where the hell am i and what is that strange thing on my desk”.
———————–
Actually, I think touch screen technology is the answer to this problem. In the future there should be no physical buttons, just a screen that you can touch and select options with.
I disagree. I always see older people standing helpless in front of our newer subway ticket mashines which are indeed touchscreen based.
i actually tried to explain the usage of an iPod to my grandma so she could listen to her operas without torturing her neighbors and, guess what, she did not understand it.
The problem is your grandma or i could just say as well my grandma don’t have experienced the initial design consistency.Once you have learned rather trivial things like a joystick,mouse,keyboard a multi-functional knob such as the one you described isn’t a big hurdle to take .In fact the knob may seem logical.
What is a simplification for one person may pose another abstraction level for another.
I don’t think there’s anything at all wrong with having choices; the problem arises when all choices are thrown at the user at once.
The trick is to let there be one obvious path for the uncaring/uneducated user to follow. For example, in this case there should be two obvious choices:
1. Shut down my computer
2. Switch to a different user
Make these obvious and easy to access buttons so that the computer illiterate user can easily find them and identify them for what they are. Provide the ability to perform all 9+ functions in a separate area that the more computer literate person would have access to. Even better, let the computer literate customize the UI location and accessibility of these options.
… that people *need* choices. Language seems to always offer different varieties for the same meaning. Oracles always say: you have a choice.
One mans’s meat is another man’s poison. Those functionalities are not invented for nothing, and I personally found a use for *all* of them.
Man is founded on choices. Defaults are what really matter. But, like the matrix, people are surely going to destroy anything that restricts them. Its only a matter of time.
I have a phone that can play videos.
At first, it was cool because it was easy. Just push the button and it plays. Push the button again and it stops. It was cool until I wanted to continue to watch the movie I stopped. I had to start from the begenning again (when I just had 15 minutes left to watch).
Hibernate is not suspend is not log off is not power off. You are iritated by choice until you need it.
Desktop/Tower
Remove power cord from back of machine.
Laptop/Notebook
Remove battery from said laptop/notebook.
See, wasn’t that easy?
:-p
It seems like we have a lot of proponents of choice here, even in such a case as shutting a computer down. While I like choice myself, I dislike *unnecessary* choice, and I consider this choice unnecessary. To resolve this issue, I have to ask the following questions:
* Why would you want to restart your computer? Booting another OS doesn’t count, as a computer should have *one* OS that does everything. That’s the purpose of an operating system. Rebooting to have changes take effect doesn’t count either, as an operating system shouldn’t need this (in the seldom case where it does anyway, about once a year at most, one can as well turn it on again).
* Why would you want to shut down your computer instead of hibernating it? Crashes and memory leaks again don’t count, since these problems should be fixed where they occur and not with this kind of band-aid.
* Why do you need “log off” and “switch user”? What’s the useful difference between these?
* Why must there be a difference between “hibernate” and “sleep”? With a well-designed computer (including hardware), taking off from hibernate doesn’t take more than a few seconds. Who would want to put a computer to sleep and have it wake up in less than a few seconds?
Without a convincing answer, these questions leave me with only three options:
– shut down / hibernate
– log off / switch user (only useful in a multiuser system anyway)
– lock (= I don’t want another user to log on while I’m away)
Booting another OS doesn’t count, as a computer should have *one* OS that does everything.
Ok, if you’re so ignorant to understand why it isn’t always so, I don’t know what to tell you.
Why would you want to shut down your computer instead of hibernating it? Crashes and memory leaks again don’t count, since these problems should be fixed where they occur and not with this kind of band-aid.
But it’s a damn useful band aid: in the meantime, go on and correct *all* memory leaks/bugs of all softeware in this world, we will be much grateful.
Even if you are superhuman and do it, maybe I want to leave no trace of what I was doing on that computer. Did you consider it?
Why do you need “log off” and “switch user”? What’s the useful difference between these?
An example. If I’m in a corporate enviroment, something bad happens from my PC and logs shows I was logged in –> they can suspect me (even if I was somewhere else) and the like –> BAD
I wasn’t logged in –> GOOD.
> Ok, if you’re so ignorant to understand why it isn’t always so, I don’t
> know what to tell you.
Nice to see some real arguments… Instead, you could name real-world systems that boot into different OSes time and again *other* than the crappy PCs which are more tinkering than engineering.
> But it’s a damn useful band aid: in the meantime, go on and correct
> *all* memory leaks/bugs of all softeware in this world, we will be much
> grateful.
> Even if you are superhuman and do it, maybe I want to leave no trace of
> what I was doing on that computer. Did you consider it?
The original Unix could kill processes without leaving memory around and without leaving traces of what you did (except files that were created, but these won’t disappear with rebooting either). There’s nothing superhuman in pointing out that programmers were already able to do it right decades ago.
> An example. If I’m in a corporate enviroment, something bad
> happens from my PC and logs shows I was logged in –> they can
> suspect me (even if I was somewhere else) and the like –> BAD
> I wasn’t logged in –> GOOD.
Finally, a good point.
Sleep and hibernate tell you nothing. Now, if sleep said what it did by the option, and so did hibernate, the user would be better informed to start using it.
As a desktop PC user, I use every option but sleep. and like having the options.
However, if the interface just said what each option did, having them would no longer be a problem–we can have our cake and the new users could eat it, too.
So, let’s see:
* Switch User: leave the current user on and running. Unix-like thinking.
* Log off: quit all your stuff
* Lock: keep people from seeing your stuff
* Sleep: suspend with quick recovery, but some power use
* Hibernate: suspend with long recovery, but no power use
* Reboot: obvious
* Shut down: obvious
Now, I don’t disagree that just showing a list of a bunch of options with no extra information is bad; but I think that adding information is a better way to handle it than removing choices.
I’m also a freak about Fitts, so adding info would make bigger targets, which woud be good, too, since data density isn’t a problem for such a menu.