In order to demonstrate our superior intellect, we will now ask you a question you cannot answer.


During the development of Windows 95, a placeholder dialog was added with the title, “In order to demonstrate our superior intellect, we will now ask you a question you cannot answer.” The dialog itself asked a technical question that you need a brain the size of a planet in order to answer. (Okay, your brain didn’t need to be quite that big.)

Of course, there was no intention of shipping Windows 95 with such a dialog. The dialog was there only until other infrastructure became available, permitting the system to answer the question automatically.

But when I saw that dialog, I was enlightened. As programmers, we often find ourselves unsure what to do next, and we say, “Well, to play it safe, I’ll just ask the user what they want to do. I’m sure they’ll make the right decision.”

Except that they don’t. The default answer to every dialog box is Cancel. If you ask the user a technical question, odds are they they’re just going to stare at it blankly for a while, then try to cancel out of it. The lesson they’ve learned is “Computers are hard to use.”

Even Eric Raymond has discovered this. (Don’t forget to read his follow-up.)

So don’t ask questions the user can’t answer. It doesn’t get you anywhere and it just frustrates the user.

Comments (54)
  1. But the real question is, what was the question the dialog asked?

  2. Mike Dunn says:

    What is the air-speed velocity of an unladen swallow?

    [African] [European] [Cancel]

  3. James Day says:

    Another handy tip. When you ask the user a security-related question which they have told you to ask them, don’t tell them "things may break" when they refuse permission.

    IE6 does this when you set running ActiveX to prompt. Say no and it harasses you with a warning that the page may not display as expected. You already knew that when you chose to refuse permission. If you didn’t know it, the correct time to say so is when you ask the question, not when it’s too late to do anything about it.

    IE4 doens’t pester you for choosing the safer route, but it is a little dated now.:)

  4. mike says:

    I note that Raymond carries on bashing Windows just for the fun of it:

    … the old Windows box. It blue-screened a lot …

    Yup, Win98 did do that now and again.

    It’s also interesting to hear a seasoned developer express the kind of frustration that the average user experiences frequently. One of the difficulties for developers, I think, is that they have a really hard time putting themselves into the position of the person who has no clue what’s actually going on and needs some help. But lo and behold, everyone is a newbie at something, and it’s good to have these reminders now and again of what that feels like …

  5. Raymond Chen says:

    ActiveX warning: Well except that lots of people never see that they changed the setting. They just go to the Security page and whack the "master slider".

    I suspect the dialog was added after lots of people complained that "Web pages aren’t working; stupid IE can’t browse the web what a pile of dung", unaware that at some point in the past they had cranked the security level to High.

    But it annoys me too.

  6. tekumse says:

    Since no other browser uses ActiveX – it’s hard to imagine waht other browser will be working fine on such pages.

  7. Anonymous Coward says:

    If you are going to complain about UNIX, do it properly! The Unix Haters Handbook has all the gory details: http://web.mit.edu/~simsong/www/ugh.pdf

    Funnily enough it doesn’t cover printing. In general many UNIX/Linux apps don’t have the ability to print, although recent ones are way better.

    [When writing up technical specs I always prided myself in how many different sections of the UHH I could quote :-]

    I also wish all programmers who went anywhere near a user interface were forced to read stuff by Alan Cooper first. Here is an excellenet first article: http://www.cooper.com/articles/art_goal_directed_design.htm

  8. I’ve seen so many dialog boxes with "Ok, Cancel" in them and a message reading "Click Ok for Yes, Cancel for No".

    Drives me insane.

  9. Mat says:

    Sure would be nice if there were more "rich" dialog boxes. Have a simple y/n, but provide pertinent info in the box as well. Like:

    ====================================

    Do this thing?

    [yes] [no]

    This thing is generally a basic action. If you do this, that will happen. That serves as a method for completing the task of creating a file and other stuff.

    Would you like more information? [help]

    ========================================

  10. Mike Dunn says:

    There’s one thing that I REALLY miss in recent apps (notably MS Office XP) – flyby help in the status bar as you’re going through the menus.

  11. (6) says:

    I recently was asked to fix a bug in our DHTML client. The bug went something like "In the Windows version the dialog says ‘Yes’ and ‘No’ but in the DHTML version it says ‘OK’ and ‘Cancel’ (with a horrible click OK for yes style message)". Of course this is impossible to fix down to IE (and Netscape) not allowing us to change the buttons.

    So although the ‘OK for Yes’ message is horrible (and confusing) it is sometimes unavoidable.

    Roughly on topic, how nice would it be for MessageBox to offer a ‘Don’t show this again’ functionality. That would then fix that annoying Active-X popup and save developers so much time (grr… it hate that Active-X dialog so much… :) )

  12. (6) says:

    Just to clarify/spam, by offering the functionality I meant present the ‘Don’t show this again’ option, rather than offer the actual functionality to prevent the dialog being displayed (we’ve gotta keep non-Microsoft employees in a job too you know :p ).

  13. Jack Mathews says:

    Mat: That kind of information makes people click the default button even faster. More text is scary to people.

  14. Jack Mathews says:

    Mat:

    Dead wrong. You’re out of touch with real users.

    a) Wizard’s don’t work. People NEXT through them as fast as they possible, never ever looking back.

    b) If people read boxes, they would take the time to check "I never want to see this again." I can’t count the number of times I’ve had people bitch about a box coming up over and over, then I just check the box for them. Luckily, Microsoft does this by default in IE 6.

    >> What percentage of the time do you educatedly guess at a dialog when you first think you need help, but then think "help" isn’t helpful, so you guess? <<

    Less than 1%.

    >> The percent of that vs. actually going to help <<

    Never. The only time I go to help is in Visual C++. How often does my mom/dad/sister/aunt/uncle go to help. NEVER.

    Dialogs just do not work. Plain and simple. People do not ever read them unless they KNOW an action they’re doing is going to pull one up.

    "I’m saving a file, do I want to overwrite? Oh, I expected that, so I’ll read it. "

    "I’m navigating to a web page. What, a box? Screw that, I just want to get on with my work. If something was wrong, I’ll figure it out in a second"

  15. Mike Dunn says:

    The first rule of UI is: Users don’t read.

    The second rule of UI is: Users don’t read.

    Even if we did make all text in all dialogs 100% understandable by 100% of users, the non-techy people have already been conditioned to think that computers are scary and dialogs are scary and more text won’t change that conditioning. Cutesy dancing paper clips won’t change that conditioning either. ;)

    Not to mention that adding more "help" annoys the techy people (see dancing paper clip).

  16. I’ve been working in Unix and a Unix bigot for 17 years now (currently banished to Windows development) and I gotta tell ya "UNIX PRINTING IS WRONG, WRONG, WRONG." It’s never been good. In fact, it’s never been horrible — it’s two levels worse than horrible. And for most of it’s history there’s been two completely incompatable methods of horror: SYSV and BSD printing.*

    A team I was on designed some oversimplified printing methods at a Large US Auto Manufacturer that worked well. It was to the point where it was *almost* as friendly as Windows print setup. But it required the Systems group to maintain large catalogs of metadata about our network printers that were constantly out of date.

    Fundamentally, to get printing right there’s got to be more cooperation between systems vendors (IBM, SGI, HP, etc..) and with printing vendors. And it ain’t gonna happen in our lifetimes.

    When idea of CUPS (what ESR is ranting about) came along in the late 90’s it sounded *great*! But the implementation was still by programmers, for sysadmins. Apparently it hasn’t gotten any better. I think the CUPS developers had the deck stacked against them — and they didn’t help matters much either.

    * There’s just some things that Unix got wrong: Printing is universally wrong, the API and implementation of X-Windows is bad (the *idea* is right though), and file sharing (nobody has gotten this right yet). I guess admitting this may get my Unix Bigot Membership revoked.

  17. Robert Morris says:

    IE4 doens’t pester you for choosing the safer route [re: Active X],

    >but it is a little dated now.:)

    And IE 6 isn’t? ::ducks and runs::

    :-)

  18. Raymond Chen says:

    I have seen people click Cancel without even reading the message. I discussed this phenomenon last year.

    http://weblogs.asp.net/oldnewthing/archive/2003/09/01/54734.aspx

  19. Mat says:

    No, what’s scary is having to go to help, and hope that what they read is there. It can be done correctly, and it can be done very wrong. Ask yourself this:

    What percentage of the time do you educatedly guess at a dialog when you first think you need help, but then think "help" isn’t helpful, so you guess?

    The percent of that vs. actually going to help — I have to think more often we all guess. If an explanation is offered in *plain English*, it can be very helpful in getting the user to read the pertinant info.

    Think micro-wizard. What’s the one sentence you tell your mom over the phone to explain it? (I say mom cuz she’s the perfect example of forcing yourself to use friendly words.) Use that sentence as text in addition to a normal dialog box, and voila! A helpful dialog!

  20. Edward says:

    I’m curious to see what you guys think of this

    http://www.specopslabs.com/david_technology.htm

    Do ya think these guys can reimplement the Win32 subsystem in 6-8 weeks? Or are they just trying to drum up some investment capital before quietly disappearing?

    Seems a bit unlikely that they actually have anything implementable that quickly.

  21. Chris says:

    Personally, I think an alert/UI model should mirror an incredibly good secretary–very competent, organized, respectful and, above all, completely transparent about what they’re doing.

    So in a lot of these alert cases, I’d prefer some kind of a countdown automatic action. For example, if you select a bunch of folders and click delete, instead of getting the "are you sure you want to do this" you might get "Deleting files permanently in 3(…2..1) seconds) with an "Delete Now" and "Cancel" button.

    Or in many cases just "I’m doing this now…" message visible for a few seconds with just an <Undo> button on it.

    With this solution, designers can cover their liabilities but the user doesn’t need to decide. Of course, no developer ever wants to implement this because it’s an artificial deterrant to performance and a non-standard alert so you’ll probably never see it in a product.

  22. jojjo says:

    Actually, Mat, all the UI research is against you on this: The more information that is presented to the user, the less likely it is that the user will read it – no matter how easy to understand the text is.

    Why do you think licence agreements read "Bla bla bla bla…" for several pages instead of just "You give us all your rights, we give you none."

  23. Jeremy Morton says:

    Mat, it’s not just "thick-headed" users; whenever I sit and watch my girlfriend do something on her computer, I’m constently seeing her click close or cancel on dialog boxes or pop up ballons as soon as they appear. She’s definitely not a novice computer user. I usually ask her if she knows what it said, and she says no. She doesn’t care, she’s trying to do something, and they get in her way.

    Watching my parents is the same. Trying to help them on the phone is painful because they’ll say that something didn’t work, and when I ask what happened, they’ll say that something popped up and they clicked the button. It’s very difficult, even when they’re trying to get me to help them, for them to stop and read the dialog. I have to ask specifically what it says.

    As strange as it may seem to "us", dialog boxes are considered obsticals by many users, not because of their content, but because they pop-up and stop them from what they’re trying to do.

  24. Mat says:

    "People do not ever read them unless they KNOW an action they’re doing is going to pull one up."

    So basically what you’re saying is that people are going to ignore a dialog box if they are confused? I wholeheartedly disagree. Users don’t become confused *until* they don’t understand what they read.

    I’ll grant you this; design should intentionally avoid dialog boxes that give functional choices… those should be done within the program. However, I disagree on the point that users get confused because something pops up.

    Users don’t start dissing a pop-up until it’s:

    a) confusing text

    b) repetitive

    c) options don’t clearly match text question (ok for yes, cancel for no)

    At this point, users have the right to get annoyed.

  25. Dan Maas says:

    I got a bit of that today when I had to use my dad’s computer to send an email… Norton and RealPlayer kept getting in my face ("Update Me! Update Me!"). There kinds of obtrusive messages are probably OK if that’s the only application you are running, but if you are trying to do anything important and time-critical, they’re a real apin.

  26. Mat says:

    I know, Raymond — I’ve seen it and done it, too. But that’s not what I’m getting at. If a user is thick-headed enough to cancel something new, that’s at their own peril. My point is that by having well-crafted, well-written information for people that *do* read to make a decision, this will solve the middle 60%. 20% "geeks" will figure it out, 20% are "stupid" and will skip. I’m saying that for those in the middle, if it’s not: "tech tech techie tech", most people figure it out.

    A good example would be taxes… 1040EZ, no problem. But get to a 1040, and maybe a schedule or two, and people go running to H&R Block. IMO, it’s because:

    1) It’s confusing enough that the /possibility/ of an error is higher, and

    2) Errors are costly, and the entity that files is untimately responsible (individual or tax co.)

    Back when I was a student, I didn’t mind filling out 10 lines on one sheet.

    Same thing with a program. If questions are presented in such a way that you can easily be *confident* of your answer, 80% of people will tackle it. Only 20% will offer to try and *learn*.

    But to bolster the overall argument that dialog boxes should be avoided, TurboTax on the web… no dialog boxes I can remember. =)

  27. Mario Goebbels says:

    <<So don’t ask questions the user can’t answer. It doesn’t get you anywhere and it just frustrates the user.>>

    The problem I have with that design philosophy is that I might miss on questions that I’d like to answer, but aren’t asked because the system tries to accomodate the way larger share of less literate users.

    That’s a reason why I’d like to see a basic and expert mode in Windows. However trying to implement that adds a lot of additional (overhead?) code to the system.

  28. Ben Hutchings says:

    How about disabling all buttons on a message-box until the user has read the message (deteced using a mandatory eye-tracking peripheral)? (OK, so there needs to be an alternate behaviour for blind users.)

  29. Mat says:

    Just curious where I can see some of the research. I have no doubt you all speak the gospel… it just bugs me the wrong way. I guess I’ll be a contrarian in usability for now.

    What’s everyone’s take on road signs? You’re required to read those, yet we breeze through them at times. When is the last time you read a sign on your drive home from work? I wonder what that says about click-thru types…

    Anyhow, I’m not trying to be a blasphemur, just challenging the status quo of everyone ignores.

  30. Scott says:

    FireFox has some brilliant examples of bad dialogs when installing extensions. Apparently you can only put up OK/Cancel dialogs, so you get wonderful things like "Press OK to put the extension in your personal directory, press Cancel to put the extension in the application directory."

  31. Joel Spolsky wrote an interesting book that touches on this subject called, "User Interface Design for Programmers". He has the first 9 chapters online here: http://www.joelonsoftware.com/navLinks/fog0000000247.html

    Specifically, chapter 3 speaks about dialog boxes that force the user to answer a question that they should never be asked:

    http://www.joelonsoftware.com/uibook/chapters/fog0000000059.html

    It is a very interesting article.

  32. Mat says:

    Oh I like this:

    "The problem comes when you ask them to make a choice that they don’t care about."

    Thus I assert that people generally care until confused. Then they don’t give a flying hoot.

    Thanks, Jason! I’ve seen Joel’s schpiel on globalization, too. That was very helpful in my current work. Good stuff!

  33. JD on MX says:

    When in doubt, dither: A Microsoft staffer, who apparently doesn’t have a name on his blog but who may work in Longhorn UI, notes how the natural tendency is to click "Cancel" if you’re not sure what a dialog is…

  34. My current favorite UI thingie that people are going to ignore is the IE pop-up blocker.

    Basically it’s a tooltip that grows down from the IE toolbar when a popup is blocked.

    It’s presence is obvious (it’s a honking big yellow bar after all), but you don’t need to interact with it. It’s clickable – clicking on it brings up a context menu, so it’s functionality is easily discoverable, but it doesn’t interfere with your experience.

    That’s good design imho.

  35. jeffdav says:

    Larry- I was wondering when someone would mention the modeless prompt UI. We actually designed it with all of this stuff in mind. Security decisions while browsing the web have become too important to leave them in the hands of the user. So whenever IE detects something significantly dangerous is happening it choose to block on the user behalf, and displays the Security Bar. Users can use the context menu to re-enable things. We have also redesigned dialogs that have security impact to be more uniform. For example, they all have one of the little shield icons on them, with different primary colors indicating different levels of security dangerousness.

    But will this help with the overall problem? Maybe. Will users still install evil controls just to get a page to load? Probably. But hopefully we have raised the bar a bit.

  36. Timdog says:

    As to CUPS (or Unix printing in general), one of the project managers at my current company, who used to work at Sun in a similar capacity, says that he once went to a meeting comparing NT to Solaris. They went through feature by feature comparison slide show.

    Under printing, it said simply: "They can."

    Tim

  37. mpt says:

    "Thus I assert that people generally care until confused."

    You can assert all you like, Mat, but it’s not true. I say that after three years of working in Internet cafes, watching people of various ages, ethnicities, and skill levels doing their utmost to get rid of alerts without reading them. People either habitually click "OK"/"Yes", no matter what the question, or they blindly click "Cancel"/"No", no matter what the question. (Ever wonder how so many computers get infected with spyware? It’s from the people who habitually click "OK"/"Yes".)

    There are various things you can do to increase the probability an alert will be read. Remove all buttons from the title bar. Remove all text from the title bar. Shorten the alert text. Put the alert text in bold. And if it’s a confirmation alert, use a meaningful label for the action button — like "Replace", or "Install" — rather than "OK". Details like those will help, but they won’t help much.

    Though it’s tempting to think so, people don’t suddenly become less patient (or less intelligent) when they sit in front of a computer. It’s just that they’re confronted with software that is uniformly badly designed. Any software that includes alerts in its interface is insufficiently advanced.

  38. Mat says:

    But what of your assertion? Paraphrasing, "People either habitually click yes, or blindly click no."

    Well, which is it?

    Look, I get the UI design "thing" — I’ve even assisted in our product’s UI at times. But there’s something here that doesn’t add up to me…

    "…people don’t suddenly become less patient (or less intelligent) when they sit in front of a computer. It’s just that they’re confronted with software that is uniformly badly designed."

    a) I totally disagree. People get much less patient. I don’t blame them, either. Just like when you lift the phone, and there’s a dial tone, it needs to ‘just work’. If it doesn’t on a consistent basis, people will give less benefit of the doubt. (see: computers) b)Additionally, "badly designed" based on what? At what level should things be designed at, for seven year olds? (see: AOL) I’m not saying that most people are "less intelligent", but I am suggesting that good design assumes a few things… how many times I’ve yelled RTFM after answering a support call back when I did that!

    Eventually, you find that making something too simple ends up restricting the functionality of a program. Good design can close the gap between easy-to-use and powerful, but eventually no matter what, there is a gap. Once you’re at that gap, it sounds as though you’re suggesting leaning towards the simple. I’m not saying that’s wrong, but I’m suggesting that it’s an out to just say "make it simple". Our MO has been make it simple until the user wants to play with dials and such. Use presets and smart settings to minimize a user’s need to manually adjust, until a user chooses to.

    I’m just trying to not let you get away with a blanket statement, mpt. =)

    [Edited by Raymond at request of original commenter.]

  39. "FireFox has some brilliant examples of bad dialogs when installing extensions. Apparently you can only put up OK/Cancel dialogs, so you get wonderful things like ‘Press OK to put the extension in your personal directory, press Cancel to put the extension in the application directory.’"

    Even worse, Firefox extensions are fairly evenly divided on which button means which. So you’re just as likely to see "Press OK to put the extension in the application directory, press Cancel to put the extension in your personal directory."

  40. mpt says:

    "’People either habitually click yes, or blindly click no.’ Well, which is it?"

    Some people did the former, some the latter. I never worked out a pattern to it. Maybe it could be correlated with their Jung-Myers-Briggs type.

    "’badly designed’ based on what?"

    Based on the psychological fact that for the large majority of the population, alerts don’t work. If a design relies on elements that are known not to work, that’s a bad design.

    "At what level should things be designed at, for seven year olds? (see: AOL)"

    No, I said "of various ages". That was a range from about 10 to 80 (though I noticed people aged 70+ were slightly more likely to read alerts). General-purpose software needn’t be designed for different age groups. It just needs to be designed for humans, rather than for some hypothetical species that usually pays attention to alerts …

    "how many times I’ve yelled RTFM after answering a support call back when I did that!"

    … some hypothetical species that not only reads alerts, but also reads manuals! Bizarre.

    "Eventually, you find that making something too simple ends up restricting the functionality of a program."

    No, I didn’t say "Any software that includes alerts in its interface is insufficiently simple". I said "Any software that includes alerts in its interface is insufficiently advanced".

    Finally, if either you or the other Mat really think "people do either become impatient or idiots (or both) when they’re sat in front of a computer", it should be fairly easy to demonstrate. Get 20 people. Get half of them to do an IQ test while sitting at a desk which has a computer on it; get the other half to do the same test sitting at the same desk with the computer removed. If you get significantly different results, your claim will be worth paying attention to (and we could start researching whether it’s caused by radiation from the monitor, or poisonous gases given off by the plastic case, or whatever). Until then, the null hypothesis must be that computers have no such effect, just like fridges and pianos and trees don’t.

  41. Mat Hall says:

    I agree with Mat (with a name like that, how could I not? :) — people do either become impatient or idiots (or both) when they’re sat in front of a computer, myself included. Whenever I’m installing software I tend to blindly click on "Next" and "I agree" without reading anything, purely because I’ve seen the same messages hundreds of times and can’t be bothered to read them again. More precisely, I click where I assume the "Next" and "I agree" buttons are, but if ever someone put a "Delete everything" button where I expect to see a "Yes", then my machine will get hosed. If this isn’t impatient (and stupid), I don’t know what is.

    It’s not just computers in the traditional sense that cause this blindness. We have an ATM near where I work that asks a *lot* of Yes/No questions before you can get any money from it, but the position of the buttons changes from question to question. I’ve lost count of how many times I clicked "No" instead of "Yes" and had to start over. Granted this UI design is obviously broken, but if I took the 2 seconds to read the screen it would save me th 40 seconds it takes to do the whole lot again. Again, my lack of patience lets me down.

    Perhaps if the ability to click on a button was removed, and you had to press one of two randomly selected keys to answer Yes or No, people would be forced to read them. Or maybe include those distorted numbers that you get when signing up to various web services… Just do something! You get paid for this sort of thing, so it’s your problem… :)

  42. Mat Hall says:

    When I say "less intelligent" I don’t mean in general, it’s just that the user’s ability to follow instructions, read things, and make sensible decisions *about what the computer is doing* seems to be greatly reduced. Obviously if you actually became an idiot, no-one would ever be able to code anything! (Although some of the code I’ve seen in the past does appear to have been written by some kind of moron…)

  43. Jeff, I was being slightly sarcastic when I described the IE securitybar as being something people will ignore.

    IMHO, it REALLY is an example of an EXCELLENT solution to this problem.

    I’ve been working in the Windows UI&Policy group for the past year and a half, so I’ve been getting a crash course in UI design – I used to think I understood what good UI was, but now I’ve gotten a whole new viewpoint on how to do good usable UI.

  44. Raymond Chen says:

    If you talk to somebody who works at an information desk you’ll find that they’ve already run this experiment many times over: It’s not just computers. People ignore unexpected information.

    For example, if there is a sign on the door that says, "XYZ is closed today", you can bet that people will walk on in and ask "Is XYZ open today?"

    "No, it’s closed today. Didn’t you see the sign on the door?"

    "Hm, yeah, now that you mention it, there was a sign on the door, but I didn’t read it."

  45. David says:

    I like the "save changes" alert. I NEED the "save changes" alert. Sometimes I choose "yes", and other times I choose "no".

    Haven’t you ever accidentally closed a document? I want to be protected from the bad consequences of doing that.

  46. Jim Davis says:

    "No, it’s closed today. Didn’t you see the sign on the door?"

    No kidding. When I worked helpdesk in college, we had a free-standing sign about four feet high and two feet wide that said, "THE VAX IS DOWN" that we’d place in the doorway when maintenance was being performed. People would move the sign, sit down at a terminal, and after failing to log in, ask if the VAX was down.

  47. mpt says:

    "…it’s just that the user’s ability to follow instructions, read things, and make sensible decisions *about what the computer is doing* seems to be greatly reduced."

    Aha, now we’re getting somewhere. So how could we test *that* hypothesis? We need to create some kind of non-computer situation that presents people with interruptions that are just as unwanted, and just as unclarifiable, as alerts are. That way we can tell whether it’s just a computer thing, or whether humans *never* follow instructions with those same qualities.

    Maybe we could sell a kind of notepaper that comes with its own security guard. Whenever you write on the paper, and then try to leave the room, the guard stops you and says "Save changes to what you’re writing?" He won’t explain further. All he says is "Save changes to what you’re writing?" And he WILL NOT let you leave until you answer. If you say "Yes", he lets you leave. If you say "No", he grabs your paper and eats it, *then* lets you leave.

    After you’ve gotten used to instinctively saying "Yes" whenever you get up, the guard might start requiring you to answer other unwanted questions too. Questions like: "Changes have been made that affect the global notepaper. Do you want to save those changes?". Or: "Do you want to save the clipboard?" Or: "Some of the changes you have made will only take effect next time you wake up. Do you want me to put you to sleep now?" (The last would be especially effective if the guard was twirling his baton.)

  48. mpt says:

    "I NEED the ‘save changes’ alert."

    No, you need a decent undo function.

  49. Users don’t like it when they can’t cancel.

  50. Many OEMs ship with recovery partitions.

  51. The tragedy of the advanced setting.

Comments are closed.