Sometimes when a window is bring to front by the system, I click a button or other widget accidentally when my intent is to click in the window which was covered by the new window. For example suppose a critical notification pops up with some action buttons. At the same time I start clicking the area where a button will be drawn. The result is that I click the button by mistake. If the button said "dismiss" probably the notification is dismissed without confirmation and lost for ever! It seems unlikly but I come across with this kind of situations some times and are really annoying.

The same applies to dynamic layouts which update by themselves.

I wonder why this is not even discussed when there is a simple and obvious solution:

A widget which is just drawn (not redrawn) should not receive user input for some milliseconds. This is so because user need some time to recognize the widget before click it so if it is clicked just after drawn then it has to be by mistake (or the user has drank too much cofee)

I would like to receive some feedback on this thought.


1 Answer 1


This is a very well-observed question about an issue in interaction design which, as you rightly say, is seldom discussed. But it's still a very common problem.

The issues here were explored by some of the pioneers of interface design like Jef Raskin and Bruce Tognazzini, and relate to mode error, Fitts Law and habituation.

Mode error occurs in your example when you say "my intent is to click in the window which was covered by the new window". Fitt's Law predicts that you would suffer from missing the target of a button to an extent determined by its size and distance from your mouse pointer. Habituation predicts that you would not expect the popup to appear in the first place (or if you did, you would dismiss it without thinking about it), which might also compound these problems.

Your proposed remedy of freezing the mode in which the popup occurs might solve the mode error issue, and is in fact similar to that used by well-designed menu systems which have rules to prevent the menu disappearing when a mouse pointer wanders over their edges.

But it might be better to allow interaction with the background area and instead address the Fitts Law issue by showing it at the edge of the screen (where some of it might have an infinite target size). This is in fact the "growl" or "toast" model.

But I would say the ultimate solution would be not to have an overlay that can be dismissed as you describe. OK/Cancel popups suffer from the negative effects of habituation in being ignored, and the more that appear in a user's session the more likely they are to be so ignored. Far better to use undo. But that is another topic.

As an aside, I think your question shows that most designers today aren't thinking about these issues, and have certainly stopped trying to solve problems in interaction design, choosing to let them stand and proliferate instead. Worse, many don't even know such issues exist let alone how to solve for them.

Edit: I tried to write a brief welcome message at the top of this answer as I can see you are new. But for some reason it has been removed.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.