Fog Creek Software
Discussion Board

"Don't ask me again"

While I boadly agree that the "Don't ask me again" tick boxes are a good thing, in general, they fall just a bit short of being usable.

Say we have this permanent redirect dialog, and I tick the "Don't ask" box and press the "Yes" button to perform the bookmark update. What should happen in future?

"It should continue to perform updates automatically. You clicked Yes and asked it to stop asking you, indicating that you are happy with the autmatic bookmark updates."

"It should update the one bookmark but then not perform any more. By clicking this tick box, you asked it to stop doing this thing."

Both arguments seem valid.

Perhaps this is more of a rant about poor UI. The words "Remember this decision" would be better, or perhaps two tick boxes.

Any opinions please?

Bill Godfrey
Friday, June 20, 2003

Bad wording can certainly break the usability of a feature. Another example is the installation program that looks for existing files that are read-only, a later version or in a different language. If the user chooses "yes to all" when one of these is found, does that answer yes to all instances of all types of overwrite prompts, or just the specific type?

On the redirection feature, it seems to me the choice is only relevant for power users. Most of the people here, including myself, could not think of any reason why they wouldn't want to redirect. Seems to be it would be better to turn it on by default and hide an option away deep inside the power user options.

Many options happen when two or more developers discuss a feature and they can't agree on how it should work, so they diplomatically enter a compromise by putting the burden of the choice on the user. I suspect that is what happened with the redirection feature.

Also, this is a somewhat "radical" new feature, and what if they hadn't thought of all the possible side effects? By letting the user make the choice, the designers won't have to think too hard about it, and can move on to other, more fun issues.

Big B
Friday, June 20, 2003

It should apply the answer you gave to all future cases.

If you say "Don't do this" and "Don't ask me again", it should never do it again.

If you say "Go ahead, do it" and "Don't ask me again", it should always do it without asking.

Roel Schroeven
Friday, June 20, 2003

If any piece of software that uses this technique had a list of all the dialogs that used it[0], and let you tick or untick the 'don't show this' box for each one, then this would be a non issue.

But none of them do this, so... :-(

[0] -- Or, at the very least, a list of those dialog boxes that use this feature and have been shown, this being slightly easier to program and not requiring a massive list of all the dialogs in the program.

Friday, June 20, 2003

Our solution to this problem was to create a "persistent decision" message box, which looks something like this:

Do you want to automatically update some other stuff because you just changed this stuff?

<Yes>  <No>  <Always>  <Never>

Friday, June 20, 2003

"<Yes>  <No>  <Always>  <Never> "

<g> not wanting to sound critical, but you've just turned a single checkbox into 4 non-standard and possibly confusing options.

I think I preferred the checkbox :)

Saturday, June 21, 2003

I like the recent Microsoft approach(!). Please don't flame me for not knowing where they stole it from ;-)

Just ask the yes / no question. They second or third time they give the same answer, follow up with the "you keep answering the same, do you want me to stop bugging you?" question. If they say yes tell them how to reverse their decision later (the initial message can tell them the can change their mind later).

This gets round the problem in a simple and easy to understand way, and gives adds user satisfaction as it feels like the program is actually paying attention.

Oh and sorry I can't think of an example at the moment ;-)

Saturday, June 21, 2003

at the risk of being stubborn I _still_ think that the checkbox solution (with whatever wording) is a lot simpler, less intrusive for the user ("Im trying to print a file/copy a file/create a file, why is it asking me this now?"), and an all around better solution.

Saturday, June 21, 2003


The problem is that the check box is ambiguous, unless it's a confirmation message.  "Don't show me this again" makes sense if I'm asking you, "Are you sure you want to do this?"

If it's a decision message, "Would you like to also add the widgets from the other list?", what will the result be the next time I perform the action that displayed the message?

I would have to argue that:

  <Yes>  <No>

  x Don't show me this again

Is quite a bit more ambiguous than Yes/No/Always/Never.  As someone else pointed out, which option will be indicated the next time I perform the action if I check the box?  The button I click?  Whatever the system feels like?  There is no clear, explicit implication as to what the result will be.

Following Microsoft's inductive design guidelines, each user interface component should have a clear, unabiguous meaning, and there should be no question as to what the result of your action will be.

That being said, I think SteveM's point is a good one.  Unfortunately, to implement the feature he outlined takes quite a bit more effort from a design standpoint than our persistent decision approach.

Saturday, June 21, 2003

how about:

Yes    No

[]  Assume yes in future

Where yes is the developers best guess as to the likely preferred position of the user.
Something like that is still much less intrusive and confusing than a series of dialogs, or 4 buttons on one dialog.

Its important to remember that 9 in 10 times the user _really_ is not going to care much.  They are trying to create a birthday card for their grandpa pete and have just tried to (for instance) d & d a jpeg file from the desktop onto it.

the _last_ thing they care about is whether or not the program changes the stored name of the jpeg to reflect the taste of the developer. (for instance)

So the most important thing in this position is that we let them get on with what they are doing.  they may not _want_ to make a decision right then about whether the program should always do x,.

Saturday, June 21, 2003

"So the most important thing in this position is that we let them get on with what they are doing.  they may not _want_ to make a decision right then about whether the program should always do x,."

This is a good point. Often I want it to stop bugging me for a but, but am in a hurry don't want to bother with the instructions on how to change it later. Maybe there should be an standard option on the help menu "things I've asked it to remember, but might want to change" or a "go away for a day/week and then check again" checkbox,

Jack V
Saturday, June 21, 2003

How about "View->List of my decisions" ? :)

To show list of questions and answers you have set default for them.

Monday, June 23, 2003

"How about "View->List of my decisions" ? :)"

regardless of the actual dialog used to make these decisions I agree that something like this would be exceptionally useful on occasion.
The hard part of implementing it would be trying to explain where each decision was made...

In the dialog which displays requesting permission to update urls automatically you selected "yes always" [Change]

In the dialog which has a small icon on the left and shows itself when the application runs out of memory you selected "never lose your data" [change]

In the etc etc etc...

<g> would be a real blast....but I agree that the idea is a good one, it just beeds a good way to implement it without requiring long explanations like the one above.

Monday, June 23, 2003

*  Recent Topics

*  Fog Creek Home