Re: Comments on dialog proposal

[Gregory Merchan <merchan phys lsu edu>]

Upon further consideration, I am in accord with Colin.

There is a way to do some things which do apply instantly, and we don't
need dialogs for them.

Example:
Setting the background doesn't need a dialog. It needs a box of chits
that can be dragged to the root window. This provides clear indication
that the effect is instant; the user actually did something to what the
change is applied.

I think this can be made general: DnD or activating a menu item or some
similar user action means instant action, otherwise nothing until an
OK-click. (whatever the label actually is)

To put this another way:

If it's in a dialog box, it doesn't do anything until the user says it
should.

A final semi-humorous note: (I hope)

Consider how hard it is to make your argument when the other person
interupts everytime you stop for air. Such dialogues are usually
unwelcome.

On Thu, 2001-09-06 at 07:18, colin z robertson wrote:
> On Wed, Sep 05, 2001 at 03:57:43PM -0700, Maciej Stachowiak wrote:
<snip>
> > The advantage to instant-apply is that it makes it really easy to
> > tweak multiple settings to taste. It also removes unnecessary
> > complexity - there is no real reason why it should take an extra
> > button press in addition to just changing the control to make the
> > setting take effect.
> 
> hmm. Personally I see an increase in complexity. As I have shown
> above, there will be situations in which we can instant-apply and
> situations in which we can't. Now a user is going to have to work out
> for every dialog individually whether it will instant-apply or not.
> 
> I'm fairly fearless about what I do in dialogs because I know that
> nothing will happen until I press a button. An extra button press is
> no great price to pay to give the user an impression of safety.
> 
> Imagine the following situation: A user who frequently changes the
> image on his desktop or the font in his documents (and sees that they
> apply instantly) finds himself in a situation where he has to change
> something like file permissions, for example, which is something that
> he barely understands but knows that it has important security
> implications. Let's assume that we've decided that file permissions
> shouldn't be applied instantly (because of the security implications).
> Our user may not realise that these settings won't be instantly
> applied (or perhaps won't be sure) and will be too scared to even
> touch any of the settings for fear of breaking something. Having a
> user be cautious about pressing an Apply button is one thing, having
> them too scared to touch anything is another.
> 
> > Imagine, for instance, a font selection drop down
> > in a word processing program's toolbar. It would be crazy if you had
> > to press an Apply or OK button to get that font selection to take
> > effect, wouldn't it?
> 
> The fact that it is in the toolbar rather than in a dialog conveys to
> me a message that it will instant-apply. The Apply button in a dialog
> serves two purposes: Applying the changes, and allowing the user to
> delay applying the changes until they want to.
> 
> colin
> 
>   _____________________________                            ____
>   rtnl  http://rational.cjb.net     c z robertson ndirect co uk
>                                                    icq 13294163