On Thu, 2006-08-10 at 22:51 +0200, cedric GEMY wrote: > may be i misunderstood your actual specification, but when you define > modifiers, i only talk about multiselection modes. What about old > options with Ctrl and Shift that does the same as "Expand from center" > and "Make Square". > In fact it may confuse the user, because these options have to be > checked before the selection is drawn. > I agree with Sven on this : keep the old way, because users are > accustomed to it. I don't think that we absolutely need to continue doing things the old way. Users get accustomed to new behaviour quite fast, as long as it offers everything they need. I agree though that it is important to make "Expand from Center" and "Make Square" more easily accessible. I believe that we need to use modifier keys for this, quite similar to the way the old tools used to work. The old implementation was however rather unintuitive and hard to discover, mainly because pressing the modifiers didn't change the selection immidiately, you had to move the mouse first. This is due to a limitation in the tools code and cannot be fixed in the selection tools alone. It needs some changes to the tools infrastructure. But I think we need to do those changes now, simply because we can't implement a satisfying behaviour without them. Mitch and me discussed this with Peter a while ago and our conclusion was that it makes sense to continue to use the modifier keys pressed when creating the selection to define whether the selection is replaced, added, subtracted or intersected. Then, while creating the selection, the modifier keys can be reused to add constraints such as "expand from center" and "make square". But in order to make this behaviour understandable by the user, it is important that there is immidiate feedback, unlike with the old tools. Pressing Ctrl should immidiately center the selection on the starting point, releasing the modifier key should restore it. Same for the Shift key. If you press it, the selection outline needs to change to a square/circle immidiately. Sven _______________________________________________ Gimp-developer mailing list Gimp-developer@xxxxxxxxxxxxxxxxxxxxxx https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
