[Gimp-developer] Default values for the Fixed: Size entry

[Martin Nordholts]

> suddenly I am in Kansas, on a very cool project for 2 weeks.

I hope you had a good vacation :)

>> How about having 100x100 as default when there is no pending  
>> rectangle and rectangle width x rectangle height when there is one?  
>> This would be the case for both the selection tools and the crop tool.
> 
> we cannot base the default of fixed size on the current rect,
> because it operates on the current rect. This would freeze up
> the rect in a positive feedback cycle.
> 
> we can base it on the image size, but that would make no
> sense for fixed size.

It seems as if quite a lot of people would like there to be a way of 
fixing the size of a pending rectangle.

If we handled the default values of Fixed: Size the same way as for the 
default values of the Fixed: Aspect ratio entry (i.e. when rubber 
banding has ended, set the default width/height to the pending rectangle 
width/height) that would allow people to more intuitively fix the 
current rectangle.

Use case:

A user create a rectangle and adjusts size to whatever is the size the 
user desires in a given context.

To avoid accidental resize of the rectangle, the user would like to fix 
the current size.

Since the default values for Fixed: Size are updated to that of the 
pending rectangle when rubber-banding has ended, the values in the Fixed 
Size entry are that of the current pending rectangle (if the entry is 
not in user-overridden mode, of course).

So, to fix the current size, the user only have to check the Fixed: Size 
checkbox.

If the user wishes to adjust the size a bit, he/she can simply do that 
in the Fixed: Size entry.

</use case>

With defaults always as 100x100, it is annoying to adjust the Fixed: 
Size since 100x100 is likely to be far from what the current rectangle 
is. To adjust e.g. only the height, the user would have to manually 
enter *both* the current values in the Width: and Height: entries, which 
can be tedious with 4 digit width/height sized rectangles.

If we did this like described, there would be no positive feedback cycle 
freeze, right? Or is it something I am disregarding? To me this sounds 
as a pretty good way of allowing the users to fix the size

- Martin