Blender was designed around principles of the latter, and those still stand 20 years later.
simply take translation icon :Antlab wrote:Hi Lukep.
Sincerely I don't see the strict relationship between graphical interface and tool/action oriented paradigms. I understand what you say about the working model of Blender, but this is not directly related to how the different commands and options are showed on the screen. IMO a simple and clean use of tabs, boxes, icons, expandable menus could maintain the same working model, but making the workflow easier for a lot of people (and less intimidating for new users, and more "standard" for professionals like Sabrina's colleagues).
As said before blender way is select first, then do action (action-oriented).
that is what you get with G key, click to finish
that is what you get too with gesture, click to finish.
the manipulator widget work the same except it is a long-click drag.
that is a first discrepancy.
the new ndof device work similar
want rotation ? same action with R and different gesture
Now how the translation icon in a separate view would work ?
select the tool, select objects, then what ? long-click drag ?
you have both discrepancy with the shorcut and the manipulator.
And this is very bad.
Worse, the icon being in a different view by definition, you lose context information, and you can even lose focus.
In a tool oriented software this work because each tool invokes its modal state, which is exactly what we dont want in blender.
We want to keep blender strengths, and one is the efficient workflow, so the key driven interface will stay the main one. Any widget and UI cues added must then be compatible.
Icons in the header part of 3dview are a possibility, but there is already the manipulator ones and it starts to be crowded there. Maybe if we can merge the 2.
Now situation is a bit confuse because with time, some tool-oriented modes crept in blender.
In conclusion, rewamping the interface yes, but not at the price of current ways.