Sat, 30 Aug 2003 11:43:23 +0200
On Sat, Aug 30, 2003 at 01:24:35AM +0200, Roel Spruit wrote:
> any suggestions about how this would work? key-modifiers are kinda running
> out, so
> we'll have to go with a mode or something....
There is so much functionality that could be added to blender (and so much
alrready exists). But only a limited set of keys. And the average user can
only memorize some keystrokes.
Problems I see:
- Modes using the same keys for different things.
- The number of different modes.
- Lack of visibilty: you have to know about much functionality waiting
behind some keystrokes with no visual hint.
- Menu-bar and Space-Menu require mouse-action. Or have I missed some
I would like to suggest the creation of a table, showing all modes and
shortcuts with the asociated functionality. Is there a possibility of
generating a rough version of it out of the sourcecode? Or could such
a table be created wiki-style?
The table should help to clean up keyboard-functionality. Removing
redundancy and making things more consistent by making shure compareable
actions in different modes use the same keys.
One thing that could be eliminated right away is Shift-Leftclick for
buttons. A simple click would be sufficient. For the increase/decrease
functionality buttons would have to have three clickzones. But because
you don't know if you increase or decrease clicking in the middle right
now, this should't be a problem. One step further: change value-buttons
to be like this:
| | + |
| 100 ----
| | - |
Think of plus and minus as arrows up and down.
If you than increase or decrease and click once to often, the correction
doesn't require a mouse move to the other side of the button, but instead
only a short move up or down. For the same reason good scrollbars have
two arrow-buttons on at least one side (like with KDE).
By the way: End and Pos1 keys should work while editing button-values!
The minimalistic aproach: numerical input for value-buttons works on
It would be of great value for learning (and relearning after some time of
abstinence) if you could see which buttons/functionality is accesible at
some point. For exanple G, R and S if an object is selected. Hints like
"[G]rab for moving" or something like that could be shown right at the
point of action and foucus of attention: On the actual 3d-window.
Hard to miss and great for the newbie, but maybe to distracting.
Therefore, another possibility would be statusba / new window-type for
displaying short help messages. In Photoshop ypu can read about modifier-
keys for the actualy selected tool in the statusbar.
The menu could be enhanced with mnemonics like in standard menus with
underlined letters. So you would press"A" for "Add". Because of the
two times "Object" the menu would have to be restructered.
Another approach would be auto-completion (inspired by Houdini):
Press Space for the menu. At the top of the menu you can see your
keyboard input. A soon as you start entering letters, the menu changes
to show possible completions. I there are several possible completions
you can type on or select one of them with mouse or cursor-keys. The
great thing about this is, that the completions can be taken out of
all levels of the menu. So for "A" you would see "Add" and "Armature"
"Append" and so on.