[Bf-funboard] (Groo) Re: Object color proposal

Luke Wenke bf-funboard@blender.org
Wed, 1 Oct 2003 12:25:13 +1000


> To be tactless and blunt, knock this crap off.  We've already agreed (I
> thought) that states like "non-selected active" shouldn't exist.
But they do. They appear as black wireframes with magenta centres (rather
than yellow centres for non-selected non-active objects). For lights, all
unselected objects are yellow. (Rather than the active unselected ones being
magenta in the centre).

> There's non-selected, selected, active (a subset of selected) and
> (because I don't know the real name) currently manipulated.
Then things like ctrl-p, ctrl-t, ctrl-c, would work differently since they
can involve non-selected active objects... I mean they use actives objects
whether or not it is selected.
If non-selected active objects don't exist, how would "a" then work? I mean
say you had some objects selected then pressed "a" to unselect them all,
then pressed "a" to select them all again... would the same object still be
active? If so, then Blender remembered which object was active even when it
was unselected. I guess Blender could just not tell the user it knows this
though (by not making it a different colour) but for ctrl-t, ctrl-p, ctrl-c,
(and tab), non-selected active objects are relevant...
At the moment, if you press tab with nothing selected, it will edit the
active non-selected object... I guess if your changes were made, there would
be no non-selected active object so pressing tab would do nothing at all
(except maybe display the error "nothing selected to edit").
So anyway, unselected active objects are treated different in blender to
selected non-active objects and other unselected objects so unless this is
going to stop happening through lots of changes to the Blender code, the
user should be informed about which object is active if it is unselected.

> The reason for NOT making the centerpoint color similar to the wireframe
> color is to distinguish it from the wireframe.
My suggestion was to make the centrepoint colour basically the same as the
wireframe colour (except when the object is unselected)... you said
centrepoints should be purple for non-selected and yellow for selected...
these would still be the same colour as other wireframes! Except they
wouldn't be the same colour as their own wireframe, which would be logical.
How about the centrepoint is just lighter or darker than the wireframe (e.g.
lighter yellow or darker yellow, etc)

> While it might, I think the axis style of indicating the
> centerpoint would be helpful in determining orientation.  I like the
> idea of implementing the tri-color for each axis throughout the rest of
> the interface (good suggestion, Thorsten).
You can already make an axis thing appear for objects at the centre (it can
be toggled, except for empties) and I think it even labels the axes x, y,
and z. I think that is a lot more useful than that tri-colour thing.

> I think you're misundertanding my suggestion, I think the centerpoint
> should look something like this image (apologies for the
> quick-n-dirtiness):
http://www.mrhostbot.com/f.php/517,117/blender_centerpoint.png
I thought the centrepoint was a single colour, made up of changing red,
green and blue components... well now that sounds better... I think the red,
green, and blue could be darked and lightened depending on if the axis is
going out of the screen or into the screen. (Otherwise the axes are
ambiguous)

> Just because it cannont currently be selected and moved is not a reason
> that it shouldn't be.
The reason it shouldn't be would be because people could accidently select
and move it when they want to select and move vertices... if you could use
"centre cursor" or "new cursor" in editmode then you could move the
cursor... just left click (to move the cursor somewhere) then click "centre
cursor". You could even use shift-s to move the cursor to where a vertex is,
etc.

> I think it should.  There's a big benefit to being able to actively
> translate and rotate the centerpoint of an individual object.
> Granted, it *can* currently be done using the
> 3D-cursor, but I think being able to select and move it interactively
> would be more powerful and more naturally fit the current editing
> workflow.
I guess it would be good to do things like rotate the object centre (but not
the vertices...) but the problem is about selection... I'm assuming that the
object centre would only be able to be moved independently within editmode
(and not at other times). So what would happen if you used "b" to select
some vertices and also the centrepoint? Would the centre point be selected
as well? Because in most cases, the user wouldn't want the centre point to
be selected as well.... I think letting people select the centre could get
quite annoiying for people who are used to just selecting vertices and not
having to worry about selecting the wrong thing.
- Luke