[Bf-taskforce25] Blender spring cleaning

[joe]

On Wed, May 13, 2009 at 3:44 PM, Nathan Vegdahl <cessen at cessen.com> wrote:
> Here's my list.  Some of them are a bit subjective, but I think most
> of them are pretty sound.
>
>
> Lights:
>      - Can we kill classic shadow buffers?  Classic-halfway is
> universally superior as far as I know, and it doesn't seem to have any
> noteworthy performance/memory penalties.

It's not always better; there are issues on 90 degree corners.

>      - Does anyone actually find Lin/Quad weighted falloff useful?  I
> never use it, and just find it confusing, honestly.
>      - Sample buffers: in retrospect, can't this essentially be
> accomplished by just upping the buffer resolution and increasing the
> softness?  I feel silly that I didn't think of this during BBB when
> this feature was added.  And I feel bad suggesting its removal so soon
> after it was added, but it seems like a redundant feature, and is also
> really confusing.

True oversampling is better, I think.  Certainly easier then messing
with resolution/softness settings.  We do need to properly have the
shadow buffer system support all the filter types though (actually
will need two settings, one for PCF soft lookup and one for AA).

> Materials:
>      - The "Blinn" shader isn't actually Blinn.  This is my fault.  I
> would like to take 2.5 as an opportunity to fix this (I'll happily
> make a patch).
>      - IMO the "Cubic" button should be replaced with an actual
> cubic-lambert diffuse shader.  Far less of a hack that way, and much
> less confusing to users.  Again, I'll happily do the coding work for
> this.
>      - Raytraced reflections and refractions should be *additive* to
> the base material (M + R).  Right now they are mixed (Ma + R(1-a)).
> Again, I'd be happy to do the coding...

+1

>      - Is "Tangent V" just for anisotropic shading, or does it have
> some other purpose?  If that's all it's for, then it seems like a
> weird hack worth killing in favor of a real solution some time in the
> future.
>      - I would really like to replace the Toon shaders with diffuse
> and specular color ramp shaders (which are a super-set of toon's
> abilities).  Again, happy to do this myself.
>      - If I add proper ramp shaders, could we kill the material
> ramps?  I personally find them very confusing to work with, and it
> seems almost everything they do can be accomplished in a much cleaner
> way with ramp shaders + shader nodes.  So it would be largely
> redundant, and killing it would eliminate a lot of UI cruft.

Lots of people use material ramps though.  The truth is, the entire
material system is old and crufty.  We've tossed around ideas of
having a more modular material system, so we can keep the old system
for compatibility while designing a new one from scratch.  This should
be possible in 2.5.

>
>
> Render buttons:
>      - Backbuf image: all it does is specify a background image for
> rendering.  This is redundant, because you can easily accomplish this
> through the world buttons by toggling "paper" on and adding an image
> texture.
+1

>
>
> Object buttons:
>      - Is it just me, or are "slow parent" and "offset parent"
> redundant?  I say kill slow parent.
>      - Kill track and power track.
+1

>
>
<snip>
> Armature:
>      - Auto IK: IMO we should get rid of this toggle and just make it
> the default behavior for grabbing bones that don't translate (instead
> of rotating them as it does now).

I kindof like the current behavior, because it works nicely with mouse
gestures and right click-drag.  I wouldn't mind having auto-IK on by
default though, so long as it doesn't cause any annoying subtle
problems, and has actually been proved to be useful in a production
setting.

>      - B-bone Rest: kill this!  It's an evil behavior kept only for
> backwards compatability with older files!  DIE!!!!

As someone with such a rig, I would be sad to see it go. On the other
hand, I'm sure my rig will need a fair amount of updating for 2.5
anyway, so probably the least of my worres. :)

>      - Stride root: kill this.  It's obsolete with the new NLA system.
>      - Kill the "W" button next to the bone shape object field.  If
> someone wants the shape to always be wireframe, they can just model
> the mesh with only edges.

That's a good idea, though we'd have to make the armature code respect
the draw object's draw type first.

>
>
> Misc:
>      - Modifiers should always be "real".  None of this "make real"
> nonsense (ala Armature modifier).  It's just weird and confusing.
+1

>      - Killing radiosity: +1.
+1, especially the interactive thing.  Though I understand some people
use it for baking stuff for game/interactive engines.  :-/

Joe