[Bf-committers] Delta Scale (patch to multiply rather then add).

[Campbell Barton]

On Mon, Nov 28, 2011 at 8:28 AM, Michael Fox <mfoxdogg at gmail.com> wrote:
> ummm yeah, delta loc adds, delta rotation adds, so logically and
> intuitively, delta scale should add,

The idea of delta (from what I can tell) is to apply a difference in
transform channels,
rotation does not add - if you look at object_rot_to_mat3() it
converts the rotations into a matrix and multiplies them together.

If you want to double the existing 1.0 scaled object `intuitively`
you'd set the delta to 2.0 right? :),
like if the object was made the child of another object which was then
scaled by 2.0x.
But with adding delta scale this makes the object 3.0x, worse, you
cant consistently double the existing scale, you always need to do
some mental arithmetic to work out what double would be.

Applying  scale to a point multiplies so IMHO its logical that delta
scale would multiply too.

Another downside of adding scale is you can get negative scales
without meaning to (flipping the object), where as with multiplying as
long as both are > 0.0 it wont happen.

> multiplying is just going cause issues, when you scale something in the
> UI, then it changes when you playback,
>
>  adding scale is more intuitive when animating, so you don't constantly
> fret over what you are actually keying and that key is going to behave
> properly when I playback

don't understand why there would be any trouble here.

> ass for the bug, delta scale when its being calculated just needs to
> subtract by 1 that was scale of 1+0=1, changing it to multiply is just
> too severe and heavy handed

are you suggesting delta scale be modified to subtract 1 before being applied?

> On 28/11/11 05:20, Campbell Barton wrote:
>> This bug report highlights that delta scaling is adding rather then
>> scaling the objects scale.
>>
>> https://projects.blender.org/tracker/index.php?func=detail&aid=29111&group_id=9&atid=498
>>
>> I've attached a patch which changes delta-scale to multiply to scale
>> which IMHO is much more logical,
>> incidentally do_displacement() in convertblender.c is already treating
>> dsize as a multiplier which is odd.
>>
>> The patch attached to the bug report updates existing files with 2 caveats.
>> - animated dsize wont be converted.
>> - objects scaled to 0.0, with a non-zero delta scale can't be
>> converted, think this is really an obscure case though.
>>
>> Would this be acceptable to apply before 2.61?
>>
>
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers
>



-- 
- Campbell