[Bf-committers] Weights range in vgroups

[Campbell Barton]

On Sat, Jul 30, 2011 at 6:09 AM, Daniel Salazar - 3Developer.com
<zanqdo at gmail.com> wrote:
> Maybe, all i know is I tried setting via python and it clamped them..
> i really wanted to store unclamped values that time
>
> Daniel Salazar
> 3Developer.com
>
>
>
> 2011/7/29 Αντώνης Ρυακιωτάκης <kalast at gmail.com>:
>> On 29 July 2011 21:29, Daniel Salazar - 3Developer.com <zanqdo at gmail.com>wrote:
>>
>>> VGroups are clamped internally to 0-1 range, I agree it would be nice
>>> to overwrite it but even assigning the value directly via python i got
>>> a clamping last time i tried. If you manage to make it work, I'd say
>>> it is useful.. dunno about safe though considering the uses of vgroups
>>> all over blender
>>>
>>> Daniel Salazar
>>> 3Developer.com
>>>
>>>
>>
>> Maybe it's got to do with RNA property hard/soft limits?

To go against popular opinion:
-1 on allowing non-clamped vertex weights.

Advantages for 0-1 weights are.
- Simple for tools, weights can be inverts without worrying about
becoming negative or corner cases (like inf weight when the values
become very large).
- Simple for weight painting (how to visualize weight of 10000?)
- A bit easier to troubleshoot weight problems if you know that the
range is clamped.

One case where unclamped weights may be useful is the proximity vgroup
modifier, but I would rather this just map min/max distance to 0-1,
don't think this is a bad loss of functionality.

-- 
- Campbell