Open lexaknyazev opened 3 months ago
Would there ever be an extension object that wants to keep the node
property defined, but replace path
with some new enum?
(I can't think of one, I'm just pondering what the proposed change might disallow...)
That extension would extend the list of valid enums (schema-wise, any string is valid anyway).
The proposed rewording looks fine for me.
I might not really be able to nitpick on this level of the language. But the suggested wording is
When the
node
property is defined, valid values for thepath
property are"translation"
,"rotation"
,"scale"
, and"weights"
.
and I think that the concern that Ed brought up might be avoided with
When the
node
property is defined, then"translation"
,"rotation"
,"scale"
, and"weights"
are valid values for thepath
property
or
When the
node
property is defined, valid values for thepath
property include"translation"
,"rotation"
,"scale"
, and"weights"
.
I think that this could make clear that these values are always valid when the node
is there, but other values might also be valid, regardless of whether the node
is there or not, depending on details of possible extension. (To be confirmed by someone who might have a better understanding of the details...)
The current spec says:
Additionally, the JSON schema for the
path
property allows any string value for extensibility.The
KHR_animation_pointer
spec says:This extension spec does not (and cannot) cover the inverse case: if the
path
is set to"pointer"
but the extension object is not present.Proposed base spec update
@emackey @javagl WDYT?