This is good overall, but the name seems a little off to me. When we use ignoreDepthValues elsewhere it's something that the user had the opportunity to set. Here it only functions as a way to communicate the system's capabilities, so a name like willIgnoreDepthValues or alwaysIgnoresDepthValues feels more appropriate.

That said, during the TPAC call we also discussed that it may be more useful to communicate that a system prefers to have depth values provided, so perhaps we should consider flipping the logic and using a name like prefersDepthValues or, more generally, usesDepthValues.

That said, during the TPAC call we also discussed that it may be more useful to communicate that a system prefers to have depth values provided, so perhaps we should consider flipping the logic and using a name like prefersDepthValues or, more generally, usesDepthValues.

Thanks! I updated the spec so it reads usesDepthValues

Should XRProjectionLayer use the same attribute name?

I could go either way on that. It's nice to have them be aligned but the context is a bit different and the attrib in the XRProjectionLayer has already shipped in at least one browser, right? I don't think that makes it unchangeable, but I also don't feel the need to break things in this case.

I could go either way on that. It's nice to have them be aligned but the context is a bit different and the attrib in the XRProjectionLayer has already shipped in at least one browser, right? I don't think that makes it unchangeable, but I also don't feel the need to break things in this case.

Yes, we already shipped it and three is checking it.