Naming conventions: Vec3::as_dvec3 vs Quat::as_f64

[LPGhatguy]

Hello! Today I'm starting a port of our game from using f32 coordinates to using f64 coordinates for everything in world space. We're switching a lot of Vec3 values to DVec3, and Quat to DQuat.

As part of that work, I noticed that conversions in the vector types and the quaternion types have different naming conventions! I thought that was incongruous, so I wanted to open this thread to ask about that.

I like the naming convention in the Quat types. I think as_<numtype> is more clear than as_<vectortype> in code that's converting both vectors and quaternions.

One consideration is that there are many more numeric types supported for vectors, as well as the 16-byte aligned SIMD variants with the A suffix. Normalizing the naming scheme to match Quat would make those other conversions stand out.

What are you thoughts on this? Do you think it would be valuable to change either of the naming conventions, i.e. adding Quat::as_dquat or adding Vec3::as_f64?

[bitshifter]

I think the history of these methods may be lost as they started on a dev branch that has been merged and deleted. From memory though, they started out as as_f64, etc and were later renamed to as_dvec3 and so on. It looks like I missed the Quat methods as part of that rename.

It was a while ago that I was thinking about this, but I think I ultimately felt that as_dvec3 was a lot more explicit about what was happening than say as_f64, the later is a scalar value so probably people would assume it would create a vector of f64 but maybe it's not completely obvious. Then there's also the extra align(16) types to deal with and it seemed more convenient to add something that converted directly to those, although as_vec3a might be a bit misleading there as there's more going on than simply casting the properties.

I think it would make sense to align the Quat methods with everything else and deprecate the as_f32 and as_f64 on quat types.