Nicholas Frechette

See here: http://marc-b-reynolds.github.io/quaternions/2016/05/30/QuatHDAngleCayley.html Seems like a great fit and easier to support than dropping the largest component.

Host runtimes sometime have the need to do a lot of work in parallel and animation can be memory intensive depending on the data. It is thus useful to know...

Machine training often uses raw or near raw animation data for training. While fast to sample, it consumes a huge amount of memory which can make it difficult for the...

In certain scenarios, an animation is made on something but we wish to play it back at a different scale (e.g. giant monster). When this case occurs, it is necessary...

Once RTM and ACL support Slerp, we should use it. glTF supports sparse samples which doesn't work too well with Lerp for rotations.

Make sure we properly support unicode paths. Add a regression test.

Long paths require a `\\?\` prefix to work. Make sure we have a regression test for it.

http://seriot.ch/parsing_json.php It provides a large corpus of test json files along with whether or not they should parse. sjson-cpp should be able to parse json at some point but for...

Perhaps allow a strict SJSON check or allow JSON key/value pair parsing, most parsers do so such as https://bitbucket.org/Anteru/sjson/overview Basically in SJSON keys are of the form: `key = value`...

It is often desirable (e.g. config files) to load up an SJSON file in memory, modify it, and write it back. Conceptually, it is a hash of hashes that contains...