ParallelCamera implements some specific behavior which has not been documented anywhere. This might lead to confusion when trying to use it outside of what it was designed to do.

Things to add/cover:
- A small explanation on how ParallelCamera achieves its goal (using ortho transform, etc)
- ParallelCamera not using nearClip/farClip properties, but estimating clip dimensions from view width and height
  - Reasons why it does that and potential drawbacks of this approach