MathJinni works in radiant mode (As do the trigometric C functions in the background). I'll put support for other modes on the TODO list. If you have any specific whishes regading the way it should be implemented, mail me.

MathJinni does use C double internally. So rounding errors that expose decimal to binary conversion problems for floating points are possible and will be seen in edge cases. I assume that rounding numbe routput to a sensible limit may be wise (10 decimals behind colon for example). I would be interested in getting opinions about that!

I you have other improvement ideas for or have found bugs in MathJinni also do not hesitate to mail me.

Gruß...Tim