It's not really a bug or a big deal. But I found the documentation for the `slerp` a bit unclear. It states:

an interpolation for a constant motion between other and *this t in [0;1]

This makes it seem like at `t=0` the return value will be `other` and at `t=1` the return value will be `*this`.

Perhaps it would be better to reverse it to read:

an interpolation for a constant motion between *this and other, for t in [0;1]

Thanks for the report. Actually the situation was even worse because slerp was documented twice. The other explanation was more consistent. Fixed there: https://bitbucket.org/eigen/eigen/commits/43047fa7fb05/.