I have noticed an error in the documentation in Isight in the section describing the Kriging approximations.

The equations for the correlation between a new point we want to predict and all the points in the design of experiment (for all the options: Exponential, Gaussian, etc) should have a negative number in the exponent (Isight 5.9 Component Guide, page 789).

(For reference, see for instance Eq. 18 in this article, or Eq. 7 in this other article).

I am concerned about this, in particular I'd like to enquire if it is possible to know whether this error is in the documentation only or if it has carried into the code as well. Is there any way (for an academic license without support) to pass on this request. I need to use the Kriging surrogate model but i need to be sure it is reliable (and initial comparisons with other models do not indicate that it is).