erratum p 256

[cottrell]

I think the equation has some typos. Consider $\mathcal{H}$ as identity and you do not recover the basic GPR posteriors. Probably you mean $\xi$ as train points and $\xi'$ as test and the inversion term should be like $C(\xi, \xi)$ etc.