Orthogonal Deviations in Dynamic Panel Data

[paul84]

Dear Members,
Dear Gareth,
it is not clear to me what is done in EViews if one specifies “orthogonal deviations” instead of “first differences” (Arellano Bond) as the transformation method in estimating a dynamic panel data model. Arellano and Bover (1995) describe “forward orthogonal deviations”, which is the paper cited in EViews User Guide II (e.g. p.575). Are those “orthogonal deviations” really "forward" in EViews 6?
Thank you,
Paul