Potential bug while calculating baselines

Hello,

At line #108 of the file ps_load_initial_gamma.m, it says:
bperp_mat(:,i)=bc.*cos(look)-bn.*sin(look); % Convert baselines from (T)CN to perp-para coordinates

However, on the article

Bähr, H., Hanssen, R.F. Reliable estimation of orbit errors in spaceborne SAR interferometry. J Geod 86, 1147–1164 (2012).
(Reliable estimation of orbit errors in spaceborne SAR interferometry - Journal of Geodesy)

on page 3 it says:
B||(t, theta) = Bh(t)cos(theta)+Bv(t)sin(theta)
where Bh stands for horizontal and Bv for vertical

Notice the +/- signs.
Is that a bug?

Oded Horowits