I overlaid multiple VECTOR/ASPECT plots of a single velocity field and accidentally plotted the same vectors twice. To my surprise they had different angles!
If you look at the arrows at z = 95.74 m in the attached plot, the blue and red arrows are supposed to be identical but they are very different. (The nearest data points are at 84.2 m and 108.00 m, which are far enough from 95.74 m we are looking at.) Both are plotted with ASPECT. The sample script and data are also attached.
I also wonder about a totally peripheral issue. I thought PPL WINDOW,OFF refers to the plotting area. But the attached script and plot suggest that the window-edge truncation is carried out according to the box of x=X1:X2/z=Z1:Z2 rather than the axes. The blue arrows are truncated at z=300m. I would think truncating graphics elements at the edges of x= and z= isn't very intuitive.
FERRET v7.3 (optimized)
Linux 2.6.32-696.13.2.el6.x86_64 64-bit - 12/04/17
on a Linux machine.
!==== Sample script =============
set mode grat:dash
let fac = 1e3 ! velocity-length scale
define symbol y1_ 31S
define symbol y2_ 27S
let z2_ = 900
let samp_vec_in_ = 0.35
let samp_vec_ = fac/200
!go finish-print try-vector-overlay 3 false
!=== END of script ====