[Wien] treshold in rstruc.f

[Stefaan Cottenier]

> FYI, probably you have to slightly adjust the atomic position of Si8
> (not 7).
>
> You might know a better correction for the position, but it seems to
> pass 'x symmetry' after slightly changing z=0.06274433 to z=0.06274333
> (last 4 changed to 3).

If that works, it's probably just by chance. Si8 is a 12i position in 
space group 166, with x, y and z all undetermined. Small (yet 
consistent) changes in z to all 12 Si8 atoms should not make any 
qualitative difference.

I suspect the reason for the failed test is just accumulation of 
rounding errors (whence why adjusting the treshold should be fine). By 
slightly changing some of the coordinates, a particular rounding error 
could be avoided. But there is no (symmetry) reason why these changed 
positions should be any better than the original ones.

Stefaan