L_a->(x+1=2t_1,y+1=3t_1,z+1=4t_1)
L_b->(x+1=3t_2,y=4t_2,z=5t_2)
or
L_a->p=p_a+t_1 vec v_a
L_b->p=p_b+t_2 vec v_b
with
p_a=(-1,-1,-1) and vec v_a=(2,3,4)
p_b=(-1,0,0) and vec v_b=(3,4,5)
calling now hat v_c = (vec v_a xx vec v_b)/norm(vec v_a xx vec v_b) and also making
Delta p = p_a+t_1 vec v_a-(p_b+t_2 vec v_b) = p_a-p_b+t_1 vec v_a-t_2 vec v_b
making the scalar product by hat v_c we have
<< Delta p, hat v_c >> = << p_a-p_b, hat v_c >> because << hat v_c, vec v_a >> = << hat v_c, vec v_b >> = 0 and finally
d = abs(<< Delta p, hat v_c >>) = abs(<< p_a-p_b, hat v_c >>)
In our case study we have
p_a-p_b = (0,1,1)
vec v_a xx vec v_b = (-1,2,-1) and
hat v_c = (-1/sqrt[6], sqrt[2/3], -1/sqrt[6]) and
d = sqrt[2/3] - 1/sqrt[6] = 0.408248