is there a way to eliminate the crossing of edges?
I tried many ways, nothing helped.
digraph graphname {
graph [splines=ortho,];
node [shape=box,];
l;
l_a [shape=diamond,label="",height=0.20,width=0.20];
l_a_s [shape=point];
l_a_i [shape=point];
l_a_ii [shape=point];
l_a -> l_a_s;
{rank=same; a -> l_a -> l}
{rank=same; l_a_i -> l_a_s -> l_a_ii}
l_a_i -> i;
l_a_ii -> ii;
l_c [shape=diamond,label="",height=0.20,width=0.20];
l_c_s [shape=point];
l_c_t [shape=point];
l_c_n [shape=point];
l_c -> l_c_s;
{rank=same; l -> l_c -> c}
{rank=same; l_c_t -> l_c_s -> l_c_n}
l_c_t -> t;
l_c_n -> n;
}
some more details:
All you need to do is to reorganize your node/edge definitions a little bit, no additional edges needed.
for example:
digraph graphname {
graph [splines=ortho,];
node [shape=box,];
a;
l_a [shape=diamond,label="",height=0.20,width=0.20];
l;
l_c [shape=diamond,label="",height=0.20,width=0.20];
c;
{rank=same; a -> l_a -> l -> l_c -> c}
l_a_s [shape=point];
l_c_s [shape=point];
l_a -> l_a_s;
l_c -> l_c_s;
l_a_i [shape=point];
l_a_ii [shape=point];
l_c_t [shape=point];
l_c_n [shape=point];
{rank=same;
l_a_i -> l_a_s -> l_a_ii;
l_c_t -> l_c_s -> l_c_n;}
l_a_i -> i;
l_a_ii -> ii;
l_c_t -> t;
l_c_n -> n;
}