Hm. To me, that was clearly not an option.

My doubt centers around this:

0 _      _ 3
   \    /   
1 _ \  / _ 4
   \ \/ /   
2 __\/\/__ 2
    /\/\    
3 _/ /\ \_ 0
    /  \    
4 _/    \_ 1
[download]

The crossovers between: the 0 & 3 wires; and the 1 & 4 wires are "vertically adjacent". Ie. As close together vertically as two crossovers can be. It also means that the striaght through 2 wire has a large break which cannot be filled, leaving its continuity unclear.

Both of these things can be addressed like this:

0 ___    ___ 3
     \  /
1 _   \/   _ 4
   \  /\  /
2 __\/__\/__ 2
    /\  /\
3 _/  \/  \_ 0
      /\    
4 ___/  \___ 1
[download]

This makes for a much clearer diagram, but that means that all the crossovers are 'on the wire lines' rather than between them. It also may be seen to contravene the "It is considered elegant not to make the grid wider than it has to be." rule.

So maybe it would need to be something like:

0 _____  _____ 3
       \/  
1 _    /\    _ 4
   \  /  \  /
2 __\/____\/__ 2
    /\    /\
3 _/  \  /  \_ 0
       \/
4 _____/\_____ 1
[download]

But that still leaves some crossovers occurring on wire lines rather than between them. So maybe:

0 _____  _____ 3
       \/  
1 ___  /\  ___ 4
     \/  \/  
2 __ /\__/\___ 2
    /  \/  \
3 _/   /\   \_ 0
      /  \
4 ___/    \___ 1
[download]

The 2 line sucks and it destroys the natural symmetry, but it more clearly complies with the rules -- except perhaps the no wider than necessary. Unless it is necessary to be that wide in order to comply with the other rules :)

It is an interesting programming problem, but it would suck to implement a brilliant solution, only to be failed because the problem description was ambiguous.