Since 'dipoles' share many commands with 'bonds' they are suited very well for displaying directed connections between atoms/pseudoatoms. One can especially switch very easily between both simply by changing the keyword 'bond' to 'dipole' or vice versa.
What is missing is an option to draw a circular dipole arrow pointing to the atom/pseudoatom itself.
I know that this doesn't fit the original idea behind 'dipole'. But since 'dipoles' are advertised in the scripting documentation for using them as arrows also in a non-dipole context I hope this option might still be considered for implementing it.
Rolf, I think you mean an arrow that is not a dipole but would be used for something else. I think this is what DRAW is for. All sorts of curved arrows there, right?
Bob, you are right that the arrows I want to draw are not actually dipoles. They are rather 'directed bonds', including bonds linking an atom to itself. You can see an example picture here: "http://www.phy.duke.edu/research/photon/qelectron/proj/Boolean_chaos_3nodes.png". Since I will need the same functions for 'directed bonds' as for 'non-directed bonds', 'dipoles' in Jmol look like the most suitable construct available. And they are also less troublesome with changing spacefill sizes (e.g.: 20%, 50%, etc.). It actually is already possible to define a dipole by providing the same atom twice, but nothing is drawn.
Besides I couldn't figure out a way to use the current 'draw' command options to get a circular arrow like I would need it. Regarding all the advantages of the 'dipole' command I thought it would be at least worth a try to ask for this option for the 'dipole' command instead of the 'draw' command.