The question is, has anyone then checked how much of that spending went to on open software to pay for or buy software or services? It could be actually running a system that used open software versus closed software.