Andrew is correct, but the best bet is to start over with a more conventional developer or a different paper. Start learning lith be sticking to things that others have shown to work then branch out. It is hard to know whether it is your technique or your experiments otherwise.