Developing visualization applications is non-trivial and poses special challenges. This is due to the fact that typical visualization software processes a large amount of data resulting in large and sometimes very complex data structures. Traditional tools for debugging are of limited use because they de-couple the information they report from the spatio-temporal domain in which unexpected problems occur. We present a set of guidelines targeted specifically at debugging visualization software. The guidelines are inspired by experience in developing applications in both industry and research. Specific examples where the guidelines are applied are given throughout. In general, the key is to exploit the strengths of computer graphics and visualization itself in combination with some more well-known good practices.