The notational part of this debate is definitely not agreed upon. Andrew Gelman and Christian Robert both argue for p(data|H_0) in comments on those blog posts.