Two recent papers by Kawarabayashi, Thomas and Wollan, "A New Proof of the Flat Wall Theorem" (arXiv:1207.6927) and "Quickly Excluding a Non-Planar Graph" (arXiv:2010.12397) provide major improvements over Robertson and Seymour's original proof of the structure theorem for finite graphs that exclude a given graph. The first paper redefines the notion of a flat wall. Unfortunately, this new notion is too strong. As a result, the new Flat Wall Theorem in that paper is incorrect. A counterexample is given in Appendix A. A follow-on lemma in the first paper, about the transitivity of flatness, is also incorrect, a fact that was noticed by Dimitrios Thilikos et al in arXiv:2102.06463. However, that error is derivative and not the main issue. This paper provides a weaker definition of the notion of a flat wall, provides a correction to the proof of the new Flat Wall Theorem and a new proof of flatness transitivity. The notion of a tight rendition as presented here differs from Thilikos' definition but is defined much more simply, and the notion of a proper cycle is introduced. The notions of certificates and tilted walls used by Thilikos turn out of be unnecessary and transitivity is preserved in its original simplicity and generality. Most importantly, it looks like the new weaker definition of flatness is all that is really necessary to carry through the proof of the structure theorem in the second paper of Kawarabayashi, Thomas and Wollan.