The problem with Alex Browns validation test is, that he is using a document from OpenOffice.org 2.4 and validates it against the ISO-approved standard. Jesper Lund Stocholm claims that he has created some documents with various versions of OpenOffice.org that doens't validates correctly. From my knowledge, Jesper hasn't put any documentation to support witch validation method he used nor put forward the bespoken documents. He has not even told us the seriousness (number of errors) of the 'variations'.