Sue One of the limitations we work with are the available features of our scanning software. Those of us who use K1000, if we choose to strip headers, will also remove the page numbers. Given that as background, if given the choice between full headers with page numbers or stripped headers, I'd opt for the full headers. Also, at play here, of course, are the limitations of our validating tools. Hence, as I validate in K1000, I face the same choice. And then, of course, are the vagueries associated with BookShare's stripper which seems to have a life of its own.