PDF/A hell

It didn’t suffice that Adobe Acrobat 7 used and expected a PDF/A namespace from an old specification draft (http://www.aiim.org/pdfa/ns/id.html). Now, it turns out that the official ISO document (ISO 19005-1:2005) had to be corrected, too. The document uses “http://www.aiim.org/pdfa/ns/id” as the PDF/A namespace. After a problem report today from a client trying out Acrobat 8 on a supposedly PDF/A-compliant file generated by Apache FOP, I found out that the PDF/A committee has published a corrigendum in which they are correcting the PDF/A namespace to “http://www.aiim.org/pdfa/ns/id/” (note the trailing slash). Furthermore, I think there’s an additional problem with metadata handling since the “Subject” is not mapped to dc:subject anymore, but to dc:description[“x-default”]. Looks like I just got some more work to do…