Don't confuse the specification of the XMP serialization with the use of XMP as
separately defined by the ISO standards such as PDF/A.
The correct answer here is specified in the PDF/A (ISO 19005) standards
(specifically ISO 19005-3:2012 for ZUGFeRD e-invoices).
ISO 19005-3:2012, Table 8 "PDF/A identification schema: defines
pdfaid:conformance to be a "Closed Choice of Text" with values "A", "B" or "U"
- so always uppercase for PDF/A XMP conformance levels.
And, yes, the XMP Metadata value for PDF/A conformance levels is always
uppercase yet the monikers used to identify PDF/A files (such as "PDF/A-3b")
always use lowercase (as also stated in ISO 19005 standards, in Clause 4
Notation). This is a common point of confusion.
> -----Original Message-----
> From: Tim Allison <talli...@apache.org>
> Sent: Friday, 4 April 2025 9:30 PM
> To: users@pdfbox.apache.org
> Subject: Re: PDFA extension value type case sensitivity
>
> Not that Tilman needs it, but to support his point, we use JempBox in
> Apache Tika because it is much more relaxed.
>
> On Fri, Apr 4, 2025 at 5:55 AM Tilman Hausherr <thaush...@t-online.de>
> wrote:
>
> > Hi,
> >
> > It gets weirder, I found the XMP specification, sometimes they write
> > "closed Choice", sometimes "Closed Choice".
> >
> > I could change it, but it's possible you'd get other problems with
> > xmpbox anyway. You can try Jempbox which is also from us, but has a
> > smaller API and is more relaxed (xmpbox was created mostly to verify
> > PDF/A files).
> >
> > Tilman
> >
> >
> >
> > On 04.04.2025 11:06, Peter Nowak wrote:
> > > Thanks for your quick reply.
> > >
> > > I feared that it has to be case sensitive 😅
> > > The PDF generation unfortunately is not in our hands, we just get the
> > > files delivered to us by customers.
> > >
> > > Then I need to do some custom xml parsing to get the
> > > /<zf:DocumentFileName>zugferd-invoice.xml</zf:DocumentFileName>
> > > /out and can't use xmpbox for it.
> > >
> > > Thanks for the clarification
> > >
> > > Best regards,
> > > Peter
> > >
> > > Am Fr., 4. Apr. 2025 um 09:33 Uhr schrieb Tilman Hausherr
> > > <thaush...@t-online.de>:
> > >
> > > On 04.04.2025 09:13, Peter Nowak wrote:
> > >> I did not find a concrete answer to it in the xmp specs.
> > >> Are the value types case sensitive? We received a ZugFERD pdf and i
> > >> tried to read the xmp metadata with xmpbox but it fails because of
> > an
> > >> unknown value type. The problem is the casing ("closed choice of
> > >> Text"), not the actual value. (xmpbox expects an upper case C in
> > >> choice)
> > >
> > > ChatGPT says it has to be case sensitive (I hope I asked the
> > > correct question):
> > >
> > > Tilman
> > >
> >
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org
For additional commands, e-mail: users-h...@pdfbox.apache.org
