CONTENTdm in our case needs a tab delimited file with the following fields (with the xmp tag in parenthesis): title (<XMP-dc:Title>), date, collection / source (<XMP-photoshop:Source>), description (<XMP-dc:Description>), rights  (<XMP-dc:Rights>), subject, creator	(<XMP-dc:Creator>), contributors, notes, filename (<System:FileName>). For those that I do not have XMP tags for I know that I won't be pulling that information out of the image metadata. 

In our case the image description is being entered by technitians directly "into" the image file using Adobe Bridge so I'm fairly confident that we'll be able to pull back out what we need. 

If you know people at NIST I'd be happy to expand my contact list here on campus! :) 

Thanks for all the help,
Andrea Medina-Smith
Metadata Librarian
NIST Gaithersburg
[log in to unmask]

Be Green! Think before you print this email. 

-----Original Message-----
From: Code for Libraries [mailto:[log in to unmask]] On Behalf Of Simon Spero
Sent: Monday, January 14, 2013 7:57 PM
To: [log in to unmask]
Subject: Re: [CODE4LIB] XMP Metadata to tab-delemited file

XMP uses a subset of RDF/XML, with a few  limitations thrown in to make reification and provenance tracking impossible, but hey who needs metadata.

I'm not sure if XSLT is particularly well suited to anything, but it ought to be possible to cruft something up.  I would still recommend following Owen's suggestion of using an RDF toolkit of some kind to take hide the details of any sequences etc.  I can point you at a few people at NIST who might be able to give some advice.

What does ContentMFDM expect in it's tab separated files?


On Mon, Jan 14, 2013 at 2:56 PM, Owen Stephens <[log in to unmask]> wrote:

> I'm not familiar with what XMP RDF/XML looks like but it might be 
> worth using an RDF parser rather than using XSLT?
> Graphite ( is pretty easy to use if 
> you are comfortable with PHP
> Owen
> On 14 Jan 2013, at 19:09, Kyle Banerjee <[log in to unmask]> wrote:
> > On Sat, Jan 12, 2013 at 1:36 PM, Michael Hopwood 
> ><[log in to unmask]
> >wrote:
> >
> >> I got as far as producing XMP RDF/XML files but the problem then
> remains;
> >> how to usefully manage these via XSLT transforms?
> >>
> >> The problem is that XMP uses an RDF syntax that comes in many 
> >> flavours
> and
> >> doesn't result in a predictable set of xpaths to apply the XSLT to.
> >
> > XSLT is not a good tool for many kinds of XML processing. In your 
> > situation, string processing or scanning for what tags are present 
> > and
> then
> > outputting in delimited text so you know what is where is probably a
> better
> > way to go.
> >
> > kyle