 |
 |
Guenter: Very glad that you are working with XMP & hope to hear more about it down the road. We do need a tool that can extract XMP in batch. All I know about are some older tools that extract XMP metadata from individual files, such as the SWAD-Europe: XMP Metadata Extraction Demo at http://www.w3.org/2001/sw/Europe/200206/imagemeta/extract/extract and a JPEG XMP extractor at http://chimpen.com/xmp/extract . HP's Jena might be of some use with this. It has an RDF/XML parser that is compliant with the most up to date W3C working drafts and is distributed as open source. That about exhausts my technical knowledge of the package, but there is more info at http://www.hpl.hp.com/semweb/jena2.htm . Bill >>> [log in to unmask] 2/19/2004 12:38:46 PM >>> Hi everybody, Nancy Hoebelheinrich prompted me to respond to this message (I just re-joined this list after an 18 month hiatus). If some of the things I say have been said already on the list during the time it took me to subscribe, my apologies for any repetition. Adobe XMP contains a number of built-in namespaces - the most pertinent for this discussion are probably Dublin Core, what looks like a proprietary Rights Management Schema (very basic), and an EXIF Schema. You can find more details on which schemas are built in and what the covered elements are in the XMP Specification at http://partners.adobe.com/asn/tech/xmp/download.jsp. XMP metadata could definitely span a whole range of METS sections, from descriptive to rights to technical to digital provenance. However, XMP can potentially do much more than the built in stuff - if you don't like what's built in, you can define your own namespaces and populate them with metadata. We have been exploring this as one possible route for capturing technical metadata according to NISO Z39.87 in RLG's Automatic Exposure initiative. The community could define what in XMP lingo is called a "custom panel" with all the NISO Z39.87 elements, which would be populated through technical metadata XMP harvests out of fileheaders. This won't give us a complete set of NISO Z39.87 elements, but at least it makes what's there visible / accessible. We'll have a lot more info on all of this up soon on the Automatic Exposure website at http://www.rlg.org/longterm/autotechmetadata.html (which right now is woefully out-of-date). One of the issues still is how you get the data out of the XMP XML RDF buckets on a batch basis. Currently, you can save the metadata for individual files, but there doesn't seem to be a straightforward mechanism yet to extract metadata out of whole directories of files. We're talking to an XMP Product Manager about what the options could be. Hope this helps! Cheers, Günter >Bill Lefugy raises a good point about PDF/A and XMP's use of RDF for >metadata... not having read the XMP specs in any detail, does anyone know >what sort of metadata XMP will really include? Descriptive I know, and >rights, but also technical? I'm wondering if XMP metadata would always fit >nicely in our dmdSec metadata bucket, or it if would actually span several >of the METS metadata categories. >MacKenzie Günter Waibel Program Officer/RLG **note new address** 2029 Stierlin Court, Suite 100, Mountain View, CA 94043 USA voice: +1-650-691-2304 | fax: +1-650-964-1461 [log in to unmask]