Gunther and Perian - To give you an idea of some "problems." Right now I am managing a scanning project for 19 California libraries whose images will reside in a digital repository of the California Digital Library. We are embedding tech metadata for preservation purposes: hardware used, vendor, owner(library), and a short descriptive field. We had hoped to also include color information, but have opted to scan a target with each object. We are using jHOVE to validate the Tiff format, including the tiff header format. Initially if rejected several image files, (50%) as did Adobe's program Elements. The problem, as we determined it, was that when the vendor used Adobe programs to process the files, they interjected their own header information in such a manner as to "corrupt" the files. This was a problem also noted by Northwestern University Library. Removal of all header information that was not explicitly requested solved the problem. The Vendor is batch loading the metadata from a spreadsheet that the libraries' provide. They will be "correcting" one batch, a library added a column for their own purposes which through everything off, but they will be doing it by simply overwriting the existing.
I believe maintaining the tech and some administrative metadata with the file has a lot to say for maintaining the integrity of the image file, but second Guenther's concern that storing all the descriptive metadata with file may not be wise for the reasons that he lists. Also for those looking for a vendor, we competitively bid this project (using RLG's great template for an RFP) and have been very happy with both the process, service and product that we have received from the winner, Northern Micrographics/Brookhaven Bindery 2004 Kramer Street La Crosse, WI 54603 Phone: 800-236-0850 ext. 107 Fax: 608-781-3883 [email protected] http://www.normicro.com If anyone requires more information, I will be glad to correspond with them privately. Trudy On 2/1/06 8:54 AM, "[email protected]" <[email protected]> wrote: > Hi Perian, > > I'd be curious to hear what mechanism you'll use for embedding descriptive > metadata into the image files - it has been an idea often pondered and > rarely implemented, as far as I know, but I'd be happy to stand corrected > by others who have pulled it off. > > There are various technical challenges in embedding the metadata (where > does the metadata sit within the file?), and various challenges of > procedure and maintenance (e.g. scalability - how do you get descriptive > records into huge numbers of files automatically; how does the data get > updated if it changes?). In theory, you could use Adobe XMP to embed any > type of metadata into a digital file, including descriptive metadata, but > at the end of the day, the logistics of getting the data in and keeping it > up to date may outweigh the obvious benefit of having such a tight link > between file and description. Maybe embedding a unique identifier within > the file which can be resolved in your database of record (your Collections > Management System, for example) makes more sense. I'd be curious to hear > from others how they maintain the link between image file and descriptive > metadata - I suspect the identifier route, often via the filename (i.e. the > unique id is in the filename), is what most people these days practice. > > As for mixing different descriptive approaches: if your goal is to have > these digitized images searchable from within one interface, you'll > eventually have to find a common denominator among different descriptions. > If you're starting from scratch in describing them, it would seem > worthwhile to consider whether you can achieve adequate description using > just one descriptive approach. If not, you'd really want to keep in mind > that these different descriptions will have to map to one another down the > road. > > Cheers, > > Günter > > *** > > Günter Waibel > Program Officer/RLG > 2029 Stierlin Court, Suite 100, Mountain View, CA 94043 USA > voice: +1-650-691-2304 | fax: +1-650-964-1461 > blog: www.hangingtogether.org > [email protected] > > > > <[email protected] > g> > To > 01/27/2006 03:32 [email protected] > PM cc > > Subject > Please respond to Embedding metadata into digital > [email protected] images - which schema? > > > > > > > > > > > Hi again: > > > As I've mentioned before, I'm working on drafting our digitization > procedures here at the Judah L. Magnes Museum, a Jewish cultural history & > art museum. I'm trying to determine the best metadata schema to embed in > the digital images. I'm curious as to what, if any, schema other cultural > history & art organizations use (although we do also have some > archeological and archival objects in the collection) in their images. > > > For a mixed collection such as this, is it a good idea to mix schema? I > would like to use MARC for the archival objects that are mixed in (as our > archive center is already using MARC for everything), but I might end up > using Dublin Core or SPECTRA (or whatever) for the other stuff. Or is it > wiser to just pick one and "make it fit" as best I can? > > > Thanks, by the way, for all of your help earlier regarding your digitzation > procedures - your helpful links have caused my collection of useful guides > to grow considerably. > > > Perian Sully > Collection Database & Records Administrator > Judah L. Magnes Museum > 2911 Russell St. > Berkeley, CA 94705 > > > --- > You are currently subscribed to mcn_mcn-l as: [email protected] > To unsubscribe send a blank email to > [email protected] > > > > > > > --- > You are currently subscribed to mcn_mcn-l as: [email protected] > To unsubscribe send a blank email to > [email protected] > --- You are currently subscribed to mcn_mcn-l as: [email protected] To unsubscribe send a blank email to [email protected]
