The Cover Pages: Image Metadata Aggregation for Enhanced Searching

Date:      Mon, 12 Mar 2001 14:11:53 -0800 (PST)
From:      Chuck Thomas <thoma134@tc.umn.edu>
To:        Multiple recipients of list <xml4lib@sunsite.berkeley.edu>
Subject:    Image Database/DTD Question

Hi Clayton and list,

I see so many mentions of TEI and EAD used for image description. Generally speaking, you can PROBABLY fit anything you want to say about digitized images into these two encoding schemes somewhere, but it's done at the loss of specificity. In other words, lots of information has to be grouped into more generalized tagging.

Other available metadata schemes could include Dublin Core, a MARC-based dtd, and numerous others. Based upon my years of previous work at trying to adopt one descriptive scheme for all digital library content, I'd day there is no one satisfactory descriptive scheme available to DL implementors for everything they digitize. This being the case, when we set about digitizing lots of digital images at the University of Minesota Libraries, we looked around for a metadata scheme that provided satisfactorily precise description, but was also compatible with other metadata schemes. DC was too general, EAD was not made for item-level description in this way, TEI's usefulness was limited, and MARC was too complex unless only librarians were using it.

Continuing our search, the VRA Core and the Getty's Categeories for Description of Works of Art both seemed like good candidates, as did Britain's Visual Arts Data Service metadata scheme. Upon further consideration, however, I decided that VADS and the VRA core were SLIGHTLY not extensive enough, Getty's CDWA was too complex to promote widespread adoption on campus. You should also know when I say "visual resources" I mean a wide range of visual content, not just fine art or photographs.

Therefore, we created our own customized DTD to support a campus-wide IMAGES (Image Metadata Aggregation for Enhanced Searching) project. A copy of the dtd is attached. It's a fairly straightforward standard, and was deliberately designed as a "flat" sturcture (not very many attributes in its present form) to promote widespread adoption as a campus standard for describing scanned image collections. The tags include descriptive, administrative and structural metadata.

A sample of the encoding and output with a stylesheet can be viewed with IE 5.0 or higher by going to

http://digital.lib.umn.edu/IMAGES/1poster.xml

This particular example does not use all of the tags available with images.dtd, but it uses most of them.

As to the matter of interoperability, I have quite successfully been able to dumb down the information with stylesheets to output Dublin Core metadata as needed for groups of records, and thus am poised to contribute to national metadata harvesting initiatives when the time comes. At the present time, we have thousands of scanned posters, photographs, fine art, etc. described in an xml implementation of the dtd, and are making a purchase decision about the best software for the metadata aggregator.

Comments on the DTD, or similar experiences, are welcome.

Sincerely,

Chuck Thomas
Digital Projects Coordinator
Elmer L. Andersen Library, Room 6
222 21st Avenue South
Minneapolis, MN 55455
612-625-0028
thoma134@tc.umn.edu

Prepared by Robin Cover for The XML Cover Pages archive.

SEARCH Advanced Search ABOUT Site Map CP RSS Channel Contact Us Sponsoring CP About Our Sponsors NEWS Cover Stories Articles & Papers Press Releases CORE STANDARDS XML SGML Schemas XSL/XSLT/XPath XLink XML Query CSS SVG TECHNOLOGY REPORTS XML Applications General Apps Government Apps Academic Apps EVENTS LIBRARY Introductions FAQs Bibliography Technology and Society Semantics Tech Topics Software Related Standards Historic	Date: Mon, 12 Mar 2001 14:11:53 -0800 (PST) From: Chuck Thomas <thoma134@tc.umn.edu> To: Multiple recipients of list <xml4lib@sunsite.berkeley.edu> Subject: Image Database/DTD Question Hi Clayton and list, I see so many mentions of TEI and EAD used for image description. Generally speaking, you can PROBABLY fit anything you want to say about digitized images into these two encoding schemes somewhere, but it's done at the loss of specificity. In other words, lots of information has to be grouped into more generalized tagging. Other available metadata schemes could include Dublin Core, a MARC-based dtd, and numerous others. Based upon my years of previous work at trying to adopt one descriptive scheme for all digital library content, I'd day there is no one satisfactory descriptive scheme available to DL implementors for everything they digitize. This being the case, when we set about digitizing lots of digital images at the University of Minesota Libraries, we looked around for a metadata scheme that provided satisfactorily precise description, but was also compatible with other metadata schemes. DC was too general, EAD was not made for item-level description in this way, TEI's usefulness was limited, and MARC was too complex unless only librarians were using it. Continuing our search, the VRA Core and the Getty's Categeories for Description of Works of Art both seemed like good candidates, as did Britain's Visual Arts Data Service metadata scheme. Upon further consideration, however, I decided that VADS and the VRA core were SLIGHTLY not extensive enough, Getty's CDWA was too complex to promote widespread adoption on campus. You should also know when I say "visual resources" I mean a wide range of visual content, not just fine art or photographs. Therefore, we created our own customized DTD to support a campus-wide IMAGES (Image Metadata Aggregation for Enhanced Searching) project. A copy of the dtd is attached. It's a fairly straightforward standard, and was deliberately designed as a "flat" sturcture (not very many attributes in its present form) to promote widespread adoption as a campus standard for describing scanned image collections. The tags include descriptive, administrative and structural metadata. A sample of the encoding and output with a stylesheet can be viewed with IE 5.0 or higher by going to http://digital.lib.umn.edu/IMAGES/1poster.xml This particular example does not use all of the tags available with images.dtd, but it uses most of them. As to the matter of interoperability, I have quite successfully been able to dumb down the information with stylesheets to output Dublin Core metadata as needed for groups of records, and thus am poised to contribute to national metadata harvesting initiatives when the time comes. At the present time, we have thousands of scanned posters, photographs, fine art, etc. described in an xml implementation of the dtd, and are making a purchase decision about the best software for the metadata aggregator. Comments on the DTD, or similar experiences, are welcome. Sincerely, Chuck Thomas Digital Projects Coordinator Elmer L. Andersen Library, Room 6 222 21st Avenue South Minneapolis, MN 55455 612-625-0028 thoma134@tc.umn.edu Prepared by Robin Cover for The XML Cover Pages archive.