Ask Us!
University of Washington Libraries

William Brumfield Russian Architecture Digital Collection

Explore Collection

Image and building descriptions (metadata)

Professor Brumfield, while scanning, used Adobe Photoshop to embed XMP metadata in his TIFF files. Mostly he identified building names and their locations by city.

The XMP metadata was extracted using the JSTOR/Harvard Object Validation Environment (JHOVE) then perl and XSLT for further extraction. The resulting XML was imported into an MS SQL Server relational database, which was later migrated to postgresql. Both databases were accessed using a Microsoft Access front end with an ODBC connection to the database. Data entry forms and queries were run from Microsoft Access, and across ten years many people contributed project metadata.

The database data dictionary (Download PDF) was based on VRA Core 4 properties. The goal was always to output and reassemble the relational data using a local VRA Core 4 implementation. This was accomplished in fact by converting the relational tables to XML and writing XSLT and XProc to output VRA Core 4 XML according to our VRA Core 4 implementation model. The resulting VRA Core 4 was finessed until it validated using the VRA Core 4 strict schema.

Using the VRA Core XML, a tab-delimited text dataset was created using XSLT and XProc. This required creating records for vra:image only, merging related vra:work data into the image records. The resulting text was uploaded into CONTENTdm, the content management software used for the public digital collection.

The VRA Core XML is also being used to generate an RDF-XML data set for download to anyone who would like to repurpose our descriptions elsewhere. As of August 10, 2016, this is a work in progress.


Back to top

Follow Us:

Twitter Flickr