It was our original intent to create a database and user interface that could hold every bit of information necessary to create the metadata for each resource. The database diagram on this site represents the attempt to do just this. The database was designed and implemented completely. The first phase of the user interface was also completed and screen prints can be found on this site.

As we progressed through this development, the enormity of the task became apparent. Making one application to track the details, hold every bit of metadata, and make each element editable and repeatable is a Herculean task.

We have now returned to the idea that the tracking database should primarily track the progress of each resource through the digital library. It does include some basic bibliographic data necessary to identify the resource, and it still does aid in the creation of metadata. However, we do not expect it to hold every bit of data.

When we begin working on a resource, this resource is either entered into the application manually, or imported from a MARC record. Whenever a MARC record is available, this is converted into MARC-XML and stored in a central depository. An example MARC-XML file can be found here.

When the scanning of the resource is complete, the Quality Control Application is used to create structural metadata for the book. This allows users to add the divisions and division hierarchy for each resource. When the resource leaves the Quality Control department, a METS file is created from the data in the tracking database, the MARC-XML file, and the data entered by the tech. A sample METS file can be found here.

The resource is OCR’d and then placed on line, using only the METS file to load into Greenstone and the University of Florida Digital Collections [UFDC]. During this process portion of the METS file are converted to the Greenstone Archive [GSA] format for loading. The sample GSA can be found here and the loaded resource used for each of these examples can be found online here.

Development continues on the application with changes likely to be seen to the database as well. New resource types, such as newspapers, will be added with their own unique properties. Additionally, audio and video has not yet been implemented. As with all large projects in the changing environment in which we operate, the evolution of this project continues.