Amanda Wilson of the Ohio State University Libraries delivered the final presentation of SAA2006 session 305 (Extended Archival Description: Context and Specificity for Digital Objects), Dynamic Duo: Enhancing Access through Dual Description with EAD and TEI. She described a proof of concept project designed to explore if EAD and TEI can be used to support a humanities professor who has students learning how to digitize and add markup.
She provided the following list of example sites:
-
LEADERS Project ‘Linking EAD to Electronically Retrievable Sources’ – transcriptions, original images and metadata
-
Barren Lands Digital Collection – University of Toronto, using EAD to enhance item level descriptions.
The professor’s goals for this year’s project are to create a home page that includes a collection description, document the scholarly process and follow markup rules. Amanda got a big cheer for saying she was “not sure if you can keep scholarly process in EAD – but hey, I’ll try anything”.
For each item being digitized the professor wanted to include all of the following:
-
Markup
-
Reading View
-
Diplomatic View – summary of all marks
-
XML source (aka TEI file – AACR2 bibliographic record could be created from this data)
How can she replicate the process the class was going through to support these goals? First, she picked the software created by DLXS that was already being used on site.
During the course of her research, she had to come up with methods to do all of the following:
-
Retrieve metadata and images
-
Convert image to TIFF
-
Figure out how to connect from EAD to the TEI information
-
Load TEI and EAD files and images into DLXS
The solution would need to support the community and the work they have already done. Amanda’s vision was to permit addition of item level description to the EAD with no additional editing AND load the TEI with little or no modification. It was a challenge to massage the TEI to validate against the Document Type Definition (DTD). She also wanted to link back to the original site created by the professor’s students in order to retain extra information that had no place in the EAD.
At about this point I was wishing (for the umpteenth time at SAA) that there was internet in order for online demos.
There were definitely challenges. For example, DLXS is aimed at eBooks, while TEI has additional fields such as those required to support properties related to “hand” (as in who wrote the scanned content). This makes it hard to change the TEI format to fit into DLXS. Using TEI in DLXS does permit searching for individual items, but they may need to do additional massaging to get the data to line up with the fields expected.
The conclusion of the presentation was that it can be done. It is possible to integrate the professor’s transcriptions using EAD and TEI within DLXS, but there will need to be more discussion with the faculty member about their requirements. The ultimate aim is for a federated search that is integrated into the institution’s central search. Those who are working on similar projects may be interested in moving to using a standard, but only up to the point at which they start loosing the data that is important to them for their research focus.