1. Integrating Non-MARC Metadata into a Traditional Technical Services Department
At the Univ. of Tenn., we have been about the work of integrating non-MARC metadata production into our traditional technical services department.

2. Melanie Feltner-Reichert
Our topic
Integrating Non-MARC Metadata into a Traditional Technical Services Department:
Perspectives from Librarians at the University of Tennessee Libraries
Melanie Feltner-Reichert
Marielle Veve
During our time together today, Marielle and I would like to share our perspectives as librarians in the department. Marielle Veve is the Cataloging Librarian for our department. And I, Melanie Feltner-Reichert, am the Metadata Librarian.

3. Why integrate? Priority to digitize our Special Collections
Recent adoption of Metadata Object Description Schema
Development of metadata creation tool
Keep skills of Technical Services staff current, competitive, and cutting-edge
First of all, we’d like to give you a little background as to why we are integrating metadata into our department. First, our library administration has a strong commitment to the development of digital collections, and among all our digital projects, we have set as top priority the digitization of materials from our own Special Collections.
Additionally, in early 2006, we have moved to creating richer descriptive metadata records to enhance access to our digital content. We had been using local metadata formats and Qualified Dublin Core, but we have adopted a richer, more flexible schema that can more adequately describe the cultural heritage materials we are digitizing. That schema is the Metadata Object Description Schema. It’s a Library of Congress standard that shares many similarities with MARC. Given its similarity, our catalogers’ knowledge of MARC transfers nicely to working with MODS.
Then in the 2006, we experimented with a couple of methods for creating the MODS records, and learned that we needed a simple tool to make the process easier. So we developed our own metadata creation tool in-house, to allow record creators to focus on the description itself rather than technical issues. This opened up the work of metadata creation to catalogers who not skilled in XML (Extensible Markup Language), but who have much to bring to the task given their long history of bibliographic description.
So, with a richer metadata standard and a simple tool both in place, we were ready to bring more people on board to create metadata records.
And then finally, we chose to integrate metadata into our Technical Services staff in order to keep their skill set current, competitive and cutting-edge in this world of exploding digital content.

4. Preparation Introduce the project
Emphasize connections between MARC and MODS
Participation on a volunteer basis
Team building
So what were the first steps we took to prepare the team for metadata integration?
An obvious first step was to introduce the project to the team as a whole. At one of our weekly team meetings, I described the project, its mission and the standards we’d use to do our work. I also demo’d our Metadata Workbook (the metadata creation tool we developed) to illustrate the process. And also demo’d the digital collections to show the end product.
During that same presentation, I emphasized the connections between MARC and non-MARC metadata creation, to highlight all the strengths the catalogers would bring to the task, and also emphasize that while they would need to learn new tools and standards, many of their current tools would apply in the non-MARC environment. --building buy-in and ownership; team building and rapport
After this presentation, we called for volunteers to form the metadata team. Every team member was given a choice as to whether to participate or not. No one was required to participate. This allowed us from the very start to build buy-in and ownership among the team. Because it was their choice to participate and we later included them in the planning of both the training and implementation stages, the team was able to take ownership of the project as a group. To provide an informal setting for discussing the project and giving input, I hosted a brown bag session where the team members were able to raise questions, give suggestions for structuring the training, and express concerns about the projects’ impact on their workload. This session set the tone for the group, allowing us to gel as a team, and build rapport with each other.

5. Training 12 hours of training New standards New tools
Emphasizing hands-on practice
New standards
Metadata Object Description Schema
Text Encoding Initiative
Encoded Archival Description
New tools
Metadata Workbook
Administrative Database
XML editor
Once we’d gathered the input from the team, we set about designing the training. Because the team expressed a strong preference for “learning by doing”, we structured the training sessions with lots of hands-on practice time. Over the course of a single week, we spent 12 hours together learning how to apply MODS to our Special Collections materials. Each session started with a short presentation to introduce a new standard or tool (about 1/3 of the session), then closed with hands-on application of what had just been presented (about 2/3 of the session). The following day’s session would open with a quick review of what was covered the previous day, then followed the same pattern on presentation, then practice.
What were some of the new standards that the team had to learn? Certainly, the most important was our metadata format, MODS. Then we also had to learn about TEI (Text Encoding Initiative), as well as Encoded Archival Description. While the team would be creating MODS records, these other two standards were critical because we were cataloging text-based materials from our Special Collections holdings. The first items the team would catalog were letters from a single archival collection. These items were already digitized and transcribed in TEI encoding. To catalog the items, the team would receive the scans of the letters, as well as the transcription in its TEI XML encoding. So, it was imperative, that they be able to navigate a TEI document.
Then finally, there would be key descriptive context for the archival collection as a whole that the team could glean from the Finding Aids. We were able to present the encoded finding aids, however, in html, so the team only needed a cursory understanding of the structure of a finding aid, and did not need to be able to navigate the XML encoding.
Next, there were new tools for our non-MARC metadata creation:
Our metadata workbook: It is form on the web in which we input the metadata content, then the workbook generates the XML markup. The workbook also performs some important quality control measures, for instance, ensuring that required fields are completed, that fields with controlled values conform to the accepted values, and that dates conform to the w3c formatting we require.
A second tool we use is the Administrative Database. This is a tool in which we track all digital content submitted for review/consideration for inclusion in our collections. Additionally, it generates unique file names for each digital object as their associated XML records, the values of which are then transferred to our Metadata Workbook automatically when it’s cataloged.
And lastly, because the team would receive TEI XML files, they needed to become familiar with an XML editor.
So there were a number of new tools and standards for the team to get familiar with

6. Training
Applied tools and skills used in traditional cataloging to non-MARC metadata
Shared standards and tools
MARC Genre List
MARC Relators
LCSH
LCNAF
OCLC Connexion
But I’d just like point out briefly that a number of tools/skills *did* carry over from their traditional cataloging work, and here are a few of those shared items: marc genre and relator lists, library of congress subject headings, library of congress name authority files, and also, oclc connexion. So the team was not entirely in *new* surroundings. Much of what they already knew came to bear on their with non-MARC metadata.

7. Implementation Pilot phase. Timeline: Two-week pilot period
Task: Catalog 12 letters in MODS.
Each team member received image files and transcription (TEI) for 12 letters from a single archival collection
Objective: Practice record creation process. Refine workflow and tools.
So, once the building blocks were in place, and the team knew how to navigate the tools and apply the standards, we launched a pilot phase of implemented all we’d learned.
For this pilot, we carved out a two-week time frame. During which each team member was asked to catalog 12 letters in MODS. To do this, each team member received the image files and transcription files for their letters, which were all from a single archival collection.
The objective of this pilot phase was the practice the record creation process. And to identify glitches in the workflow and tools, so we could address them before launching into a routine production phase.

8. Implementation Evaluation of pilot phase. Assessment of records
Gathering feedback from team
Once the pilot phase was completed, we spent a few weeks assessing the pilot experience. The records went to the Metadata Librarian for a first review for adherence to the guidelines, then they were passed on the Programmer to ensure there were no technical difficulties with the records (examples) and could be ingested into our delivery system. Records go to Metadata Librarian and DLC Programmer to assess for adherence to guidelines
Another critical component of this evaluation period was to gather feedback from the team. What were some of the problems you encountered? Do you need additional training and resources? Were there functionality issues with the tools? Are there gaps in the documentation and guidelines?
Through this conversation, as well as examination of the records, we were able to identify a number of issues we needed to address before moving on to our production phase. Marielle will now talk about those issues, how we have sorted through them, and wrap up the presentation.

9. Surprising outcomes Functionality issues with tools
Errors in TEI (transcription)
Duplication of effort in TEI Header and MODS record
Many inconsistencies in names not in LCNAF
Workflow needs modification
Issues that emerged out of the pilot phase.

10. Changes to Workflow New approach to authority control:
Authority files should be done first, before records are cataloged
Authority files should be done by only one person (Cataloging Librarian) to avoid inconsistencies
Metadata team members then ready to catalog using MODS

11. Changes to Workflow New approach to TEI Header:
Adopt Minimal TEI Header for initial transcription
Letter is then cataloged in MODS
A fuller TEI Header can then be generated from the MODS record
Established reporting process for issues with the tools and guidelines, as well as enhancement requests.

12. Expanded collaboration
Prompted collaboration among Metadata Librarian, Cataloging Librarian and Archivist
Crosswalk development
EAD -> MARC21

13. Current status Experimenting with workflow Revising pilot records
Cataloging new resources (priorities set by Special Collections)
Using TEI markup and image files; as well as new authority files established by Cataloging Librarian

14. Future directions
Expand metadata duties (branch out into other markups: EAD)
Expand integration (all department)
Train other team members to establish AF
Continue improvement of tools
Survey team members

15. Contact Melanie Feltner-Reichert Metadata Librarian mfeltner@utk.edu
Marielle Veve
Cataloging Librarian