I had some excellent comments from my last post and I wanted to return to some ideas that came from these comments. One noted that the quality of metadata is a ticking time bomb. In both comments, issues about the place and increasing reliance on technology was raised.
Metadata quality is a hot issue. Where I work, it is a struggle to be able to respond to the number of requests and provide good metadata, especially if it isn’t there to begin with. Let me explain. Part of my job is to take EAD records and create MODS records for our digital repository. The collection level portion of these EADs are full and impressive records. The amount of work put in by our archivists, curators and their pages is tremendous. The issue, however, is that the item level descriptions are less than full. I have tried to add common subject and genre terms and any other information that can be set to a default. The reason is that I typically transform hundreds of EAD records at a time that in turn create even more MODS records. In the last 3 months, I’ve created over 20,000 MODS records. As the only person on this project, it becomes more than a challenge to go back and enhance all of these records. It is not that these records are less than minimal. They still allow for discovery and accessibility. However, they are not really unique. I would say these records are average and not above average or even good in terms of providing good or above average unique descriptions. Now the good news is that help is on the way slowly and surely. Also, I keep annoying my supervisors for little helper elves to handle enhancement. And this brings me to my second point about technology.
My main focus is to get these records into the digital repository efficiently and in a timely manner. There are 7 who work in Archives. Then there is the metadata that needs to be transformed from our partner institutions; here I work with a variety of people. Then there are others around campus including those who need help with metadata. All of this is to illustrate that I deal in bulk produce! I write transformations creating little MODS records by the hundreds. This is an automatic approach that many use. Some do this in Oxygen, the xml editor, others in MarcEdit, and others rely on programmers who have written scripts to create these records. These various technologies have been and continue to be a lifesaver. I and other metadata librarians rely on it every day to do our work. However, when push comes to shove, enhancements are still in the realm of a person adding good quality metadata. A good example is that the transformations transform the data that are present. If that data is wrong or inaccurate, that is brought over into the new xml structure. Recently, I saw one record that I transformed en mass where the title was in German and not English. In the original data file, there was only the English title; the German was never recorded. It was only thanks to a person who looked at the digital resource to discover this and add the German title to the metadata.
All of this is to say that technology and human knowledge and expertise must work together. Technology certainly can take us far and provide solutions to problems that seem amazing. For the Digital Commonwealth, programmers have written a script that controls place, name and topic headings in spreadsheets. For the this script to work, someone familiar with thee controlled vocabularies had to inform the programmers such that the requirements of the task were and continue to be met by the script. The Digital Commonwealth also has a post review by people to ensure the script and other automated processes worked as expected.
The discussion can’t be about whether technology will supplant people. Technology are tools created by us. In a sense, the better we understand these technologies, the better we can leverage it to assist us in our work. We can avoid the ticking time bomb of metadata quality by ensuring that people don’t get lost, that metadata librarians add value and quality to repositories along with technology, and that even if all of these metadata are not seen that it is thanks to people and technology that cool visualizations like timelines and mappings are possible.