Tuesday, November 12, 2013

Comparison of OAI Automatic Metadata Harvesting versus Manual Metadata


I had to draw from my memory from LIS 640 for cataloging and what I remember about MARC record creation, as well as researching AACR2. Of course, much of the language of cataloging and metadata can be difficult to understand if you are not using it in everyday context. Sometimes I just had to make my best guess as to what information to put in the various fields. I harvested the metadata for the Atsina Fly Dance photograph by Edwin S. Curtis from the Library of Congress into DSpace. Then I created the manual collection and attempted to create what I felt was the appropriate information. I took many of my clues from the actual metadata on the Library of Congress web link as well as referencing the actual MARC record and trying to apply AACR2 methods. When I compared my file to the automated file I was pleased that I was either correct or close on many of the fields. However, the harvested metadata had more depth and more fields of information than my record. Of course, there were some fields in my manual record that were simply wrong or the information lacked the detail of the harvested record.

So what do I take away from this exercise? For one thing the automated harvested method will in the majority of cases be better.  I personally know that our Metadata and Serials Librarian where I work relies heavily on copy cataloging and downloading MARC records and metadata generated by the Library of Congress. One could apply the axiom that it is better to have too much information than not enough. Once a record has been harvested the metadata could be manipulated in order to make it conform to a particular integrated library system if needed. At least that is my theory.
Another reason this assignment seems important is that my understanding of the function of metadata is to better facilitate retrieval of information over a vast scope of networks. The DSpace communities allow for a common language and ease of access to aid in research. The automated harvesting is also a vast time saver if one was attempting to aggregate sources and information.

Here is the link to my DSpace Community: