I just spent a very pleasant weekend in Asheville, NC. I interviewed an
academic librarian. One of the areas of interest and discussion was what are the technology challenges of a librarian?
Librarians, as it turns out have very similar technology problems to the rest of society? Who knew? One of the benefits of librarians over most IT people is that they understand what metadata is already.
Metadata describes data.
During our conversation, it was very apparent that while many curators and librarians deal in physical “data”, the data is increasingly being digitized. The digitization presents some very interesting issues:
- In what format will the data be stored?
- What metadata will be stored?
- For example, if a piece of history is written on papyrus, the library has to decide not only how to store both the papyrus, but also pictures of the physical nature of the papyrus, the makeup of the papyrus, any ink residue, etc., etc.
- Over time, more and more metadata gets collected, and more and more determination of relevancy needs to be decided.
- What format will the data be store?
- If it is initially stored in (say) JPEG, what happens when that JPEG format is updated to include more metadata or resolution? Will the original be preserved?
- How will the metadata (or copies of the artifact) be cross-referenced?
- A perfect example would be examples of the signature of signed pieces? How will a facsimile of that signature be used to compare against others?
Ultimately, most librarians actually have more problems with storage than most IT professionals. Of course, this should not be a surprise. Records managers (the precursor to IT professionals) have many of the same issues.
Whether it is an academic librarian, a public librarian, art curator, the problems posed are identical. What my conversation ultimately illustrated is that despite all the technology advances our society has progressed, the basic issue of classification has yet to be solved.
Computers can only classify as detailed or as quickly as we, as humans, can define. While machine learning helps, computers can only be as “good” as humans ourselves!