In one of my classes this semester, we’re learning about Research Data Management. We’re exploring policies, funding requirements, best practices, history, theory, and practical details for the research data management life cycle. I’m having a great time and learning a lot.
One of the things we’re learning about is how to define a good organizational strategy for file naming and folder hierarchies. Just like call numbers in a library describe where a book goes on the shelves, files created as part of the research process have to live on a computer disk somewhere, and these file and folder conventions help produce organizational structure to make it easier to locate and manage all of the files produced during a research project.
However, having a fixed structure for filenames and folder hierarchies has its downsides. If I choose to put the date first in the filename, sorting by filename won’t have all of the files for a specific experiment together. If I separate the raw data into a different folder from the processed version of that data, it’s more work to connect those two files. Any structure that we define will be optimized for one use case, and detrimental to others. It’s a trade-off.
When learning about RDA last semester, I got to take a deeper look into the Resource Description Framework, a standard for creating linked data. Having the ability to create semantic connections between research data files would be powerful. A raw data file could be linked to its processed version, which could in turn be linked to the code that did the processing and the figures created from the results, which could be linked to the article drafts it’s referenced in. Creating these kinds of links isn’t possible just by arranging files into a hierarchical folder structure.
I’m sure I’m not the first person who has considered the RDM organizational problem from this perspective. But now I have a burning desire to see what kinds of tools researchers have created to manage these semantic relationships. Down the rabbit hole I go.