It’s official: we at NINES are revising the genre lists we use for our RDF.
The original impetus for this change was a thought-provoking ARC meeting we attended a few months ago, in which we discussed expanding Collex to include several new nodes—MESA, REKn, ModNets—which will aggregate digital objects from the medieval, Renaissance, and Modernist periods, respectively. Once we began to consider objects outside the eighteenth and nineteenth centuries, we found that a many of our genre tags became unhelpful or even irrelevant. How, for example, could we call a manuscript a descriptive generic term, when such a large portion of the text-based medieval objects would be in manuscript form? Would such a term be useful any longer? What about genres like “family life,” “life writing,” and “ephemera”? How were we defining these terms and could we sustain a consistent definition across time periods?
These questions led me to wonder, in a more abstract sense, about how and why we use metadata at NINES. Every time we aggregate a new collection of digital objects, we append to each piece several descriptive elements, so-called data about data. These elements allow us to preserve certain characteristics about each object—its form, its subject, its applications, etc. Tagging such elements also allows us to catalogue and search for items by group or type from among the ever-growing collection. (It would, after all, be difficult to sift through 992,000 digital objects without any sort of filtering system.) It also gives us a sense of what NINES can make available to our users, as well as the strengths and weaknesses of our various collections. At the moment, for example, we offer our users 17,869 objects on the subject of “Education,” but only 196 on “Architecture.”
The danger, however, is that tagging an object allows us to preserve only some of its characteristics. Those that we do not tag we implicitly dismiss as “less important.” We create our list of genre elements with care, but we cannot anticipate every eventuality. In many instances, we will undoubtedly fail to prize important elements, which will then be lost in the shuffle.
What’s more, in separating and delineating the objects available to readers, we also make judgments about the users viewing them. Every time we choose to tag something as “Education” or “Architecture,” we surmise not only about what that object represents, but also about the person to whom that object will be helpful. Someone working on a project about education, for instance, will likely not search through the “Architecture” genre. As aggregators, though, we cannot know how different projects will employ individual objects. By tagging the objects ourselves and at our own discretion, we potentially limit not only their meanings, but also their applications. We can, of course, tag a particular digital object in several different ways, but the crux of the issue remains: by making judgments about the nature and meaning of the digital objects in these collections, we influence the way our users view and interact with those same objects.
Now, as we go about reconsidering our metadata for NINES and for the other ARC nodes, we have an opportunity not only to reconsider the objects in our archives, but the way we want viewers to study them. If we can change our genre elements, how will that change the way others’ view them? How can we create the best system to maintain a well-categorized, well-differentiated collection of objects while still allowing our users the freedom to make their own choices about their purpose and implications?