Structuring the Metadata
By Molly Roy, Ph.D. How do we render embodied performance art as data? This is an ongoing question that emerges across different phases of the project, in both philosophical and practical terms; a question to sit with and think through, in relation to how we want to represent the work, or put differently, how to structure the metadata. There has been some push and pull between wanting our system to be compatible with others—to be transferable, searchable, locatable—which means aligning with prescribed standards and structuring our data as others have, and on the other hand, wanting it to convey the rich specificity of No Boundaries and the perspectives of critical dance studies. The Dublin Core metadata schema was our starting point, and a series of modifications have followed.
In our first data model, the creator of a work was classified as the Artist, which we changed to Choreographer, to bring in more specificity and the particular vocabulary of dance. Moreover, everyone involved in a performance project is an Artist. It didn’t feel accurate or appropriate to list an Artist, then musicians and designers—all artists as well.
While switching from Artist to Choreographer was a simple choice, the question of what role to assign to those who dance or perform in a choreographic work proved to be trickier, and we went back and forth between the descriptors Dancer and Performer. There is no definitive right answer here. It can be easy to go around in circles, and it may seem to some extent like splitting hairs. Nonetheless, the small details of representational structures carry weight and can shape conversations over time, and therefore every element warrants attention. Initially, we leaned toward Dancer, as one of our guiding principles has been to foreground the dancing body, to intentionally uplift dancers within the archival space. However, as we started importing information from programs, the No Boundaries performance programs typically listed performers rather than dancers, and we began to wonder if it made more sense to align with those original information records instead. There was also a question about folks who may be performing in a dance piece but are not themselves dancing per se—actors or poets, for example. They are better described as Performers. We considered including both Performer and Dancer as roles in the data structure, but that transfers burden to the cataloguer, creating unclear distinctions and requiring a certain level of knowledge about each piece and about the form, which might not be shared by everyone tasked with cataloging. Ultimately, we proceeded with Performer, though not without mixed feelings.
The more complicated the metadata schema, the more challenging and time-consuming it becomes to catalog materials, a balancing act that pulls focus toward another key question—who is, in fact, doing the cataloging? We have tried to dig into this question, not just as a practical or logistical matter, but in terms of labor, compensation, and sustainability; uncertainties around specialized knowledge or subject-area expertise, and how different people with varying degrees of familiarity with No Boundaries and/or with dance studies will categorize things differently, and make different choices about how to represent the work. It is important to be clear about what is being asked and what is being expected in this area, and what resources there are in place to appropriately support that labor.
Classification of genre was another component we struggled with. While less relevant for No Boundaries specifically, genre seemed like a reasonable, common way that people might search for dance content. Thus, if we are aiming to build a model that is broadly applicable and easily searchable, perhaps including genre would be a good idea. However, genre categories are blurry, unstable, and often problematic. Eurocentric and white-supremacist language has historically been baked into information classification systems, and the field of dance is no different in this regard, with genre terms like “ethnic dance” and “world dance,” which we do not want to perpetuate. In addition, genre can be entirely subjective, as choreographers make individual choices or feel particular affinities with how they describe their form and practice. In the cataloging workflow, having a drop down menu of fixed, available dance genres will necessarily be exclusionary or unwieldy. On the other hand, having an open text field creates its own set of problems and inconsistencies, inevitable typos, and a reliance on specialized knowledge, all of which can hinder searchability, negating the original intent in even including genre in the first place. We considered a range of options, but finally scrapped genre as a descriptive field, because in the end, it did not serve No Boundaries, even though others using our prototype might miss it.
The old adage, you can’t please all of the people all of the time, became a useful reminder in these situations, a way to pull ourselves out of the rabbit hole. Choices have to be made, and those choices will necessarily be limiting in some sense; classification systems inevitably foreclose certain possibilities. Even as we think about other users, we ultimately have to make choices that most honor the collection at hand.
That being said, we do want to connect with others, a motivation that has fueled investigations into linked open data, with the goal of improved interconnection between the No Boundaries materials and other content sources. Explorations into Wikidata have yielded exciting potential, yet Wikidata integration is not without its own share of tensions. For example, looking at the existing entity properties and relationship types in Wikidata, in order to maximize connections, we have also encountered language that we find problematic. At one point, we pulled a chunk of data into our prototype from Wikidata, and the term ballet master ended up as one of the possible roles, which we were not at all comfortable with. Needless to say, this exchange of data requires its own monitoring processes. There are also legitimate concerns around wiki platforms, and the possible interjection of erroneous or misrepresentative data. Yet such platforms also have the potential to challenge structural hierarchies in knowledge production. What is the line between shaping representation and simply gate-keeping? The long-term vision for the No Boundaries/Dancing Digital collaboration is to use these explorations to develop spaces where knowledge can be produced from below, above, and in between.