Investigating Metadata and Archival Resources
The three resources attached to this page represent an initial foray into what will likely become a more cohesive research program centered on print culture, interface, and currently emergent digital variations on those topics. The two images and single video also gesture towards the form I hope my final project takes: a curated digital archive compiling examples of textual "strangeness." This project is, necessarily, grounded in the radical decontextualization of print materials to reveal the fluidity of their structures, especially in the period 1476–1660 in England. Thus, the video "Medieval help desk with English subtitles" is a lighthearted yet sophisticated way to begin that reconsideration. The two image files ("The Altar" and an excerpt from the play Ralph Roister Doister) on display are two examples of the type of textual "strangeness" I hope to critically consider in my final project. In "The Altar," George Herbert pens a poem focused on the Eucharistic sacrifice into the shape of an altar; in the excerpt from Ralph Roister Doister, Nicholas Udall displays the chaotic nature of printing norms in the mid-16th century: speech headings appear on the right margin, despite the page being a recto leaf; speech headings appear in-line with dialogue rather than starting a new line; different type indicates the same name being used in different ways, and so on. Both images are small image captures from larger PDF documents.
I encountered various types of difficulty when populating the metadata for each of these items. As the linked video is hosted by YouTube, much of the Dublin Core metadata was directly imported. Even that, however, demanded scrutiny, and I made several alterations to the automatically generated content. The import process was particularly good at transferring subject elements, for instance, but fell short when it came to the creator and publisher elements. The default creator was listed as the username of the video uploader, rather than the organization behind the username, suggesting an automated process of field population rather than human consideration of the video's origin. In my scheme, the creator is the writer of the skit (listed in the YouTube notes), while the publisher became the organization behind the distribution of the content—in this case the Norwegian Broadcasting Corporation (NRK). Much of this information was more straightforward in the case of the two images, excepting the issues outlined below. Titles, creators, and publishers of print materials, it seems, map much more cleanly onto Dublin Core metadata than video or other forms of newer media.
By far the most difficult decision to make seems simple, but to me is a foundational question: when populating metadata, are we referring to the singular instantiation of the text before us or the original object itself, as if it were unmoored from medium? Strictly speaking, the metadata should seemingly be different for the physical edition of Nicholas Udall's play held in Eton College Library, published in 1566 and for a PDF version scanned from a microfilm copy held by most universities in North America and available digitally via Early Electronic Books Online (EEBO). The creator for all three, however, would likely be Nicholas Udall, and the date for all three would be 1566. The bibliographic record of the play in EEBO does not even bother to note when the microfilm was first produced; the sole date is 1566, referring not to the thing to which they are giving access, but to the original play behind the thing they are giving you access to. Those distinctions seem to collapse when metadata is used, at least and perhaps especially when the legacy elements are used. The DC legacy set sometimes confusingly eradicates the distinction between the thing itself and the thing in front of us, which are rarely the same. This is why, as you may notice, the titles of my two images follow the pattern of "facsimile image of xx;" in the strictest sense, it seems foolish to name an image (which is a jpeg produced from a PDF produced from microfilm taken of the original document) as identical to the thing it purports to capture. These sorts of distinctions, the layering of chronology into a set of terms for instance, or being unable to distinguish between English and Norwegian for a language element when one is written in subtitles and the other is spoken, strike me as especially problematic for a system purporting to fully capture the description of various objects.
In broad strokes, DC metadata seems useful, but also limited (and limiting) in very real ways. Abusing a system in order to bring it in line with perceived reality indicates the inadequacy in that system. Being unable to say whether the date element refers to the date of original production of what the item being described represents or to the production of that reproduction is a glaring example of this. Of course, one could use two date elements, but that would likely invite confusion quickly. That said, metadata seems to be a lesser evil than what we could have, organizationally speaking. In other words, it somewhat structures, however imperfectly, an amount of data that would otherwise be simply impossible to navigate. In practice if not in theory, the assignation of metadata elements seems fraught with error, uncertainty, and intentional deformations of practice.
Author: Daniel Powell
Word Count: 875
|Previous page on path
|2. Metadata, page 2 of 8
|Next page on path