News travels fast in the big, happy, dysfunctional family of DH/TEI practitioners. Everybody knows by now that Edward Vanhoutte served hand-made (sweet, sour, bitter, salty and umami) chocolates during his keynote at the TEI Conference in Würzburg. With the exception of the bacon praline, which tasted just the way it sounds – like a bad hangover – Vanhoutte’s confectionary treats were both daring and delicious, offering an overlapping hierarchy of sensations in one’s mouth while leading, inevitably, to typical DH/TEI questions such as: what is the nature of chocolate as a medium? What are its constraints? Can we create an ISOCat category to resolve non-standardized chocolate vocabulary? And how would we translate the TEI representation of such-and-such taste into an RDF triplet?
But there was certainly more to Würzburg than chocolates (and Silvaner, and beef tatar with Sabine…). At least two topics — really, two specters — haunted & taunted me during the conference while complementing and antagonizing each other: the specter of social editions and the specter of asocial tools.
The need for and the difficulty of producing truly social, fluid texts was what propelled Vanhoutte’s keynote into metaphoric exhibitionism. But it also lead him to ask one of the central questions: how should we manage referentiality on a large scale in a world of shifting texts and collaborative annotations? The question is no longer only: what to do with a million books, but also: what to do with a million TEI documents? And how to do it?
These questions — in addition to being essential for my meddling with text-dictionary interfaces in digital libraries — really go beyond the scope of TEI itself. They are not just about agreeing on one tag or another, writing best-practice guidelines and developing application profiles. They are really about embedding our textual, editorial and development work in larger contexts of digital infrastructures and interoperable web services.
And it is there, in the realm of hoped-for and not-yet-fully-utilized infrastructures that the specter of digital editions steps on the toes of the specter of asocial tools.
Infrastructure projects should be as much about technology as they are about people who use the technology. Unless a great deal of thought (and let’s face it — investment) is given to education and outreach, digital infrastructures will always run the risk of fossilizing before even gaining momentum.
After Andrea Rapp presented Germany’s wonderful and ambitious TextGrid project — a virtual research environment that had its official release party at the opening of the Göttingen Center for Digital Humanities in July — a very important question was put forward: what could and should be done to make sure that TextGrid gets adopted and used to its full potential outside the inner circle of hard-core, code-till-you-drop TextGridistas?
Serious hands-on workshops are indeed needed to introduce a whole generation of users not only to what is already available in TextGrid, but also how to set up and develop their own projects in the TextGrid VRE. This is different from arguing that we need “tools that your grandmother can use” — although this thought certainly was entertained by several people in Würzburg. While I am all for empowered grandmothers (mine was a member of an avant-garde modern dance group in the 1920′s, which offended conservative Serbs by prancing around — lo and behold! — completely barefoot), I am at the moment more interested in making sure that complex tools become widely accepted by their target communities. If we are ever to come close to creating social editions, we will need, at the very least, perfectly socialized tools. Tools without people are like academics who never got over the fact that they were not popular in high school: bitter and lonely.
Of course, the world would be a candy land if TextGrid workshops were a sugar-coated panacea for all DH ills.
Before the chocolates, Vanhoutte told us about two students who worked on two different digital edition projects: one in HTML, the other in XML. The HTML student did no textual encoding herself, but wrote an academic essay about the nice-looking edition she produced. The XML student wrote no essay, but did a fair amount of very useful encoding.
The HTML student passed with honors, while the XML student failed.
The XML student failed because the scholarly value of encoding itself does not seem apparent to the academic world at large. So, there is, after all, a third specter hovering above the other two: the specter of academic evaluation of text encoding practices. As long as there are no clear guidelines in PhD programs around the world about what counts as scholarly work in digital humanities, text encoding will remain on the margins of academic acceptability.