With two keynote speakers, six parallel sessions in three rooms, a poster and a demo session DHBenelux2015 had a full program, spread over two days. I was present there with my BiographyNet colleague Antske Fokkens and our two Academy Assistents, Yassine Karimi and Miel Groten, who presented their research on changing perspectives on people in biographical dictionaries over time.
 |
| DHBenelux group photo, after dinner at the Antwerp Zoo. Photo Saskia Scheltjens |
The program provided presentations of a nice mix of different kinds of digital humanists. I think we can divide them into a few categories, of course with the necessary overlap:
1) Humanist researchers that have little background in DH but try to apply ready made tools on their own datasets.
2) Humanist researchers who are experienced in using DH tools for their research but do not build tools themselves.
3) Humanist researchers that are part of an interdisciplinary team that builds DH tools (a category I belong to myself).
4) Computer scientists (a minority) that build DH tools in collaboration with humanist researchers (from either other category)
Who we may have missed at DHBenelux, but I believe this is a general problem, are researchers that have an extensive knowledge of both the humanities and the digital technologies. I was present there with a computational linguist and a (future) computer scientist and therefore would not have to be afraid of any difficult technical questions. Other teams however, were often only represented by the humanists, or only relied on ready made tools. This lead on several occasions to an apology for not being able to answer a question 'because our computer scientist isn't here.' Maybe even more worying is that tools are used without really knowing what they exactly do (black box tooling). You can compare the outcome of different tools and evaluate what seems to work best, but it is also necessary to know why some tools work better than others, to explain the differences and, if necessary, to be able to work on adjusting a tool for your own purpose. DH tools cannot be treated like text editing programs that may or may not function properly. DH tools read, interpret and manipulate the humanist data that are the core of our research. A proper tool criticism is necessary when using them.
The question remains of course, and has been asked many times before, how digital humanists, and current humanities students, can be trained in the best possible way. What is practical? What is needed? Where do we find the time and the resources? But maybe it is just a matter of waiting for the next generation to take over. At DHBenelux there were quite some very young junior researchers who already are becoming experts in both the humanities and in computer science (or only programming) and will be more comfortable in uniting the best of both worlds than most more experienced academics are now.
