My goal is to produce a digital research and teaching platform that makes accessible online the voluminous prepublication materials for ‘Circe’, the fifteenth episode of Joyce’s Ulysses (1922). A dense inter- and intratextual descent into Dublin’s red-light district, ‘Circe’ takes the form of a fantastic play, with elements from earlier in the novel returning in hallucinatory form to beset the two main characters.

The digital research and teaching platform will allow students and scholars to explore how Joyce wrote the episode and how he achieved the effects discernible in the published text. I propose to bring together the encoding model for Genetic Editions (produced by the TEI MS SIG) with more ‘traditional’ TEI P5 markup to facilitate research queries and analyses along the diachronic axis.

Cordell will draw on the Viral Texts project at Northeastern University to demonstrate how computational
methods such as text mining, mapping, and network analysis can illuminate nineteenth-century systems of
circulation, reprinting, and remediation systemically and at scale. Cordell will demonstrate how this large-
scale analysis—and the new corpus of popular texts it has unearthed—might shift our understanding of
everyday reading, editing, and writing practices during the period, leading to a alternative notion of
communal rather than individual authorship. Cordell will also show how these texts might, when
considered at scale, allow us to expose lines of influence among nineteenth-century publications, leading
to a more robust conception of the systems that underlay print culture in the period.

Ø. Eide: Sequence, Tree and Graph at the Tip of Your Java Classes
How to represent texts in computer systems has always been an important topic in Digital Humanities. Tree based formalisms including XML are useful for many purposes, but problems related to their hierarchical structures are inherent. Various solutions has been presented over the years, from questioning the existence of overlap in textual material through various workaround for overlapping structures to the abolition of nesting formalisms. In this paper I will focus on three different design principles for text representation systems, namely, linear, hierarchical and graph based. This will be connected to differences in the understanding of data between tool developers and tool users.

A. Freitas: Effective semantic information processing over heterogeneous data – how we can help to better understand and process data for the humanities

G. Howanitz: Distant Watching: A Quantitative Approach towards YouTube Videos

Quantitative methods often focus on text(s), but nowadays even literary scholars have to take other media and cross-media processes into consideration. This "medial turn" becomes especially evident in the memetic reality of the Internet, where images and videos are shared, copied, transformed and recycled in large numbers. The sheer amount of material available online necessitates a quantitative approach.

In my talk I propose a method to "distant read" (Moretti) YouTube videos. 'Distant watching' computes frame likeness charts for each video, which can be then transformed and visualized using PCA or SOM. In order to illustrate the potential use cases of this technique I present two case studies: online videos about Ukrainian national(ist) hero Stepan Bandera which are used to negotiate cultural memory in Poland, Ukraine, and Russia; and "Jožin z bažin", a Czechoslovak video clip from 1978, which went viral in 2008.