On Thursday 17 March, we held the second Enhancing DMVI workshop. The theme of the day was social networking, and the broad aim was to investigate the ways in which online communities (after the fashion of sites such as Facebook, Twitter or MySpace) might be encouraged to participate in the processes of tagging, analysing and commenting on digital representations of Victorian art. The participants (to all of whom the DMVI team extend many thanks) were an eclectic cross-section, consisting of students and staff from a number of Cardiff University departments, including English Literature, Language and Communication, Computer Science and Libraries/Archives.
The first session was an exercise designed to highlight some of the practical issues involved in tagging images. Participants were split into groups, and asked to provide keywords for a set of pictures, which had been taken from various websites which employ iconographic descriptors (‘tags’). The groups’ tags were then compared with the pre-existing sets of words – with some interesting results. In this part of the workshop, we also wanted to explore the possibility of employing established high-level categories, based on those used for DMVI, and compare the results when users’ tagged inside and outside an external framework.
The second session introduced the group to DMVI’s prototype Facebook App and online-tagging pages, and involved analysis of the some of the practical problems relating to user-generated iconographic description. As well as getting feedback on the webpage structure and point of entry, we also wanted to look at what might motivate people to come to and use such a site, and to think about the research aims that might be served by creating and maintaining an online community of taggers (not only for humanities scholars, but also for the computer scientists who would necessarily design and deliver any such system, and who have their own research interests to consider). A number of existing social-networking sites, and their methods of attracting and retaining users were discussed.
As usual with this kind of workshop, the day presented as many questions as answers, and we ended up a long way from our original assumptions. To some extent, the results were not entirely surprising: broadly speaking, tags relating to physical objects, settings, and people were fairly consistent across the groups; on the other hand, tags relating to themes, emotions, and the tone or atmosphere of a picture were less consistent. In essence, some people seem more willing than others to construct a reading around an image, and have much more confidence in bringing their own knowledge to bear in the tagging process.
What this suggests in practical terms is that tagging images is unlikely to have enough appeal on its own. People are drawn by the notion of usefully contributing to a community, and the process of analysis, criticism and feedback is a key part of that. Inevitably, this extends the scope of the project beyond simply generating keywords to help people search for images, into much wider questions of interpretation and narrative. Rather than merely adding terms to a list, it seems much more attractive to give people the option to develop readings of images, and correspondingly to comment on the readings of others.
In the end, the discussion coalesced around certain key questions:
- What is the primary objective of involving a user community? Is the focus on tagging and searching for images, or on creating and exploring a narrative around an image?
- What useful research questions do we hope to ask of the material generated by a community of taggers/commentators, and how might the material be usefully interrogated?
- What elements of user profiles need to be recorded, and what do we hope to learn from this data?
- Is the overall intention of this Enhancing DMVI project strand to generate useful data i) for humanities specialists and ii) for computer scientists, or is it rather to develop a model for the creation of an online community associated with future digital humanities projects?
These questions are not easy ones, and finding the answers may take longer than the duration of this project. With thanks to our friendly guinea pigs, however, we have made a start.