[live blog, so excuse the errors, omissions and personal perspective]
Humanities Research Institute at the University of Sheffield established in 1992: support innovative use of technology to support humanities research. Entirely project based and externally funded, with no teaching or faculty support.
Digital Humanities is an essentially practice based activity, not a theoretical one. Oxford’s practice based approach makes it ideally suited to hosting this workshop. Skills gained from practice required for DH.
How does the information age relate to the digital humanities? Modern information very similar to the record of our human past: messy. Humanists are good at analysis is this messy data, so are useful to our age. Three examples of why.
DH can bring manuscripts together to not only one place but to a place where cross-referencing and comparison can occur.
Assembly not basic. Have to think what is useful for the scholar, which techniques are appropriate, survey the state of the field and proximity to data (here Canterbury Tales digital editions project). Network analysis, used in humanities to understand relationship information not explicit in the evidence. Type of data work often referred to as web ontologies, though that has got some bad press in recent years. Preference for Pidd of calling databases, databases. Needs of these sorts of projects challenging: visualising Twitter feeds ‘child’s play by comparison’.
Locating London’s Past. Bringing datasets together, from old maps to street view! Complexity of doing this: collaboration with Museum of London, so map had to be warped to fit; which King’s Street is being talked about at any one time?; where on the Strand did the crime occur? To achieve a good data mashup, you need a limited number of data sets which you know well.
Humanities impact through creation of research standards which help organise modern data. Without standards we encounter problems of longevity. Hiding poor OCR under a resource causes more problems than benefits, eg John Johnson Collection or British Library newspapers: billions of words, around a third of which are wrong. Hartlib Papers made obsolete before it was even published because of updates to commercial software it relied on. Sustainability can be ensured through good data curation, use of open standards, open software. User testing and data curation for historical resources similar to those required for B&Q or John Lewis websites. Investment in data during the Old Bailey project ensured its longevity, both as a dataset and as a resource. But much more expensive.
Most important aspect of digital humanities. If management is bad, project fails. We learn from failures. Overambition often the key problem: early projects tried to do too much tagging, transcribing.
As a result, others sectors and the economy in general benefit from work done in the digital humanities.
Why so project/service based? Surely DH is about understanding historical phenomena using digital techniques, not just serving up data in usable formats?