It Is Happening Again: Do we keep building the same Digital Humanities web interfaces?

Sheffield’s Digital Humanities Institute now maintains about 85 project websites, some of them dating back to the early 2000s. This is a huge and growing burden in terms of developer time. In their current state they will never be cost-effectively maintainable.

The project data is sound. DH was quick to recognise the value of open data standards. But the project interfaces are built upon a jumble of forgotten and ancient technologies; many of them once flavour of the month, now obsolete and unmaintainable. In the commercial sector, websites are overhauled every few years as a byproduct of continual funded development. The academic research grant does not offer that luxury.

And in broad terms, most of these interfaces do many of the same things: Browse records in some sort of hierarchical list. Fill in a search form. View a results list. Should it not be possible to standardise the technical implementation of some of this behaviour?

Seasoned Digital Humanists will immediately be sceptical. The field has long attempted to formalise its techniques, with mixed success.

The Digital Humanities Institute Data Service was conceived as a pragmatic way to address the department's maintenance burden; a way to keep project data available even if a project website could no longer be maintained. But while developing it, it became clear that obvious parallels in interface function could be drawn across its hosted projects. To what extent do we dare imagine a standard DH interface? Does such an undertaking undermine the originality and research value of our work? Would it be inevitably destined to join the discarded technology scrap heap alongside the interfaces it was meant to replace?

This paper examines how we might go about defining and standardising the broadest, most common functions of Digital Humanities web interfaces. Doing so has both technical and project management implications. It is relevant to any developer or PI who is beginning to feel that they may have implemented essentially the same DH web interface more than once.