The Modular Edition

Digital scholarly edition projects increasingly organise their work around reusable components (also referred to as “modules”, “lego blocks”, “templates”, or even “pizza toppings”). Rather than building a bespoke publication infrastructure for each edition, DSE projects – including but not limited to Ediarum, TEI Publisher, and eDITem – define digital scholarly editions in terms of separate components that can be re-used and recombined according to the project’s needs. These components, in turn, inform a more generic, modular publication infrastructure. The modularisation approach facilitates long-term maintenance of the DSE frameworks, while still supporting a certain level of customization flexibility. Moreover, modularisation raises questions like “what counts as a component? What parts of an edition can be standardized, and what must remain project-specific?” Accordingly, it is both a practical (technical) exercise and a theoretical (scholarly) exercise. We proposed to compare how component-based edition projects define their components, and to map those definitions onto one another. However, the frameworks we looked at define the interface between their components at different levels, so there was nothing comparable to line up. We therefore worked from the other end ("bottom up"), comparing five encoded letters from four different editions. Our conclusion is that body-level interchange is more tractable than the field assumes, and that the obstacle is not where editorial theory predicts. We end by suggesting a "Body Interchange Format" (BIF): the counterpart to CMIF, without the metadata. Like CMIF it would fix a vocabulary and require resolvable identifiers, but it would name editorial acts rather than elements, and oblige each project to declare how its encoding realises them. No such format exists as of yet and we propose to change that – something that would only work through collaboration across projects.