Hugh Cayless
During one of the early planning meetings for the Digital Latin Library project (http://digitallatin.org/), in an attempt to enumerate all of the features one might wish to encode in a digital edition, we projected a page from the Harvard Servius onto a whiteboard and used markers to annotate all of the places where there was useful information implicit in the formatting and typography. We chose that text because of its complex use of print conventions to model the information in an edition.
Broadly speaking, editors have followed two approaches in presenting the text of Servius: Thilo, and more recently Ramires, have opted for a unified text, using typographical conventions like italics, bold type, and spread text to distinguish Servius (S) from Servius Auctus (DS). The Harvard Servius treats the differences between S and DS as more significant, and distinguishes them by their placement on the page. Murgia’s edition of v.5, which Bob Kaster is completing for the DLL, retains the earlier volumes’ formatting, but works harder to erase their ambiguities.
The approach that the DLL has taken goes beyond a simple transference of print conventions to the screen and uses semantic markup as defined by the Text Encoding Initiative instead (https://digitallatin.github.io/guidelines/), meaning that we rely upon a data model for our texts. Readings in the critical apparatus, for example, are represented as alternate branches in the flow of the text, so that readers of the online versions may choose to follow different paths than the editor as they read. Murgia’s and Kaster’s text, with its parallel streams that merge and then diverge again, presents a number of interesting problems. Servius makes it quite clear that digital editions require the flexibility to customize their models and presentation to accommodate the needs of the work at hand: a “one-size-fits-all” approach cannot succeed.