The Edition Indexes Itself

TL;DR The edition now has an index of people and an index of places. Neither is compiled by hand. Both rebuild themselves every time I add a letter.

Every person and place named in these letters already carried a pointer into an authority file, where I record the authoritative name, the dates, the occupation, and the identifiers that tie the entity to VIAF or to the Getty thesaurus. The site printed all of it as plain text. The scholarship sat in the files and appeared nowhere on the page.

Two pages now expose it. People and Places list every entity the letters name, in index form, each entry gathering every letter that mentions it. Names in the transcription link to their entry, and the entry links back to the letters. The place index opens with a map of the Europe of 1570 rather than of today: Franche-Comté is Habsburg, Ferrara still belongs to the Este, Corsica to Genoa, and the Empire is divided into its Imperial Circles. Places are plotted and sized by how often the letters name them, and clicking one goes to its entry. The calendar heatmap on the front page behaves the same way. Click a day and you reach the letter written on it.

The part that matters is what happens next. When I encode a new letter and push it, the indexes rebuild themselves in the same automated step that publishes the letter. A script reads the authority files and every transcription, and writes the data the site draws on. Nobody adds a name to a list. Nobody remembers to update a page. The index cannot fall behind the edition, because it is not a separate artefact that has to be kept in step: it is the encoding, read back out. As the edition grows letter by letter, so does everything built on top of it, without any further work from me.

There is no database and no server behind this. The indexes are static files, generated at build time and served as plain HTML. Minimal computing applies to indexes as much as to anything else.

The useful consequence of generating rather than curating is that the index reports its own gaps. It shows the authority records that no published letter yet cites, and the references I have not yet identified. A hand made index conceals that sort of thing. This one publishes it, which is the more honest arrangement and, incidentally, my working list.