A sustainable work flow for a multi-layer edition of the Chronicon by Romualdus Salernitanus

My experimental scholarly digital edition of the De nomine by Ursus Beneventanus tested the feasibility of the edition model theorized by Orlandi 2010, based on three different layers of text representation (graphematic, alphabetic and linguistic) and on a table of all graphemes having distinctive value in the graphic system of a specific manuscript (“graphematic table of signs”). Its work flow, however, proved to be very time-consuming. This talk analyzes the specific practices of the Ursus edition that mostly slowed down the work flow and outlines possible solutions to be applied in my ongoing digital edition of the Chronicon by Romualdus Salernitanus. Those aspects include (a) the markup of abbreviations, which can be expedited by taking advantage of the systematic nature of ancient abbreviations, (b) the markup of ancient punctuation, which will be omitted altogether, and (c) the representation of the linguistic layer. The minimal units of the latter are inflected words regardless of any specific spelling. In the Ursus edition, each word was represented at this layer by a combination of lemma and morphological information (e.g.: ablative plural of lemma “praepositio, -onis”), by means of @lemma and @ana attributes of . The values of those attributes were populated by the lemmatizer/PoS tagger TreeTagger, but needed to be reviewed manually. The Romualdus edition will initially include no representation of the Linguistic Layer. If time will suffice, a simplified represent ation of it will be added, falling back on the common practice of representing words at the linguistic layer by means of their “normalized” spelling. Lastly, while in the Ursus project the TEI-to-HTML transformation was performed dynamically by JavaScript, in the Romualdus edition a Python script will perform this task statically.