Marking up transcriptions

Fremantle

· Wikisource · Wikimedia · FreoWiki ·

I've been working more on the City of Fremantle and Town of East Fremantle Street Names Index, which is a process of (after getting a copy of the original spreadsheet that the PDF was created from) copying the data into a wiki page and then going through it line by line and adding links to the streets on FreoWiki, the Wikidata item for the street, and the FreoWiki page for the person/thing that the street is named after. It's not really about retaining the original formatting (some of the descriptions are accidentally spread over multiple table cells for instance), and it's a born-digital document so it's not really about transcription — but it is a process of annotating, with links, and so it's got me wondering. There's so much tabular data in the world that would benefit from being moved into a more structured form, but in doing so it's vital to also retain the original layout and verbatim text. Ideally the transcription process makes it easy to not have to handle each row twice (once to transcribe an annotate it, a second time to work on the annotated data) but this doesn't really feel very easy with the Wikisource-esque workflow.

It's a matter of going through a document, and adding links like I am above, and then on Wikidata adding statements based on (and citing) the document. Could a tool such as the new annotation gadget do this in one easy step? Select text while editing, get a popup with a field for searching for a Wikidata item and once one is found another field for adding a property (with the current page as a reference URL). Then it'd insert the link with a template such as wdl. It doesn't really feel like something that'd be easily generalizable, but maybe I'm wrong.