An experiment — and some of these places are wrong.
Nobody checked this map before you saw it. That is the point of the demonstration, and it is
also the problem with it. Two examples found by clicking on it: “the Court of
Admiralty”, mapped to a quarry in Dorset; and [CSP] — the catalogue’s
shorthand for the Calendar of State Papers, cited whenever it vouches for a man’s court
office — mapped to the Chartered Society of Physiotherapy in London, where it briefly
became the sixth-busiest place in Tudor litigation. Both are fixed. Both were invisible in the totals
and obvious on the page.
Errors of that kind are not a defect in the idea; they are the reason the
Workbench is being built. A machine can find candidate places and narrow them
down. Choosing between the candidates is a job for someone who knows the county — and
giving that person good tools is the actual problem worth solving.
Named but not located in
Names the extractor found that WHG’s index could not match inside this
county. Matching is strict — a name must match a record exactly — so this list
holds three kinds of thing: real places the gazetteer does not yet know, names the sources
spell differently from the index, and people, which is what searching inside a county is for.
Loosening the match to catch the first two reliably placed the third, so it was not kept.
How the Workbench helps. It stops for a person wherever you want
it to: its review stage puts the rival candidates for an ambiguous name side by side, takes a
rejection, lets you search by hand for one the index missed, and records why you decided as you
did — and a decision survives a re-run of the extraction. Every place on this map skipped
that step. Anyone publishing from it would not, and the not-located list above is where they
would start.
Built with the World Historical Gazetteer Workbench.
How this was made ·
WHG
Catalogue data © The National Archives, under the
OGL v3.0.