The OCDH has been involved with the International Image Interoperability Framework™ (IIIF) since its beginning. We are adept with Annotation to supply supplementary data to data resources that are owned by others. We have our own public open annotation store and Web API known as RERUM to support this ecosystem of assertions. The scope of IIIF encompasses media like Image, 3D, and A/V files within the Internet of Things. Our challenge of People + Places + Things + Events extends beyond this scope but is still driven by the standards and best practices employed by such a framework. We focus on Web Annotation which IIIF uses to supplement the data resources under its scope. We realized using IIIF resources as subjects for geographic Web Annotations could set some foundation for standardized, interoperable, usable, and geospatially described linked open data in the broader Internet of Things.

At Stanford University in a room full of people who specialized in geospatial resources I learned of GeoJSON. This is a base data format utilized by a large set of map-related user interfaces. It is a structured data object containing coordinate type and latitude/longitude information. I also learned that IIIF has documented annex support for GeoJSON. I was left with a purpose: make the assertions with GeoJSON so they work immediately in UIs that already exist and fit within a standard data format that data scientists are using. That way, near-term integration with IIIF resources would be possible. This was the first major step and one from which we did not look back.

I also learned about the kinds of challenges that await. Geographic assertions can have differentiating purposes which means GeoJSON data nodes ought to be able to note a purpose. We found a repetition of sought-after ends: geolocate, geocode, georeference, georectify, and co-locate. I was also enlightened by the fact that humans have created varying geographic coordinate systems that use varying spatial reference systems. Adoption of certain specifics was not uniform. This means when I provide you with a latitude and longitude from one given system, that may relate to a different location in the world on another. Sometimes the difference is only a couple meters, sometimes much more. This means in some way the data must note which system it uses, or the coordinates end up more arbitrary than intended.

The conference focused on the challenge of making these assertions on IIIF resources that are images of maps. There are margins, scale, languages, coordinate systems used, and other variations that make images of maps tricky. Harder still, inset maps are common. That is a map on a map, in the corner or something, to give some context to further the purpose of the map. It is also important that conversion between pixels and coordinate values is standardized and usable across user interfaces. Clearly, encoding coordinates in a standard format was not the hard part.