Standards and Metadata, Part X
Welcome to the tenth (!) instalment of my series posted here, on my blog and my original website. In this year’s AGI GeoCom17 lightning talk session, I “challenged delegates to think about the developments of spatial grids and the structure of spatial data models”:
A romp through geospatial gridding the last 30 years pointed to OGC’s newest proposal for Discrete Global Grid System. Their announcement starts with: The goal of DGGS is to enable rapid assembly of spatial data without the difficulties of working with projected coordinate reference systems. The OGC DGGS Abstract Specification standard defines the conceptual model and a set of rules for building highly efficient architectures for spatial data storage, integration and analytics. It concludes that: One of the core contributions of a DGGS is geospatial data fusion on demand. In a multiple provider environment, fusion is only possible with an information system architecture based upon open standards. The OGC DGGS Abstract Specification provides a platform to enable interoperability within and between different DGGS implementations while promoting reusability, knowledge exchange, and choices in the design of individual DGGS implementations.
This is no small deal, as for example Australia’s CSIRO propose a datum update due to the dynamic nature of earth’s crust. And this applies to the northern hemisphere as well as described by Blue Marble Geographics.
Not to deny the excellent work by surveyors worldwide — a good starting point on this vast topic is EPSG — such a grid may help interoperability immensely by abstraction the spatial objects.
One practical example is what3words plan to address the world on a 3x3m.sq. grid in parts of the world where no street address exist so far. Not unlike their founder who in the music business early in his career had trouble locating people and shipped kit, my personal experience was a failed pizza delivery in Kuwait City, because my street address sans post code remained a mystery… until I met them at the mosque that luckily was kitty-corner to my flat!
Case in point, it wasn’t the GPS coordinate that mattered to my driver, as much as the location of my flat w.r.t. local landmark he could relate to.
Back to my AGI lightning talk, Spatial Object Models carry the notion of concept vs. occurrence in geologic data models— details of lithology can vary over time as classifications change, but the location of said samples will not — and I would add today: in distinguishing the metadata from geolocation, why not take it one step further an put the geolocation in the metadata?
Grids proposed by OGC may address that, and a lot more; to reiterate my challenge to AGI GeoCom17:
- Are coordinates, projection, datum etc. always à-propos?
- Are there better frameworks for real-time crowd-mapping?
- Do robotics ‘think’ or work in Cartesian space?
- Are base 2 or 10 the best there is? [recent news item]
- How about ‘good enough’ or ‘close enough’ computation?
This closing image — how do starling swarm like that? — could be the ultimate challenge in geo-robotics… to automate geo-location in real time!
