2018 ESIP Summer Meeting

A number of JSON serializations exist for representing Earth Observation data however further work needs to be undertaken to align efforts, reduce overlap and expand/evangelize usage and software implementations. Some examples include TopoJSON, JSON-LD, hdf5-json, GeoJSON, CovJSON, CF-JSON, NCO-JSON, STAR JSON and there are several others.
Due to initiatives such as the W3C + OGC Spatial Data on the Web Working Group (which resulted in a significant advancement of the CovJSON standard) this issue is gathering so much interest that NASA has recently initiated a dedicated Earth Science Data Systems Working Group to investigate, evaluate and provide a formal NASA recommendation for use of JSON Encodings for Spatial Data.
HDF5/JSON preserves the data and metadata of any HDF5 dataset through a round-trip encoding (i.e., HDF5 -> JSON -> HDF5). Thus HDF5/JSON is automatically 100% lossless. NCO-JSON is a compa-rable turnkey solution that serves for netCDF a similar role as HDF5/JSON for HDF5. Although netCDF can be implemented as a subset of HDF5, the two APIs and their vocabularies are so differ-ent that using or extending HDF5/JSON to represent netCDF files would be unnecessarily complex. CF-JSON, ERDDAP, and STAR JSON all implement the NCO-JSON dialect, designed to represent any data stored in netCDF format. Differences include that CF-JSON is designed to extend to higher-level CF constructs while STAR JSON includes a library for potentially faster conversion to and from JSON. NCO-JSON also provides lossy options to reduce JSON verbosity and size and increase legibil-ity.
Critical production-grade software infrastructure such as OPeNDAP also provides several mecha-nisms for serializing JSON data retrievals. Very recently, OPeNDAP developers have provided the ability to retrieve CovJSON responses from OPeNDAP queries. We anticipate that this functionality will enable a new generation of high performance applications.

Some stakeholders and interested parties in this area include:
· Web application developers tasked with designing and developing applications which con-sume EO spatial data.
· Parties interested in serving and consuming Spatial data on the Web.
· Developers at data centers who currently distribute/expose endpoints/resources which serve spatial data
Two back-to-back sessions will therefore cover (i) data modeling issues; providing an opportunity to evaluate the list of candidate JSON encodings and exploring model semantics, (ii) applications which leverage JSON serialization for EO data, and (iii) use cases which could explore and possibly benefit from use of JSON serializations for improving application performance.
The first session will offer four 20 mins presentations with the second workshop offering a hands on investigation into then extension of existing systems’ ability to return richer JSON encodings.