October 2023: What future work do you recommend for testbeds, pilots, code sprints and other OGC initiatives?

[ghobona]

Based on the outcomes of the October 2023 Code Sprint, what future work do you recommend for testbeds, pilots, code sprints and other OGC initiatives?

What work (e.g. new standards/profiles/best practices) would you like to recommend to the Standards Working Groups?

[tomkralidis]

OGC API - Records

• Extension for faceting
  OGC API - Pub/Sub
• use case experimentation for various OGC API standards workflows
  • Records: aggregation/metadata ingest/publication
  • Processes: subscribing to async jobs channels
  • Collections: publishing data updates (e.g. publication of new NWP model run for EDR or Maps (cc @solson-nws @chris-little)
  • associating OGC API collections from AsyncAPI (x-ogc-api-collectionId)

[chris-little]

@ghobona I agree with @tomkralidis . In particular, I think that the use of PubSub with Collections or a Collection is very important.

[IvanSanchez]

I personally would like a it more of SensorThings, and OGCAPI Tiles.

Also, multiband rasters in OGCAPI maps/tiles/coverages (by "multiband" I mean non-RGB and non-RGBA; either single-band elevation, or two-band slope, or 12-band Sentinel multispectral, etc etc). Including the questions: does it make sense to have it maps and tiles? How to symbolize multiband data?

It'd be very interesting to see a "GIMI vs GeoTIFF" shootout. I recall some GeoTIFF/COG people in a previous codesprint, so let's put them together in the same room with the NGA people in charge of GIMI.

[bradh]

Also, multiband rasters in OGCAPI maps/tiles/coverages (by "multiband" I mean non-RGB and non-RGBA; either single-band elevation, or two-band slope, or 12-band Sentinel multispectral, etc etc). Including the questions: does it make sense to have it maps and tiles? How to symbolize multiband data?

Extension to this - tiled complex, signed, and floating point data. The viewpoint is where the tiles are a way for machines to process an area of interest (potentially in parallel), rather than as pictures to show to the user.

[chris-little]

@ghobona Perhaps a standard workflow is something like:
Observation->Collection->Exposing availability->Selection of data->Process->Generate visualisation (which may be maps or not, such as graphs)->Dissemination of results->Archiving.

Though Archiving may be applicable at any stage, and also PubSub, and Visualisation. And a lot is hidden in Process.

I suspect Q/C is applicable at every stage too, though the forms may be widely different.

[ghobona]

Thanks everyone!

[GeoCarsten]

Hi, The MUDDI SWG would like to continue with data encoding related work, though we are currently debating whether this is better done as part of a larger codesprint or standalone. Both approaches hae different pros and cons.