New Work Items

[pvretano]

In today's SWG meeting we starting discussing new work items for the SWG ... besides maintaining and enhancing Part 1. The list of possible work items discussed included:

• Faceted Search
• Transactions/Harvesting
• OpenSearch

The purpose of this issue is to get people thinking about new work items for the SWG. Please feel free to update this issue and add your thoughts and/or proposals for new work items. If you happen to have a fully formed proposal in your back pocket, please feel free to add it to the proposals folder.

In the mean time, here is a summary of what was discussed in the SWG today (30-SEP-2024) ...

Faceted Search:
The faceted search items is quite far ahead. There is an existing PR #372 to add the extension to the repo so it seems that is something that the SWG will definitely want to persue.

Transactions:
Transactions would rely on OGC API - Features - Part 4 and perhaps OGC API - Features - Part 5 and most of the work would be to decide which payload(s) (or more specifically which payload type(s)) should be posted to the server to create records in the catalog. The natural payload type would be a Record document as defined in Part 1 but other payload types might include a STAC item, a GeoDCAT document, etc.

Harvesting:
Harvesting is a bit more complicated because there are a couple of interpretations of harvesting.
One interpretation is that you point the harvest operation to some resource that you want to make discoverable and the catalog goes out, inspects the resource and creates one or more records in the catalog to make that resource discoverable. So, for example, you could point to a directory of EO imagery products. The catalog would go out, inspect each EO product in the directory, read the metadata about each EO product and then create one or more records in the catalog to make each EO product discoverable.
Another interpretation of harvesting is that the catalog harvests records from other catalogs for the purpose of federated searching. This seems like a more specialized version of the previous interpretation where the "resource" being harvested is a catalog.
Anyway, for this work item we would need to figure out the exact use cases we want to cover and go from there. There is an existing issue, #48, where there is some discussion about harvesting.

OpenSearch:
OpenSearch was originally a conformance class in Part 1 but we took it out and drafted a new part because it was felt that this capability was not "core". Quite a bit of work still needs to be done to get this new part in shape but the question is, is it worth the time and effort? My personal opinion is no. OpenSearch was a thing between 2005 and 2021 at which point Amazon repurposed the name "OpenSearch" for something else unrelated to the original OpenSearch. See here for further details. So, while OpenSearch was something that was supported by CSW the question of whether we want to move it forward still remains. Thoughts?