Capture requirements for JSONPath

[mmccool]

IETF has a group attempting to standardize JSON path, we should capture our use case and requirements (eg security...) and provide our input.

Edit: see https://datatracker.ietf.org/group/jsonpath/about/

[farshidtz]

Is it the same jsonpath group mentioned in #56 (comment)?

I think operators for ISO 8601 timestamp comparison are necessary. String operators are not very handy and often require complex regular expressions (related to #80).

Geospatial queries are nice to have, in case they are part of a TD (related to #79).

OData examples from OGC SensorThings:

• Time comparison: resultTime ge 2010-06-01T00:00:00Z
• Geo distance: geo.distance(Locations/location, geography’POINT(-122, 43)’) gt 1

[mmccool]

Notes:

• Yes, as mentioned in issue 56 so I added a link to the IETF group to the description
• a joint meeting with SDW/OGC is planned for Dec 10, we can discuss what should be supported there; this is closely tied to how geospatial data is represented in TDs or directory metadata about TDs (and there is still the issue of "dynamic" geolocation data reported via a property...)

[mmccool]

Generally I have a concern that arbitrary JS should not be allowed in queries for security reasons (including DoS attacks by specifying pathologically expensive queries). Instead there should be a specific set of comparison operators and query types supported. I hope this aligns with what the IETF group is thinking.

[farshidtz]

From Discovery call:

• Closing this issue. Discussion can continue on JSONPath as a search query language #234