You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If we reduce e.g. a 2D coverage to a single point value (e.g. requesting an elevation value for a specific position), can this be encoded in CIS JSON? Does it need to be encoded as a 1D coverage?
From the ISO 19123 / AT 6 standpoint is there such a thing as a 0D coverage with a single set of range values?
extracting a value in the first place would be a value, so the server would return this value (probably in ASCII).
However, for reasons of closure a value might just as well be considered a 0D coverage - so with no axes, an empty axis list, no coordinates, etc. Could certainly be encoded in JSON, for example. (Note that not all formats need to support all result dimensions - eg, TIFF cannot return 5D - so this is nota "bug".)
Bottom line: behavior in this corner case is implementation dependent, as different applications may expect different things.
Coming from the Observations&Measurements perspective, my first reaction would be say that this is an observation ;)
Staying within the coverage philosophy where the core data model consists of a Domain and a Range (with some details in between providing context), I'd say the result needs to be a bit more than just an ASCII number (the Range), as without something like the Domain providing the where (in space, time or other conceptual axis), it's really just a number.
I guess one could reduce the Domain to just the origin, then the 0D logic pans out
if you did trimming then indeed the result is a coverage, with all its axes, that contains a (single) location, or to be exact: lower bound = upper bound.
if you arrived via slicing then there is no spoon - err - coordinate any longer as all axes got removed. Meaning, if you keep a coverage view: zero axes in the domain set.
coming back to the original question: a coverage that contains only 1 direct position will contain 1 range value, regardless of any regular/irregular/etc bells and whistles, and for any number of coverage dimensions.
So IMO can be closed.
Activity
pebau commentedon Jun 5, 2022
extracting a value in the first place would be a value, so the server would return this value (probably in ASCII).
However, for reasons of closure a value might just as well be considered a 0D coverage - so with no axes, an empty axis list, no coordinates, etc. Could certainly be encoded in JSON, for example. (Note that not all formats need to support all result dimensions - eg, TIFF cannot return 5D - so this is nota "bug".)
Bottom line: behavior in this corner case is implementation dependent, as different applications may expect different things.
KathiSchleidt commentedon Jun 5, 2022
Coming from the Observations&Measurements perspective, my first reaction would be say that this is an observation ;)
Staying within the coverage philosophy where the core data model consists of a Domain and a Range (with some details in between providing context), I'd say the result needs to be a bit more than just an ASCII number (the Range), as without something like the Domain providing the where (in space, time or other conceptual axis), it's really just a number.
I guess one could reduce the Domain to just the origin, then the 0D logic pans out
pebau commentedon Jun 5, 2022
hm, depends:
pebau commentedon Mar 6, 2026
coming back to the original question: a coverage that contains only 1 direct position will contain 1 range value, regardless of any regular/irregular/etc bells and whistles, and for any number of coverage dimensions.
So IMO can be closed.