search

opengeospatial / HY_Features

http://opengeospatial.github.io/HY_Features/
Other
8 stars 1 forks source link

HY_IndirectPosition model changes. #196

Closed dblodgett-usgs closed 7 years ago

dblodgett-usgs commented 7 years ago

Remove the HY_RiverReferenceSystem class. An HY_IndirectPosition now has relationships to HY_FlowPath linear element and HY_HydroLocation referent. This allows assignment of a position (referenced along a flowpath) to a HY_HydroLocation and to express this as distance from the known referent (HY_HydroLocation).

sgrellet commented 7 years ago

+1

one comment though (I hope I don't open a dangerous box): in many hydro dataset I know we have those HydroLocation layers (often point based). This is ok for referencing but often not ok from a data management point of view as they duplicate information from other datasets (I'm thinking about dam, hydrometric station, Wells etc...) then lead to inconsistencies within instances... . => have we ever discussed having the 'referenceLocation' pointing to anyFeature ? or an attribute allowing this within HY_HydroLocation ?

'HY_DistanceFromReferent' : we may want to mention which version of the FlowPath element is used. We have attributes regarding accuracy and precision, that's good. But when reference 'blue lines networks' are regularly updated/consolidated (ex:each year), if we don't know which version of a given FlowPath was used, you have no way to reproduce the exact positioning (ex : if a meander was added in one FlowPath between two versions). We could say that's handled somehow on the way you assign identifier to FlowPath instances but I feel this should appear more visibly (extra attribute ?).

dblodgett-usgs commented 7 years ago

This is a good point. The HY_HydroLocation should be implemented in such a way that it points to another feature, for sure. While the HY_HydroLocation is typically a point, it almost always is a reference point that is taken from some other source. Hopefully, data systems have the ability to link to the actual location of the hydrolocation rather than copying it's geometry into an instance in the context of the hydrographic data. Will look for a chance to make this point.

Your second point is well taken, and we've intended to make it by saying that it must be one and only one flowPath. Will make sure it's made in a more firm way in the spec.

dblodgett-usgs commented 7 years ago

Removed river reference system from the spec.

dblodgett-usgs commented 7 years ago

Section 6.6.1 and 6.6.2 still needs some rewriting.

dblodgett-usgs commented 7 years ago

@IrinaDornblut took the first stab at the rewrite. I'll finish it up. Latest in #214

dblodgett-usgs commented 7 years ago

will finish this up pending #222

dblodgett-usgs commented 7 years ago

I think section 6.6 is in pretty good shape. It could use another set of eyes, or another pass from me later. Will leave this open but put it in the review column.