Transformation Type

[thehabes]

Our prototype is missing this information. What are some ways we can supply this information?

[thehabes]

@bert where might this go? If the type specifically applies to "pixel_coords", it may be worth making "pixel_coords" a JSON object like

pixel_coords:{
   "type": "TransformationType",
   "value":[12, 12]
}

[thehabes]

If the type is put on the GeoJSON body, we would have to careful because GeoJSON already uses type. type can be a @set. So..

{
  "type": ["Feature", "TransformationType"],
  "properties": {
    "pixel_coords": [
      2840,
      3153
    ]
  },
  "geometry": {
    "type": "Point",
    "coordinates": [
      -78.137998,
      39.217154
    ]
  }
}

or extend GeoJSON with something like transformation.

{
  "type": "Feature",
  "transformation":"Type",
  "properties": {
    "pixel_coords": [
      2840,
      3153
    ]
  },
  "geometry": {
    "type": "Point",
    "coordinates": [
      -78.137998,
      39.217154
    ]
  }
}

I am more in favor of keeping our stuff as scoped as possible, so I would prefer going with something like in the previous comment where this new field/type are defined alongside the "pixel_coords". The less extending we have to do, the better.

[bertspaan]

What about something like:

{
  "body": {
    "type": "FeatureCollection",
    "transformationType": "polynomial",
    "features": [
    ]
  }
}

[thehabes]

Yes, that would work out. transformationType would be considered a foreign member to the GeoJSON. We could add a context entry for transformationType and scope in some expected values. We would then add all of that to the vocabulary document.

Is it always true that all Features will use the same transformationType as noted by the FeatureCollection?
Or, another way to ask: Is there a situation where you would need different transformationTypes for Features in the same FeatureCollection? If so, we may need to accommodate that immediately and put transformationType on the Feature.

Here are the rules for foreign members : https://tools.ietf.org/html/rfc7946#section-6.1

I believe what you and I are suggesting follows all the rules.