Closed sgillies closed 10 years ago
+1 nice Personally haven't specified collections of collections in GeoJSON but would be relevant in more cases in order to group by aggregated "when" for GeoJSON-LD.
This sounds good to me. It's basically what I'm looking for! :)
Great. Here's an example of a feature with temporal bounding box.
{ "type": "Feature",
"tbounds": {"earliest": "-1000", "latest": "-900"},
"when": {"@type": "Instant", "datetime": "-950" },
... }
If you were to put this feature on a timeline by itself, it would suggest the bounds for that timeline view. It could also be used in the sense of Simile's fuzzy periods.
Now we need a good name for this object. To be continued on Tuesday, because I'm going to be gardening the rest of the weekend :)
Let's call this "circa".
I don't see a reason why circa shouldn't be just another Interval. I've made it so in the example, see Fuzzy instant.
Circa? The definitions I have seen for circa are consistent: an approximate date. Using it to define an interval is going to cause confusion.
How about during
?
In #18, @ekansa suggests a temporal bounding box, which also came up in the discussion of #9. See @kgjenkins at https://github.com/geojson/geojson-ld/issues/9#issuecomment-41921750.
@kgeographer points out that they're not as necessary for features as spatial bounding boxes are because there are no temporal objects comparable to polygons and multipolygons. They do seem quite useful for feature collections, however, giving a quick summary of the temporal range of all its features.
Would something like this work?
And would feature tbounds let us express that a feature might be relevant to times outside the range specified by its "where" item? The return of fuzziness, in other words. By the way, I've done this in GeoJSON before: setting a feature's bbox to be wider than its actual geometry so that my map applications will zoom not to the geometry but to the wider bbox.
I'm not wedded to the name "tbounds". Suggestions are very welcome.