FeatureCoverage

A feature in Ilwis-Objects is anything that can have one or more geometries and has one distinct identity. In the Ilwis-Objects terminology an ice berg tracked over time (assuming it doesn’t split) can be represented as one feature. It may change shape and/or location in time but it is still the same ice berg. The track of geometries is referred to as the ‘index ‘, similar to the same notion in the raster data model.

A feature at a certain index is referred to as feature-element and is similar to a traditional elements of vector maps. Apart from geometries, features also have extensible set of other properties (the attributes) that depend on the nature of the feature and context in which it is used. Organization On a programmatic level sets of features are organized in a feature coverage. Feature coverages are collections of features that share one or more properties (e.g. within the same bounding box). Features are thus described by

• One or more spatial locations.
• Per index entry a geometry ( which maybe an empty geometry)
• Index dependent attributes
• Index independent attributes.

The consequence is that feature-coverages are described by

• A set of features (duh)
• A common spatial reference system else the spatial location would not make sense
• common index-domain or else the index values make no sense. Usually a value (related) domain like value, time or scale but others are also possible.
• A common attribute table.

Indexes are mapped onto a domain. A few notes here:

1. The traditional feature layer representation is of course still possible. It is just a feature set with only one index and one geometry type.
2. A feature set is not necessarily mono-type, not even when they are restricted to one index. Features happily coexist with different geometries. This doesn’t prevent mono-type sets though.
3. As with rasters, movement through the feature set is controlled by an iterator. This iterator can (optionally) discriminate between the various geometry types(and other properties).

Every feature is tied to a record in an attribute table through its featureid. A featureid is a run-time unique id that is generated for every feature during the run time existence of this instance of Ilwis-Objects. No meaning or persistence is attached to the featureid. It is only a glue between feature-set and attribute table. The index may have an attribute table were attributes are stored that are solely dependent on the featureid + index key combination.