sequentialArray interleavedArray runLengthEncoded nTree
IndexScheme is an abstract superclass within which the possible schemes for referencing gridded data are defined. The inheritance hierarchy for IndexScheme is shown below.
The values sequentialArray and interleavedArray imply that no form of compression is used on structured data. The data is organized as either sequential or interleaved. For example, a non-compressed raster cover may be described by an arraySequential. A satellite image that is delivered with seven interleaved bands would be described with the value arrayInterleaved. The value runLengthEncoded provides for data compression through the grouping of locations with identical values. When applied to StructuredData, the attribute contentValues contains an integer representing the number of identically valued locations followed by the value of the location as specified by the attribute structure. This structuring scheme begins with the first location and progresses through the last location.
The value nTree allows for the spatial referencing of locations using a linear tree of n dimensions with variable length locational codes (Samet, 1989). When applied to StructuredData, the attribute contentValues contains an integer representing the locational code (Morton number) of the identically valued locations followed by the value of the location as specified by the attribute structure. Only locations (or leaves) with data are specified; i.e., void areas are not explicitly listed. Also, the locational codes should be sorted hierarchically, not by numeric value. For example, the following is allowed: 1 511 521 61 62. The following list is not allowed, as it is not ordered correctly: 1 61 62 511 521.