Writing a reader plug-in involves a complicated set of computations to
take into account the source region of interest, the source subsampling
factors and offsets, the destination offset, and the size of the destination
image. In order to simplify things for application developers, a new
convenience method should be added to ImageReader that automates these
computations.

  If this method determines that the region to be read is empty, it
will throw and IllegalArgumentException. This exception needs to be
propagated upwards to the read and readAll methods whose implementations
in each plug-in may call computeRegions.

  Also, the existing utility methods getSourceRegion and
checkReadParamBandSettings should be made static in order to be consistent with
the other utility methods for checking the validity of the ImageReadParam
against the image being read.