On Tue, 23 Jul 2002, Wilson, Andrew, J wrote:
> Thanks Ugo, Bill,
> By the time the seismic data gets to me it is, without
> exception, regularly sampled. That is the sampling interval
> in each dimension is constant (but may be differnt between
> Typically we index the values according to inline, crossline
> and time. Think of time as equivalent(ish) to depth. So all
> three of these can be integers.
> Is there any computational advantage to using an Integer3DSet
> over a Linear3DSet?
No, its just there for the simpler constructor signature.