Concerning "discrete" vs. "continuous" data

I realized belatedly that we did not discuss this concept in class, though it
is an important one.  By "discrete" data I mean "spatially discrete" in the
sense of having discrete boundaries between areas with different attribute
values: for example administrative boundaries such as counties, surficial
geology, vegetative communities, land use, etc.  In contrast, spatially
"continuous" data are those for which clearcut boundaries between areas with
different attribute values cannot be assumed:  for example, elevation varies
continuously across the landscape, as do many elevation-derived values; air
temperature or other meteorological data also change continuously across the
landscape.

The point of distinguishing these spatial data types at the time it falls in
the outline of the course is that Vector GIS does a better job of modeling
discrete mapped data, and in most cases Raster GIS does a better job of
modeling continuous mapped data.

This is not to say that you cannot use a Vector GIS to model data we're
assuming to be continuous and vice-versa.  Furthermore, there are some spatial
variables for which discrete vs. continuous is debatable, for example soils
and vegetation at some scales can be thought of as continuous.

Decisions about the most appropriate choice of spatial data model must make
certain assumptions and we must understand those assumptions and their
implications.

Jerry