Since this object is designed to work with geographic datasets, there needs to be some capability to import data from GIS data sets, and to export the output to the same GIS format. In this implementation, file input/output is handled by a separate datafile object, which is responsible for knowing the format of the GIS data files that it is importing, and converting them to and from the format that the GIS-CA object expexts.
One of the most important aspects to mention of the datafile object is what it does not do: it will not convert data from one spatial resolution or data format to another. It is designed to simply take a raster data set, read the data set grid cell by grid cell, and load those values into cell-objects of the same resolution. It will not handle raster files of differing resolutions. This is a very deliberate omission: combination of data sources of varying scales or degrees of resolution, while a common function in most commercially available GISs, introduces possible errors, which can be difficult to trace after the fact. It is desired in this case that if a modeler wishes to incorporate data of varying resolutions that s/he explicity conduct the necessary resampling of the varying themes before they are introduced into a spatially explicit model such as this one.
As of this writing, the GIS-CA datafile object reads and writes GIS files that are in GRASS or Arc/INFO GRID export format. The input/output code is segregated from the rest of the code in such a way that a user wishing to write his/her own import/export filter will be able to do so with a minimum of coding.
To better illustrate the utility of modeling landscape processes with this framework, an example of a landscape simulation using the GIS-CA toolkit will be presented here. This is a work in progress, and presented here for illustrative purposes only.