IO-Part

SAMT can handle different io-formats. Firstly it can read and write the ASCII-grid from ARCGIS. This format contains a header with the number of cols (int) and rows (int), the x- and y-corner (double), the cell size (float) and the nodata-values (int); and the ASCII-representation of grid cells (float). The reading of ASCII-grids is slow because of the translation of ASCII to the internal float representation. To overcome this problem was the main reason for introducing the HDF-file-format. HDF contains the same header information and additional information like the creation time, the modelers name etc. but it stores the grid cells in a binary format. This is much faster to read. HDF should be used to store a set of constant grids. In the underlying library there is no method to delete a grid from an HDF-file. Of course you can delete an HDF-file and construct a new one. For this reason the HDF is used as a container for basic grids (in the case of a project file) or as a container for one grid to exchange information between SAMT and external programs.
The hdf-filesystem can also handle meta data. For example the time stamp, the name of the modeler, the name of the model that produced the map are stored. This is useful to control the simulation (using the time stamp) or to manage a big set of maps. To use hdf files, please read the hints in the status line.
Additionally SAMT can read and write point data. Point data are used to characterize points in a landscape, for example points can be nesting places of birds. The structure of point data is very simple. It is an ASCII file with x y value (double double double). The x and y are the coordinates, value is the data at this location.

SAMT can read fuzzy models in a generalized file structure. This structure is influenced by the MATLAB fuzzy toolbox. A new interactive modeling tool is written in C++ and is part of SAMT. This tool can generate fuzzy models in this file structure.

Last but not least SAMT can read trained neural networks. This networks can be applied to input maps and produces an output map.
To open an neural network or fuzzy model the the action (FUZZY_Open, NN_Open) in the IO-Window has to be clicked. This opens a file dialog like this:

The selected model will be loaded in SAMT and can be used.

An important interface is the database connection to MySQL. MySQL is a free database wich is often used as a web-database under Linux. MySQL contains a free ODBC-interface for windows, so that ACCESS can be used as a front end for MySQL. This is a good way to integrate different users. In SAMT the database can be connected and it is easy to join tables with spatial data. To use this database connection an init-file ".my.cnf" must be created. This file has the following structure:

[database_name]
user database_user
host hostname
database database_name
password password

If the database connection was successful the button "MySQL_Get" is black in the menu table. An alternative to the database is a simple csv-file that can be used to include table information in SAMT.