Name: A name for the dataset. Actually, a directory with that name will be created to hold all the files and folders for the dataset. A dataset is always a collection of files.
Path: The Path where the folder for the dataset should be created in. This path can always be the same but the resulting directory (Path + ’\’ + Name + ’\’) should be empty.
Data type: One of the four dataset types: “Image”, “Elevation”, “Vector” or “Annotation”.
Storage method: Can be “Combined” or “File per tile”.
In combined mode, the dataset will consist of only a handful of files and no subdirectories. The main data file may become very large and some older file systems like FAT32 are not able to store files larger than 4 gigabytes.
In file per tile mode, a new file is created for each tile. A tile is the smallest portion of data that is loaded or “paged in” at once. The files are automatically organized in a directory tree. A tile file is typically smaller than 100 kilobytes. The number of files may become huge (about 4 million files for an image dataset of 100 gigabytes) with larger datasets and some file system might not handle this very efficient. Also, more storage space is used because of cluster waste and moving or deleting the dataset will be much slower. The benefit of this storage method is that it is better suited for web streaming and more robust towards data