The CurlyWhirly data format is a tab delimited text format with a header which specifies the category groups which categories are found in, as well as the axis names for the data set. The category column header should start with the prefix categories: and the header line for the column which contains your data point names should be label.
A header line should resemble the following:
categories:size categories:age label PCO1 PCO2 PCO3
You can specify a database which links into the dataset in an optional comment line before the header line. You do this by specifying a URL which accepts the name of a data point as a parameter. Ideally this URL should be to specify more information about the data point. The format of the URL should resemble:
A full sample of the start of a file may look something like the following:
URL=http://mylookup.address/name= categories:size categories:age label PCO1 PCO2 PCO3 large young point1 -0.3552 -0.1541 -0.0363 medium old point2 -0.344 -0.0556 -0.1038 small young point3 -0.3411 -0.0542 -0.066