Objectives

The design has these objectives:

• use a well understood format (we choose “comma separated values” a.k.a. CSV)

• files editable with a text editor and commonly available software (Excel, Google Sheets, Calc etc.)

• avoid duplication of data as far as possible to minimise possibilities for inconsistency

• include the minimum amount of data needed by Cantabular

• contents of files are self-describing (you can tell from the header line what you are looking at)

• use similar formats for variables and mappings

• avoid having to load very large files to edit most data

• make version control / diffs easier and more comprehensible by splitting into multiple files

• use full stop as a separator because doubling up to escape is readable unlike hyphen or underscore

• can be used to represent geographic hierarchy via large mappings