This spike is to determine how to capture version information within the Qube config files for csvcubed; the main part of this spike though is to figure out whether we need to capture three types of version information to accurately reflect the status of any particular dataset.
There are three areas which we would need to cover in order to properly convey comparability across editions within a publication.
[ ] Versioning of data
[ ] Semantic versioning of what constitutes a new release (i.e. a new time period in a time series dataset)
[ ] Semantic versioning of what constitutes an update
[ ] Changes to existing codelist members, or addition of new codes
[ ] Versioning of schema
[ ] Layout changes of source files, does this matter?
[ ] Layout changes of the DSD breaking cross-release "mechanical" comparability
[ ] Versioning of methodology
[ ] Is there a statistical break to the releases which renders comparability between periods impossible
robons
commented
2 years ago
This spike is to determine how to capture version information within the Qube config files for csvcubed; the main part of this spike though is to figure out whether we need to capture three types of version information to accurately reflect the status of any particular dataset.
There are three areas which we would need to cover in order to properly convey comparability across editions within a publication.