Is both dataset version & released needed?

mdoering commented 4 years ago

version (string) and released (date) seems very much the same property of a dataset. Should we better only keep the release date?

Version: 2019-03-06 Release date: 2019-03-06

Version: Nov 2017 Release date: 2017-11-18

Version: Jan 2014 Release date: 2014-06-18

Version: Feb 2018 Release date: 2018-02-15

mdoering commented 4 years ago

ping @gdower @yroskov

yroskov commented 4 years ago

We need BOTH:

version, (1) as applied by data provider inside its own project (it might number only (ver. 1.34, or ver 5.), number and date (ver. 21 of Feb 2018), or date only (DD/MM/YYYY or MM/YYYY or YYYY), and (2) as data provider asks users to cite his data;
release date, a date when export for the CoL was generated (single format DD/MM/YYYY, better YYYY-MM-DD). We may receive different exports (with different dates) of the same version of database, and those exports might be different for various technical reasons. Date in the version and Release Date might be identical, close each other or very much different. For example, when we'll update ILDIS, version will remain as ver. 12, May 2014, but release date will be 2020-02-10. If dataset version important for both users and CoL management, Release Date is more important for CoL management. Dataset version goes to the bibliographic citation; release date remains only in digital metadata.

mdoering commented 4 years ago

thanks, that makes sense. Release date was not cear to me and sounds a bit misleading. maybe export date would be better?

CatalogueOfLife / backend