LTER-LIFE / FDFDT

FAIR Data for Digital Twins
0 stars 0 forks source link

Use of persistent identifiers #9

Open CherineJ opened 5 months ago

CherineJ commented 5 months ago

In Darwin Core Archives records are assigned IDs, which link records across the different tables. It is recommended in the Darwin Core documentation to use uuid for that. This can easily be done, but it remains challenging to ensure that these IDs are persistent. It is not completely clear how to deal with the IDs if entries change.

For data sets, the identifier would stay the same if records within the data set change, so that the identifier always refers to the latest version, but there should be versioning in place to trace back such changes. For records within files or data sets, this is however more challenging and no clear recommendations exists how to handle this.

Kim and Cees will forward this question/issue to Wim. There are also documentations on persistent identifiers (cf. here and here, but they also don't give clear recommendations for our concrete problem.

For now, we will stick to our human-readable, hierarchical, unique-within-datasets ID system that we have used for the bud burst and cricket data sets.