Dataset identifiers

[pieterprovoost]

The spreadsheet template uses dataset identifiers to link together entries. Can we expect datasets to have DOIs at this stage, or do we want to add a DOI column so it can be added later without having to change identifiers throughout the sheet?

[kmexter]

DOI would go in the access sheet, as it is about how to access the data/metadata....but do you think it should be moved to here? Thing is, not all datasets will have a doi (or any PID)

[pieterprovoost]

DOI is commonly recommended as an option for a dataset PID (see GO FAIR, Google, Science on Schema) so I think it's perfectly fine to use as an identifier to link together the elements in the spreadsheet. If a DOI is not available then there are other PID options like a UUID, see https://www.go-fair.org/fair-principles/f1-meta-data-assigned-globally-unique-persistent-identifiers/.

In any case, I think we should have an "identifiers" column which allows adding additional identifiers for a dataset (handle, ARK, DOI, UUID, etc). This example from Google has DOI and ARK for example: https://developers.google.com/search/docs/appearance/structured-data/dataset

[kmexter]

The identifiers column is purely for internal use - especially in the sheet "links to other data/publications" which allows you to identify which datasets came from which other ones. I don't want to change this. In the sheet Access there is now a single column "DownloadURL" in which the DOI (/URL) to the dataset landing page should be added. 1- is that sufficient? 2- do we want more than one dataset DownloadURL? it is a lot easier if we just ask for one.......

[kmexter]

We need to decide if we are going to have the "distribution" in our json files or not, as currently we are only asking for a landing page, not a download page

[kmexter]

Given that we are now seeing that datasets are coming from diverse sources - some are from data repos/catalogues and have a DOI that points to a data landing page, some are URLs that point to a data selection tool (not data itself), some are publications, some are even GH repos. I think we are now at the point where we can look at what we are getting and think about how to include this info. Hopefully this is a discussion for our WP1 meeting on Sept 4

[ptagliolato]

If we opt for distinguishing different cases (PID landing page, direct download URL, some web page/web form to request data, and even some service to retrieve the data from), the last dcat version offers specific properties to accomplish this. See examples at https://www.w3.org/TR/vocab-dcat-3/#example-landing-page