It not clear if the DOI as persistent identifier is mandatory for the metadata as well as embedded with the data
or if the mandatory part is only applicable to the metadata and optional for the data.
Before the DOI , repositories maybe have a different persistent identifier stored in the data and metadata
that was not a DOI.
Adding DOI it would require work from the repositories an somewhat changing the metadata is
simpler compare to changing also the data
Data maybe distributed in different repositories. It would be good to have words/guideline on who should
assign the DOI. An obvious example is an international mission where the data are collocated in the
repositories of the country participation to the mission.
It should be clarified if the assignment of the DOI is at the repository level or different approach.
A repository may have the capability to search data sets from different mission where each data set has already
a persistent identifier. Repositories may have a mechanism to assign a DOI to the collection (dynamic DOI).
There are questions of whether or not if these dynamic DOI is recommended or it would be better to recommend
users to record in the publication the DOI for each data set used in the publication
From "Angelini, Lorella (GSFC-6610)" lorella.angelini-1@nasa.gov:
It not clear if the DOI as persistent identifier is mandatory for the metadata as well as embedded with the data or if the mandatory part is only applicable to the metadata and optional for the data. Before the DOI , repositories maybe have a different persistent identifier stored in the data and metadata that was not a DOI.
Adding DOI it would require work from the repositories an somewhat changing the metadata is simpler compare to changing also the data
Data maybe distributed in different repositories. It would be good to have words/guideline on who should assign the DOI. An obvious example is an international mission where the data are collocated in the repositories of the country participation to the mission. It should be clarified if the assignment of the DOI is at the repository level or different approach.
A repository may have the capability to search data sets from different mission where each data set has already a persistent identifier. Repositories may have a mechanism to assign a DOI to the collection (dynamic DOI). There are questions of whether or not if these dynamic DOI is recommended or it would be better to recommend users to record in the publication the DOI for each data set used in the publication