search

iobis / env-data

ENV-DATA related issues and documentation
2 stars 0 forks source link

Different datasets contain different data from the same sample #18

Open Daphnisd opened 5 years ago

Daphnisd commented 5 years ago

Samples can be spit up into different taxon groups and be analysed by different research groups. Therefore it's often preferred to store it as different datasets (different people involved and thus different data attribution).

Can this data be stored as different datasets where the different eventIDs refer to the same sample? If so, how will OBIS know that the eventID's refer to the same sample? You can assign the same eventID, but how does OBIS know whether the same eventID is used deliberately or accidental?

Daphnisd commented 5 years ago

EurOBIS will be sending some such datasets in the coming year to OBIS (and another node asked me about the same issue). We will a list of all stationVisits which could possibly be published separately if useful?

During a workshop in February 2018 one option which was suggested is to have URI's as the eventIDs. While this is an interesting option, the overhead for the OBIS node to make these URI resolve is quite high.

pieterprovoost commented 5 years ago

We currently have no plans for linking resources across datasets. Of course you are welcome to reuse URIs in multiple datasets to allow the user to analyze the combined data. When a user will query data based on eventID (to be implemented), records from both datasets will be returned, OBIS doesn't need to "know" that that these are linked.

Daphnisd commented 5 years ago

An option to generate URIs could be: generate a DOI for the samplelist/cruise report and format all eventIDs as https://doi.org/**DOI**#**ID**. DOI would be the DOI of the stationlist. ID would be the id of the sample/event in the samplelist. This type of URI (using the #) will still resolve to the landingpage of the DOI.