search

openphacts / FederatedPhacts

Experiments in federated querying
3 stars 1 forks source link

Should we use identifiers.org in between different parts of queries #8

Open Chris-Evelo opened 5 years ago

Chris-Evelo commented 5 years ago

The thought here is that if you carry out a first query you will get identifiers for the queried resource returned in whatever format that resource prefers at that time. However, you do know that what you get really is an identifier for that resource. So it would be relatively easy to convert the result into identifiers.org compliant identifiers, and you could probably do that in a generic way that will survive identifier schema changes (the most meaningful part is typically at the end). Using these for the next part of the query would have the advantage of making the whole query persistent even if the target URL schema changes. However, it does come at a cost (the extra step of going to identifiers.org) and it needs identifies.org itself to be up to date.

Chris-Evelo commented 5 years ago

Nick is going to ask Manuel about identifiers.org status since it moved to the cloud.

Chris-Evelo commented 5 years ago

Nick indicated that the original version of identifies.org could actually work in both directions. So yes it could convert whatever you got from the first resource into the persistent identifiers.org URI and the use that to query the next one (where it would essentially replace it with the then-current generic URI for the selected service, and could even find an active service). Manuel's answer left us uncertain about whether the current one can though.