GBIF API changes

Package Use Questions
,
1

Hi all rgbif users out there (including pygbif users, if any)

GBIF wants to get feedback on proposed changes to their API with regard to the eventDate field, as well as related fields day, month, and year.

Give any feedback here https://github.com/gbif/gbif-api/issues/4 Or discuss here, and I can relay information over there

There are two questions at hand:

eventDate

I lean towards making eventDate contain date ranges, which it currently doesn’t. Does anyone have strong objections against this?

day/month/year fields

I vote to get rid of day, month, and year fields, but would anyone not like this? Does anyone use these fields?

definition of eventDate

Another issue is this: https://github.com/gbif/gbif-api/issues/3 That is, how is eventDate defined. Any thoughts on this?

2

Yes, I agree, I’ve wondered about these fields for a while. I can’t think of any back compatibility issues, though they may exist.
If this change is made Darwin Core should insist on ISO 8601, rather than just recommending it. This is one field where flexibility is no use to anyone.
You didn’t mention dwc:endDayOfYear and dwc:startDayOfYear. Is there a reason to keep these?

3

Are you saying you want those or you want to get rid of them? Note that this discussion only concerns the API (see What and how to return Occurrence eventDate? · Issue #4 · gbif/gbif-api · GitHub):

This is strictly related to the response of the API.

4

I don’t see any reason to keep them. I can’t quite think why they were there in the first place.

5

@qgroom Again, do you mean remove them from the Darwin Core dump files? Or from the API? Those field are not in the API anyway.