So, we actually already do this for citoid citations. The issue for this particular case is that we get the results from Zotero, and Zotero is telling us that the full text url is that doi link; if the url is missing from Zotero we fill it in with the resolved location, otherwise we accept whatever have there.

I guess we could check every Zotero citation for unresolved doi links, but is this really a big deal? Isn't the idea of doi links to prevent linkrot anyway?

I'm going to decline this but feel free to comment if you think otherwise ^-^

In T106857#2571504, @Mvolz wrote:

Isn't the idea of doi links to prevent linkrot anyway?

Yes, I believe so so why are we even adding URLs in the first place?

In T106857#2571544, @Josve05a wrote:

In T106857#2571504, @Mvolz wrote:

Isn't the idea of doi links to prevent linkrot anyway?

Yes, I believe so so why are we even adding URLs in the first place?

1. In case there is no doi
2. Possibly to reflect where the original metadata was retrieved from, as there may be more metadata at the retrieval source that might help disambiguate it?

In T106857#2571567, @Mvolz wrote:

In T106857#2571544, @Josve05a wrote:

In T106857#2571504, @Mvolz wrote:

Isn't the idea of doi links to prevent linkrot anyway?

Yes, I believe so so why are we even adding URLs in the first place?

1. In case there is no doi
2. Possibly to reflect where the original metadata was retrieved from, as there may be more metadata at the retrieval source that might help disambiguate it?

But if there is an DOI, and the URL is the same as the doi (pointing to the doi.org domain), then why add that? Can't we make a check that if the end-link points to doi..org, and a doi is present, do not include a url? Otherwise, we would just add the same doi and link twice.