Page MenuHomePhabricator

Improve copyright guidance
Closed, ResolvedPublic


a) The tool appears to be built to encourage rapid processing of opportunities, and not to build in caution. I think more caution should be built in. For example, there is no warning on the bot working page, that adding a link to an unauthorized version is a violation of WP:COPYLINK, and that adding such links can lead to a block of the user - that the user is responsible for their use of the tool
b) There should also be a warning that it is not common for the published version to be open access, and users should check carefully to ensure that if the bot suggests a link to the final published versions, that the user should check to make sure it is OK.
C) the link to "sherpa/romeo" is not right there on the working page. Instead, a link to sherpa is provided in this page that is linked-to from the bot page. That does not encourage use of sherpa. (is it possible for the sherpa results to be presented on the bot working page?)

Event Timeline

Ocaasi_WMF triaged this task as Medium priority.Oct 30 2017, 7:02 PM
Ocaasi_WMF created this task.

a) We should also signal to users that an edit was made (T179475).
c) We could use the API (|&mode=simple) to display the information directly on the page.

Ocaasi added subscribers: Pintoch, Ocaasi.EditedNov 27 2017, 11:10 PM

@Samwalton9 @Pintoch
So this feature is great, but I'd like to make it a notch better. I think Antonin can advise and maybe Sam you can wriggle out another pull request.

What's the idea?
The current results from Sherpa/Romeo are summary results, behind which are full journal guidelines.

Instead of a link saying: "Data from Sherpa/Romeo." (which links to the S/R homepage), I'd like it to say: "Data from Sherpa/Romeo. Full details available" (such that we link to the publisher page with all of the stipulations as well).

Example page we'd want to link to (for Nature publishing):

You can see we can search by journal title, or more precisely by ISSN (journal identifier).

Antonin, does Dissemin have ISSN's in the database? Is it as simple as appending the ISSN to a url query so that we can link to the full publisher info page with a variable/magic word for the ISSN that pulls from the API?

Sam, could you do this, technically, theoretically?


I started working on this ( but broke something along the way; I can't (locally) get the to return any proposed changes now. Would appreciate Antonin taking a quick look to see if the issue is obvious!

Hi @Samwalton9 - your refactoring is very sensible but it switched from to There was a bug in that API that I just fixed (I think) so you can try again… but there will be other issues down the line: for instance, this new API also returns ResearchGate URLs, that should probably be filtered out.

Samwalton9 closed this task as Resolved.EditedDec 12 2017, 2:05 PM
Samwalton9 claimed this task.

That seems to have been the issue - it's functioning fine now, but there are indeed a bunch of URLs showing up. Changes pushed to Github (; I'll go pull and restart the site now.

EDIT: Done, seems to be working well.