Le jeu. 24 janv. 2019 à 14:29, Matthew Paul Thomas via desktop-devel-list <desktop-devel-list gnome org> a écrit :
None of that is to say that GOA shouldn’t exist. It has the potential to
save people time. “I set up an account in App A. Now it will be quicker
to use the same account in App B.” Anything more than that is noise.
While currently the case, GNOME basically is responsible for every app that uses GOA to connect to accounts, when said account uses OAuth. Which means that it takes one app to badly behave for the account provider to shut down access to GNOME, and thus GOA and all the apps that depend on it. This is why the idea of having applications use their own keys is better - not only you won't GNOME own the keys (better UX since from the account dashboard, under applications connected, the user won't just see "GNOME" but every application they have granted access to), but one misbehaving app won't trip the whole system.
Now, by adding application-provided keys, the user will have to sign in to every application in order to grant access, which defeats the purpose of having GOA as a "centralized user account repository". Therefore GOA would need to basically be redesigned from the ground up.
The underlying idea of GOA is good, however it will need to gain more features, and not less, to stay useful; for example have application be able to share content through those accounts, instead of only being able to send emails (although, maybe another implementation of a sharing feature could work with application that would provide and consume sharing endpoints).
As for Documents, I had come across Paperwork while exploring the Flathub repository, which seems like what Documents should have been to me. It has tags support, automatic OCR, and global text search, among other things.
