fuck it, i'm going to write a personal oauth2 service

pretty soon on websites you'll see a sign in with google button, and sign in with github button, and a sign in with danny button

Good morning to everyone, except that one Twitter dev who one day woke up and was like "YOU KNOW WHAT, MY APPLICATION WILL FEATURE BOTH OAUTH1 AND OAUTH2 ENDPOINTS, BUT SOME FEATURES WILL BE EXCLUSIVE TO EITHER OF THE TWO -NOT NECESSARILY THE MOST RECENT, JUST A RANDOM ONE-, AND ALSO THE OFFICIAL TWITTER LIBRARY WON'T COVER ALL THE ENDPOINTS SO PEOPLE WILL HAVE TO RESORT TO RAW HTTP REQUESTS INSTEAD OF USING MY SDK AND ALSO I'MMA MAKE DEVELOPERS FILL 2 VERY DETAILED FORMS, REQUIRING PERSONAL DATA AND ACTUAL REAL PHONE CALLS, JUST TO START DEVELOPMENT WITH 7 DIFFERENT AUTHENTICATION TOKENS, BECAUSE SOME REQUESTS WILL REQUIRE A DIFFERENT AUTHENTICATION METHOD THAN THE OTHER REQUESTS DESPITE ALL OF THEM PERTAINING TO THE SAME FUCKING ENTITY"

User: If we use Oauth2, can we audit exactly where this data is going and who sends it there, and in addition cam we audit who grabs that data from the Authenticating app and make sure it doesn't violate our requirements?

Me: No

User: Why not?

Me: Because thats like asking us to audit whether or not a user accessed files and then uploaded them to their personal drive instead of corporate. We don't mandate that application owners take responsibility for their data outside of their application, why would we require that in this case???

User: Uhhhhh

FFS the lack of understanding of application accounts here boggles my mind. I understand that the security concerns are real but throwing out all permissible contexts based on a mandate that we dont even apply to extremely permissive accounts (i.e. users compared to apps) is folly