Disclaimer: I work for Mozilla, I maintain django-browserid (and StravosK is a valued contributor many sites. This is all just my own personal opinion. I was very bullish on Persona early on, but the fact of the matter is, we failed . And not just because (as I feel is being implied) some higher up suddenly came over and asked for an unreasonable amount of adoption for a revolutionary product. We failed for a thousan…
>it had a major privacy issue of leaking your identity provider and relying party to Mozilla via a centralized iframe we host. A mailing list thread among the Persona devs and community failed to find a solution to this. Can you elaborate on this?
Basically, Persona by design needs to transmit a keypair and a signed certificate between the identity provider and the relying party. Without in-browser integration, the way this was done was using localStorage on the login.persona.org domain, making Mozilla a trusted third party. This was deemed okay as a temporary measure until in-browser support for BrowserID became commonplace, replacing the trusted third party with the user-agent.
Reading the thread again, my memory was a little off: possible solutions were proposed, but the conversation fell off before consensus could be reached, and no one had the time to drive any attempt to implement the change.