3 ms·
I miss the point why we really need some sort of "identity providers" when it's we who possess our own identities. It's the login-password pairs which are inco
by drdaeman 16y ago
I miss the point why we really need some sort of "identity providers" when it's we who possess our own identities.
It's the login-password pairs which are inconvenient (and, a lot of time, insecure) part out there. So, logically, the straight answer seem to be to get rid of this part in favor of something more usable, users can possess to prove their identities. For example, improving browser support for public-key crypto and key storage (think HTML5 <keygen> element). However, world has gone in a completely different direction, created "identity providers", who now actually possess users' identities (so you don't own your identity anymore), and, as a result, introduced more failure points to the auth process.
OpenID is neat and convenient when done right, but it was just a protocol invented to sign comments at other persons' blogs. It was never meant for personal identity management and architecturally lacks a lot of important points for that. XRIs may be considered some sort of solution for that (I-numbers are guaranteed to be persistent), but still feel somehow wrong.
- wvenable 16y ago> However, world has gone in a completely different direction, created "identity providers" That's because the question answered wasn't: "How do we solve the problem of user identity on the web?" The actual question answered was: "How do we solve the problem of user identity on the web using web services?" Once you define the problem that way, OpenID becomes the logical best solution. However, it's hardly the best way to do it in any general sense.