Hi people,

I have to apologize first, I am by no means an expert (or even any good) web
developer. Still I read some of the protocol descriptions and here I am
*seriously* baffled.

Suppose we have "a server", "a client", and "a third party". As long as a third
party is not involved, the client (or, precisely, the web browser application
that runs on hardware that is under the user's control) interects directly with
the server, and OpenID protocol is used.

When a third party (usually another web server) needs to get some limited access
to the server on behalf of the client, this third party is not fully trusted by
either the client or the server. OAuth protocol is used here in place of OpenID
in order to (1) not to share the client's credentials with the third party, and
(2) to restrict the third party to a limited subset of operations/data that the
client can perform/access on the server.

Now consider a desktop or mobile application. It is software that runs on a
device that is controlled by the user, and trust in it is no different from the
trust in trust in a web browser application! It is no "incomplete trust" issue,
no desire to conceal the user's credentials from the app, no reason to limit its
access to the operations/data on the server. In other words, it calls for the
use of OpenID protocol rather than OAuth. OAuth is designed for completely
different circumstances than the desktop/mobile application.

So, why people try to use OAuth when OpenID is seemingly a better match?
What am I missing?

Thanks,

Eugene

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
StatusNet-dev mailing list
[email protected]
http://lists.status.net/mailman/listinfo/statusnet-dev

Reply via email to