well with OpenID atleast, I think we can easily design a logout extension, where an RP can register it's logout handler with OP during login flow (checkid_immediate/checkid_setup) and the OP could call each of the RP's logout handlers (in the browser) that are registered with the current session when a user logs out from the OP. Ofcourse this also requires a new logout mode supported by OP, which an RP can call when a user wants to logout from the RP itself. The OP then can provide options to the user whether to sign out from one RP or from all.
Any reason why something like this was not incorporated into the specs yet ? thx Praveen Johannes Ernst wrote: > That might be hard from a usability perspective, and in my experience, > the underlying user requirement tends to be a variation of "I am about > to go to lunch with the guys waiting in the hall way, log me out of > all apps I'm currently logged in but take no more than 10 seconds > because otherwise they will leave without me. Or at least the critical > ones." (which is where it gets hard to design this right) Where > sessions come in is that again from a usability perspective, the user > should not have to "log out" from apps that he currently isn't logged > into (because the session expired, for example). > > > On Apr 6, 2007, at 10:51, McGovern, James F ((HTSC, IT)) wrote: > >> I would think that you wouldn't need to track the notion of a session >> but have something where the selector that tracked where the card was >> previously sent in terms of a list would allow you to graphically >> send another event. You could optionally walk a list based on each card. >> >> -----Original Message----- >> *From:* Johannes Ernst [mailto:[EMAIL PROTECTED] >> *Sent:* Friday, April 06, 2007 12:29 PM >> *To:* McGovern, James F (HTSC, IT) >> *Cc:* specs@openid.net <mailto:specs@openid.net> >> *Subject:* Re: Logout >> >> So far, neither OpenID nor CardSpace define the notion of a >> session, so no common logout is possible within the standard >> protocols. >> >> What we do in our code at NetMesh is to add a convention where >> RP-URL?lid=OPENID >> is the same thing as "submitted OpenID URL in the first form", to >> which the RP-URL responds with a redirect to the OP, while >> RP-URL?lid= >> means "become anonymous again" aka "logout". >> >> There are substantial usability issues with common logout in a >> decentralized, "internet-scale" approach, however, that nobody >> has really solved as far as I know. >> >> >> >> >> ************************************************************************* >> This communication, including attachments, is >> for the exclusive use of addressee and may contain proprietary, >> confidential and/or privileged information. If you are not the intended >> recipient, any use, copying, disclosure, dissemination or distribution is >> strictly prohibited. If you are not the intended recipient, please notify >> the sender immediately by return e-mail, delete this communication and >> destroy all copies. >> ************************************************************************* >> _______________________________________________ >> specs mailing list >> specs@openid.net <mailto:specs@openid.net> >> http://openid.net/mailman/listinfo/specs > > = > ------------------------------------------------------------------------ > > _______________________________________________ > specs mailing list > specs@openid.net > http://openid.net/mailman/listinfo/specs > _______________________________________________ specs mailing list specs@openid.net http://openid.net/mailman/listinfo/specs
