[OpenID] An Ode to OpenID

Peter Williams pwilliams at rapattoni.com
Tue Sep 8 16:46:22 UTC 2009 

I rebuilt the SSL side of his experiment, working just with a (modified) source of the tomcat server's JSSE socket factory. Rather than use FOAF and SPARQL inferences as the account linking mechanism, I tried out  same basic flow using an XRI  proxy to resolve a webid that had the form of an HXRI (to fit better with the openid foundation's overall vision). Account linking mechanism done the  XRI way uses synonym testing as it basis This contrasts with the preferences of the semweb folks who seem to use a bases of Bayesian logic or formal reasoning.

The whole concept allows something little seen in practice in the SAML2 world despite being specified: letting the users choice of "SP affiliation name" [read webid] act as a linking point for several IDPs' [read OP  providers'] Subject names [read openid identifiers]. Leveraging ssl to assert the webid does things better than SAML in my view - as the user gets both to choose the link name and to use it at different RPs. This keeps the constant power struggle of the IDP-interests vs RP-interests vs user-interests in balance.

If I might comment a little, now, the balancing act of poliics is the critical matter. I find the combination of the use of ssl handshake clientauth (to assert the webid to the RP) and the SSL server's (trusted) resolution of the foaf file (with which the RP may  then do the plural account linking of several openids onto the webid) to be very natural.

But let's not delude ourselves. OpenID delegation pulls the UCI information. Foaf+ssl uses ssl to push the UCI information.  If Google cannot bring itself to support openid delegation in practice, users may well find themselves using the likes of foaf+ssl to introduce additional providers who can add then add back in the UCI value, with or without the consent of the OP.


From: Santosh Rajan [mailto:santrajan at gmail.com]
Sent: Tuesday, September 08, 2009 1:56 AM
To: Peter Williams
Cc: general at openid.net
Subject: Re: [OpenID] An Ode to OpenID

I also got a post in private from an foaf guy, that he can do authentication from the browser. It makes sense to me, though i don't know much about foaf.

Please do post the same thing here, bcos i can't understand why you did not post it for everyone. (I dont think OPenID general, has any IP issues).
On Tue, Sep 8, 2009 at 2:09 PM, Santosh Rajan <santrajan at gmail.com<mailto:santrajan at gmail.com>> wrote:
Hey Peter,
The article was written on June 23, and was actually meant for the layman, also, I did not understand the ramifications completely at that point in time.

As I had promised in that article,that I would demonstrate with an example in a future post. But I never got around to doing that post, because that is when it hit me "the browser will manage your XRD". No matter how you conduct that experiment, the browser will always be at the center. No escaping that. And that is a whole new ball game. No discovery, no delegation. It is all about the user. (Whatever he wants to identify himself with).



On Tue, Sep 8, 2009 at 1:53 PM, Peter Williams <pwilliams at rapattoni.com<mailto:pwilliams at rapattoni.com>> wrote:
The article makes scant reference to any federated sso - where federated means talking to someone in a foreign security domain that you don't control.

At the same time, it does seem to validate the infocard message -invoking a secure window manager for sensitive operations. Can we characterize chrome (in a secure window manager) as an alternative to the Microsoft infocard support?

Let's not forget that openid and infocard signalling have a shared history.



On Sep 7, 2009, at 10:11 PM, "Santosh Rajan" <santrajan at gmail.com<mailto:santrajan at gmail.com><mailto:santrajan at gmail.com<mailto:santrajan at gmail.com>>> wrote:

Hehe, the ode was supposed to make you laugh. I am sorry it didn't.

What is actually most interesting, short of pimping my own blog, is "federated identity in browsers".

I have been working on it for a couple on months with some unbeleivable results.
A few days back readwriteweb<http://www.readwriteweb.com/archives/google_chrome_os_to_feature_single_sign-on.php> posted about Google Single sign on in the new chrome OS. Their conclusions are ridiculous, even though the actual result of such a move is very exciting.

 The results of my experiments in "Federated Identity in browsers" have proven that, there will be no "Discovery or Delegation" any more. Thats right!

Think about it. Your Browser would store and manage your XRD. Browser vendors would be "KING".

Thats when you realize what a bunch of idiots the XRI TC is made of.



On Tue, Sep 8, 2009 at 6:53 AM, Peter Williams <<mailto:pwilliams at rapattoni.com<mailto:pwilliams at rapattoni.com>>pwilliams at rapattoni.com<mailto:pwilliams at rapattoni.com><mailto:pwilliams at rapattoni.com<mailto:pwilliams at rapattoni.com>>> wrote:

The not particularly interesting ode video led me to your (much more interesting) blog site, one of whose entries talked about delegation using Opera Unite. No odes were ever found, but one odyssey did result.


1. I followed the instructions, and got myself a vanity openid: <http://home.homepw.operaunite.com/a> http://home.homepw.operaunite.com/a that delegates to <http://homepw.myopenid.com> homepw.myopenid.com<http://homepw.myopenid.com><http://homepw.myopenid.com>



This is cute, because it only seems to work when I'm online (since only then is the index.html doing the delegation available to relying party sites).


Used at Slashdot (based on Janrain's RPX integration, I believe), <http://home.homepw.operaunite.com/a> http://home.homepw.operaunite.com/a got me to the Slashdot account linking screen. It failed at Plaxo.




2. So then I went to another RP, <http://freexri.com> freexri.com<http://freexri.com><http://freexri.com> and hit the login button (using the openid signin option). Obviously, I supplied my shiny new vanity name from operaunite: <http://home.homepw.operaunite.com/a> http://home.homepw.operaunite.com/a. Things properly pinged back to myopenid, and I authorized release of an assertion and the null persona of attributes.



3. I created myself a shiny new XRI, which is called @id*<http://home.homepw.operaunite.com>home.homepw.operaunite.com<http://home.homepw.operaunite.com><http://home.homepw.operaunite.com>, using my current account which has several XRIs (you may need a new account, if this is your first time through XRI land), providing the capcha. Thereafter, I selected the option for the XRI that it would have an openid i-service attached. That is: I now have @id*<http://home.homepw.operaunite.com>home.homepw.operaunite.com<http://home.homepw.operaunite.com><http://home.homepw.operaunite.com> delegating to <http://home.homepw.operaunite.com/a> http://home.homepw.operaunite.com/a (which ...delegates to ...). If I do XRI resolution on that name, I see all  that properly reflected in the XRD markup, using local name of <http://home.homepw.operaunite.com/a> http://home.homepw.operaunite.com/a (operating under the openid (not XRI) semantics of an local name since the local name is in the SEP/link, not the XRD body).



4. I logged out of the RP, fully. Logging back in to this RP (nominally to create a synonym to next see if the polyarchical stuff really "works" with delegation), I obviously now cited my shiny new XRI @id*<http://home.homepw.operaunite.com>home.homepw.operaunite.com<http://home.homepw.operaunite.com><http://home.homepw.operaunite.com>. The myopenid page eventually showed, but wants me to  solve this:



"myOpenID is not authorized to verify that "@!B1E8.C27B.E41C.25C3!5adb.130d.2ac2.be9e" is your identifier. If it is your identifier, you can set up myOpenID to verify it. See the help page for more information."


How can I do this in my XRD? (Assume I login to the <http://freexri.com> freexri.com<http://freexri.com><http://freexri.com> RP locally, to get back my admin privileges, since Im operating a master XRI account)



5. Then I did the same at plaxo, for @id*<http://home.homepw.operaunite.com>home.homepw.operaunite.com<http://home.homepw.operaunite.com><http://home.homepw.operaunite.com>



This time I got



"myOpenID is not authorized to verify that "<http://home.homepw.operaunite.com/a/content/>http://home.homepw.operaunite.com/a/content/" is your identifier. If it is your identifier, you can set up myOpenID to verify it. See the help page for more information."



To solve THIS one, I think I need merely put the magic "proof" file (that proves control over the "domain") onto the webserver under the above path. Does this feel correct? If so, Ill add the myopenid proof file to my file system (!).


6. Remembering that my XRI also has a +contact i-service bound to it, with its own delegations built in to the HXRI variant of my XRI, I had a look at the contact page that one gets from following 302's issued by the <https://xri.freexri.com/@id*home.homepw.operaunite.com> https://xri.freexri.com/@id*home.homepw.operaunite.com.



At the end of the chain, at <http://contact.freexri.com/contact/@id*home.homepw.operaunite.com> http://contact.freexri.com/contact/@id*home.homepw.operaunite.com



We see the HTML-centric metadata:


<link rel="openid.server" href="<http://www.myopenid.com/server>http://www.myopenid.com/server" />

<link rel="openid2.provider" href="<http://www.myopenid.com/server>http://www.myopenid.com/server" />

<link rel="openid.delegate" href="<http://xri.net/http://home.homepw.operaunite.com/a/content/<http://xri.net/http:/home.homepw.operaunite.com/a/content/>>http://xri.net/http://home.homepw.operaunite.com/a/content/<http://xri.net/http:/home.homepw.operaunite.com/a/content/>" />

<link rel="openid2.local_id" href="<http://xri.net/http://home.homepw.operaunite.com/a/content/<http://xri.net/http:/home.homepw.operaunite.com/a/content/>>http://xri.net/http://home.homepw.operaunite.com/a/content/<http://xri.net/http:/home.homepw.operaunite.com/a/content/>" />





I find it strange that one resolution of the contact service at XRI proxy would delegate to the synonmy resolution service at another proxy. Additionally, the form of the local_id looks weird, but is not actually illegal, apparently.  But, let's try it at the plaxo. anyways.



After ignoring the problems with the https variant (since plaxo  doesn't know about the freexri SSL cert domain), we got


myOpenID is not authorized to verify that <http://xri.net/http://home.homepw.operaunite.com/a/content/<http://xri.net/http:/home.homepw.operaunite.com/a/content/>> http://xri.net/http://home.homepw.operaunite.com/a/content/<http://xri.net/http:/home.homepw.operaunite.com/a/content/> is your identifier. If it is your identifier, you can set up myOpenID to verify it. See the help page<https://www.myopenid.com/help#own_domain> for more information.



Well at least that's familiar. How do I now fiddle around with XRI/XRD or my contact page so I can make myopenid happy?





7 I find it strange that Slashdot could get myopenid to assert for my opera delegation uri but myopenid will not react to more advanced cases.



There seems to be a distinction being drawn, between delegation and delegation for domains. The former is detected and myopenid finds it acceptable to make an assertion, but the latter cases require additional proof over control of the domain/URL.



I wish the spec was clear on all this!



-----Original Message-----
From: <mailto:openid-general-bounces at lists.openid.net<mailto:openid-general-bounces at lists.openid.net>> openid-general-bounces at lists.openid.net<mailto:openid-general-bounces at lists.openid.net><mailto:openid-general-bounces at lists.openid.net<mailto:openid-general-bounces at lists.openid.net>> [mailto:<mailto:openid-general-bounces at lists.openid.net<mailto:openid-general-bounces at lists.openid.net>>openid-general-bounces at lists.openid.net<mailto:openid-general-bounces at lists.openid.net><mailto:openid-general-bounces at lists.openid.net<mailto:openid-general-bounces at lists.openid.net>>] On Behalf Of Santosh Rajan
Sent: Monday, September 07, 2009 12:13 PM
To: <mailto:general at openid.net<mailto:general at openid.net>> general at openid.net<mailto:general at openid.net><mailto:general at openid.net<mailto:general at openid.net>>
Subject: [OpenID] An Ode to OpenID










<http://www.youtube.com/watch?v=ztc4V3ttlso&feature=related>http://www.youtube.com/watch?v=ztc4V3ttlso&feature=related









-----



Santosh Rajan

<http://santrajan.blogspot.com>http://santrajan.blogspot.com <http://santrajan.blogspot.com> http://santrajan.blogspot.com

--

View this message in context: <http://www.nabble.com/An-Ode-to-OpenID-tp25335016p25335016.html> http://www.nabble.com/An-Ode-to-OpenID-tp25335016p25335016.html

Sent from the OpenID - General mailing list archive at Nabble.com<http://Nabble.com>.



_______________________________________________

general mailing list

<mailto:general at lists.openid.net<mailto:general at lists.openid.net>>general at lists.openid.net<mailto:general at lists.openid.net><mailto:general at lists.openid.net<mailto:general at lists.openid.net>>

<http://lists.openid.net/mailman/listinfo/openid-general>http://lists.openid.net/mailman/listinfo/openid-general



--
<http://santrajan.blogspot.com>http://santrajan.blogspot.com
<http://twitter.com/santoshrajan>http://twitter.com/santoshrajan
<http://www.facebook.com/santosh.rajan>http://www.facebook.com/santosh.rajan



--
http://santrajan.blogspot.com
http://twitter.com/santoshrajan
http://www.facebook.com/santosh.rajan




--
http://santrajan.blogspot.com
http://twitter.com/santoshrajan
http://www.facebook.com/santosh.rajan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-general/attachments/20090908/6a66663b/attachment-0001.htm>

More information about the general mailing list