My first feeling was that this is a bit intrusive but then...
Here is a picture of the authnapped OpenId form:
Here is a picture of the original OpenId login:
It is the user's decision to install and use Mozilla Lab's project "weave" or not. And this solves parts of the NASCAR problem. Why should the service provider suggest some OpenId providers using the NASCAR? Well, if he has a whitelist of trusted OPs then yes.
But the OpenId-NASCAR is a cludge anyway.
I think that there should be an XRD description of which authentication methods and providers and token formats and so on a service provider supports or requires. Then a client component - read Browser extension - could help the user to make a good decision and prevent phishing attacks and more.
The user does not care whether the protocol is OpenId or Information Card or if the token format is SAML2 or what not. A unique user experience is desired. Ease of use is required. User consent is required. Security and Privacy need to be protected.
This should be "in the browser"! Secure by default. Privacy protecting by default.
I guess I don't have to repeat that I prefer the Information Card metaphor and UI. A client component is a good thing and it should be ubiquious, build-in but replacable and configurable at the user's choice.
Identification, authentication and claims/attribute transfer is not the primary service provider's interest. Those tasks should be moved outside of the website's code into an authnapping module of the user's browser.
Authnapping is good!
Thursday, July 23, 2009
Auth-napping OpenId by Weave
Posted by
Unknown
at
12:32 PM
2
comments
Labels: information card, Information Card Foundation, OpenId, privacy, security, weave
Friday, October 24, 2008
XMLDAP XRDS
The XMLDAP relying party is now updated to provide XRDS data.
Notice the Information Card icon in the lower right corner in the status-bar of the browser.
You can start the card selector either through the sidebar - by dragging a card to the main window - or by clicking the Information Card icon. You need the latest version of the openinfocard id selector.
You may be wondering what the difference between the next and the previous image is?
I created a second card and used it at the xmldap relying party too.
The new claims were added to the previous set of claims. The claim "locality == Berlin" is new.
This image shows that the claim set was cleared. The relyingparty party has forgotten the privacy data after the "Clear privacy data" button was pressed.
THIS IS THE USER EXPERIENCE YOU WANT. DEATH TO USERNAME/PASSWORD.
(learn more)
Posted by
Unknown
at
9:35 PM
0
comments
Labels: information card, Information Card Foundation, information card icon, openinfocard, privacy, xmldap
Tuesday, August 19, 2008
Privacy in the Silly Season
In Germany there are currently lots of news stories about private data beeing stolen / misused. It started with a set of 17000 that contained information about lottery customers. Next a data protection commissioner bought a CD with information about 6 million people where 4 million dates contained bank account information. There where several thousand cases where money in the range from 50 to 100 Euro was withdrawn from bank accounts. The latest news: a call center with access to the Deutsche Telekom database abused its access rights to steal data. The number of illegal accesses and of stolen data is unknown...
Well, now politicians - you never heard of - voice their concerns and demand stronger punishment and a data protection office demands that the selling of personal identifiable data must be prohibited by law... The government has decided to wait for the investigation to be finished.
It is good that all this is in the news every hour. But we have to wait what actually changes when the politicians and people are back from vacation.
In relation to these stories I try to translate Kim's laymen's laws of identity: Laws of Identity (in German)
Posted by
Unknown
at
10:30 AM
0
comments
Labels: privacy, The Laws of Identity
Monday, May 12, 2008
IIW2008a Monday: <link rel="metadata" ...>
Reading through the proposed topics for IIW2008a I noticed that George Fletcher blogged about something that I want too.
Though calling it Identity Metasystem Markup Language seems a little too big, I think.
Anyway I posted something similar to the osis-general mailing list on May 2nd.
Using <link rel="metadata" ...> to indicate what the RP wants is a good idea, I think. This is very simple and very much simpler than embedded objects.
What I like most about this idea is that we might get rid of any RP login form etc all together. If the browser/UA/client notices this new "link" then it can display a dialog or whatever to the user and there is no need for RP generated login forms. This way we have a unified user experience at the user's choice.
What we need next: Do the same with privacy statements. There should be a "link" to the privacy statement of the RP (maybe this is part of the metadata already?). The privacy statement should be kind of machine readable too. I want the browser to be able to help the user to make the right decision here. Well, we need much more but this might be a start.
----
This topic is related to "identity selector advertising"... The selector should/could advertise to the RP that it understands to handle "link metadata" and then the RP could avoid sending <object type="application/x-informationcard" ..> because it now knows that the id selector will offer the user the option to send his claims.
Posted by
Unknown
at
5:04 PM
0
comments
Labels: iiw2008a, metadata, privacy, user agent, user centric identity
Thursday, March 06, 2008
Microsoft acquired Credentica
This is such a smart move of Microsoft! I am impressed and I am sure that Credentica's technology will lead to a privacy improved version of CardSpace. I hope that Microsoft will provide open access to this technology for others to implement identity selectors, relying parties and security token servers. CardSpace is token agnostic but when I have read the U-Prove papers correctly then there is more then one roundtrip between id selector and STS required to deliver all the nice features. The protocol between id selector and STS thus probably has to be changed.
"Nice features": Some weeks ago somebody asked me "What is this (group signatures, zero knowledge based algorithms, electronic money based algorithms) good for. Where can anonymity, privacy, pseudonyms, untracebility, unlinkability, ... be used". At that time I answered this questions on a too technological level, interpreting the word "where" to mean protocols, signatures, encryption. The better answers would have been, and the acquisition of Credentica by Microsoft points in this direction: If the technology in available on every desktop computer or mobile computer/phone then users will learn to want privacy, untracebility and unlinkability. The services a company offers must have these features then or the users will use services that provide them.
Posted by
Unknown
at
9:50 PM
0
comments
Labels: CardSpace, Credentica, Microsoft, privacy, unlinkability, untracebility
Saturday, March 01, 2008
IdentityCampBremen
| This sounds interesting. "Identitycamp Bremen will be the first Barcamp in Germany that focuses on identity 2.0, single-sign-on, reputation management, relationship management, privacy 2.0 and related issues." |
The new ePA and its pseudonym feature should be talked about then too. We should help to build the need-to-know society. We should use claims instead of identification.
Posted by
Unknown
at
8:49 PM
0
comments
Labels: identity, privacy, relationship management, reputation management, single-sign-on

