Minutes: XDI TC Telecon Thursday 2011-11-17

xdi message

Time: 1:00PM - 2:00PM Pacific Time (20:00-21:30 UTC)

1) IDTRUST BUDGET REQUESTS

We discussed potential requests for the XDI TC to submit to the IDTrust Member Section for budget for 2012. One specific proposal from Mike and Markus is at:

http://ox.gluu.org/doku.php?id=idtrust:proposal

We discussed this proposal for interoperability of a single XDI client with multiple XDI servers that are different implementations. This could also include interop of token formats for authorization, such as for OAuth and OpenID Connect.

We also discussed other possible interop-related projects for which funding could be requested.

Thirdly, we discussed working with the XRI TC so they could submit a proposal too.

The consensus was for Mike and Markus to work together to submit one or more proposals before tomorrow's deadline. They will contact Dee Shuur for details about how to submit.

# MIKE AND MARKUS to submit the proposal(s).

Kaliya was able to join the call and clarify that we should go ahead and submit the requests to Dee and then request feedback if it's not what they are looking for.

She also suggested that we submit a proposal for collaboration with the W3C and RDF communities working on Linked Data, and potentially for working with the OASIS Trust Elevation TC (see below).

# KALIYA to submit these proposal(s).

2) LINKED DATA UPDATE

Giovanni Bartolomeo sent a message with background about Linked Data [1], and had one question for Markus about how it was done in Higgins.

Markus said that he doubted how useful that work might be to the current XDI graph model because it was done using the previous model and with earlier Linked Data work.

Giovanni said it still may be helpful as background for making the case for how XRI and XDI may be useful for solving certain problems in Linked Data, particularly the context/provenance problem. Using XRIs as structured identifiers could be a good solution.

This is important because it is estimated that there are now 20 billion RDF nodes available as Linked Data, and the community is still growing. Giovanni said it may replace the notion of the original semantic web, simply because the size of the graph being created. It takes a different approach - instead of ontologies developed in the lab, it simply turns existing data into RDF graphs, and then lets different interested parties describe this real data in other RDF statements. So it is a more data-oriented, more practical approach. This is the reason that we should not only not ignore it, but seek directly to integrate with it. We should use our years of experience to help solve specific problems the Linked Data community has.

Mike asked how permission and access works in Linked Data. Giovanni said it was a good question - currently he believes that is all handled at the http/https protocol level, and not from within the graph itself. That's the new approach that XDI is taking. He said that there is also not a single identity standard that can be used by all parties sharing Linked Data graphs.

[1] Christian Bizer, Tom Heath, Tim Berners-Lee: Linked Data - The Story So Far. (2007). http://tomheath.com/papers/bizer-heath-berners-lee-ijswis-linked-data.pdf

Jay gave an update about PLOA - Personal Levels of Assurance. He explained that if you Google "LMNOP Street Identity", and then click Working Group to get to this page...

http://sites.google.com/site/streetidentitylmnop/workinggroup

...you see the list of resources contributed, and the AT&T white paper on PLOA is one of them:

http://www.idcommons.org/wp-content/uploads/2011/10/PLOA-White-Paper-v1.02.pdf

This is highly relevant to the XDI TC because it is so closely associated with link contracts and portable permissioning. We discussed how this topic could be pursued.

Jay explained that one lynchpin might be to work it into the user permissioning process where it can be shown to dramatically cut down the amount of work necessary to complete the process, as well as to be able to log and audit the process.

Jay also suggested that the combination of XDI and PLOA could go a very long ways to implementing the challenges of the new Trust Elevation TC. Drummond agreed.

Kaliya said that we should consider asking for budget for colloboration activity with the Trust Elevation TC.

Mike said that the basic implementation of OXServer is basically done. It implements JWT 0.6 tokens as proposed by OpenID Connect 1.0, plus OAuth signing. OXServer 0.2 will validate that the XRI for the sender of the message is verified, and the integrity of the message.

To demonstrate single-sign-on, they plan to implement a new server, seed.open-xdi.org, that will give everyone on the TC the ability to use the application being written for ID-Cubed, as well as against their own personal graph. So it will essentially be a personal graph server.

There is much more coming - SAML tokens, PKI authentication, ease-of-use improvements.

Mike also said that the team wrote an "object graph mapper" that gives you a mapping between a POJO (plain old Java object) and an XDI graph. This tool essentially synchronizes changes between the Java object and the XDI graph. Mike called it "Hibernate for XDI".