XDI TC Minutes

Following are the minutes of the unofficial telecon of the XDI TC at:

Date: Friday, 23 May 2014 USA
Time: 09:00AM - 10:30AM Pacific Time (16:00-17:30 UTC)

ATTENDING

Drummond Reed
Dan Blum
Peter Davis
Markus Sabadello
Hubert Le Van Gong
Phil Windley
Les Chasen
Brian Wu
Animesh Chowdhury
Joseph Boyle

GUESTS

REGRETS

NEWS & UPDATES

The XDI TC congratulated Phil Windley on rejoining the XDI TC as a member of Brigham Young University OIT.

PRESENTATIONS/DISCUSSIONS

Link contract shift

The current and new link contract patterns are documented here:

https://wiki.oasis-open.org/xdi/LinkContractPattern

https://wiki.oasis-open.org/xdi/LinkContractPattern/Discussion

We reviewed the following notes from the weekly XDI2 call on 22nd May 2014, during which we took a detailed look at the way e.g. the public link contract of a graph changes after the shift:

https://www.oasis-open.org/committees/document.php?document_id=53126&wg_abbrev=xdi

We recognized that the context node [=]!:uuid:1111<#email> appears twice in the diagram, once with the value, and once as part of the link contract. Animesh expressed concerns that from a programming standpoint, the two nodes would not be considered the same, and that there should be a $is arc between them. Animesh added that a signature of the graph in the inner root will be different from a signature on a subgraph of the outer root.

Drummond explained that the addresses of the different nodes are the same RELATIVE to their respective roots, and that therefore the nodes are considered to be LOGICALLY EQUIVALENT. All data under all roots is always understood to be the same everywhere.

Animesh asked if you delete the node under the outer root, will the node under the inner root get deleted? Drummond answered no.

Joseph asked if XDI2 stores logical links between corresponding nodes in different roots. Markus explained this was not the case; however XDI2 has code that understands this logical equivalence.

Animesh then suggested that maybe link contracts should just be identified by UUIDs, and information about the requesting party and authorizing party should be added as attributes, rather than being part of the address. Drummond pointed out that one has to be able to algorithmically determine the address of a link contract, in order to be able to use it without any a-priori exchange. Drummond added that after the link contract shift, this approach makes even more sense.

Animesh added that with the UUID approach, all aspects of a link contract could be changed without having to change its address. Animesh is concerned that XDI clients would run into problems if the address of a link contract changes. Drummond thinks any change of parties to a link contract would have to be propagated to all participants anyway. This is a feature, not a problem.

Les asked, if he didn’t want to allow access to his e-mail address anymore, would he have to update a link contract in Drummond’s graph? Drummond answered, no, the link contract would have to be updated in Les’ own graph. Such an update may however propagate through a push or pull mechanism.

Dan said this reminded him of the usage policies he had been working on. Link contract policies matter when a requesting party talks to an authorizing party, whereas usage policies matter after a requesting party has obtained data.

Dan also noted that we haven’t yet established the details of the push/pull mechanism. Drummond agreed, but mentioned that the $copy permission will be used. Hubert asked how exactly such a mechanism would be triggered. Drummond answered it would be specified by the link contract that is being used.

Report from XDI editors subcommittee

Peter has posted the following report of the 5/22 XDI Editors call::

https://lists.oasis-open.org/archives/xdi/201405/msg00035.html

XDI policies

Dan would like to discuss the current XDI Policy draft:

https://www.oasis-open.org/committees/download.php/52892/XDIPolicyDraft.docx

We did not have time to further discuss this.

XDI signatures

Peter has posted an updated XDI signature proposal:

https://lists.oasis-open.org/archives/xdi/201405/msg00030.html