Following are the minutes of the unofficial telecon of the XDI TC at:
Date: Friday, 25 January 2013 USA
Time: 9:00AM - 10:30AM Pacific Time (16:00-17:30 UTC)
ATTENDING
Markus Sabadello
Les Chasen
Drummond Reed
Phil Windley
Joseph Boyle
REGRETS
Giovanni Bartolomeo
Bill Barnhill
THE ETHERPAD LINK FOR TODAY IS:
AGENDA
***** NEWS & UPDATES *****
--- FIRST PERSONAL CLOUD MEETUP IN SAN FRANCISCO TUESDAY JAN 29
Drummond Reed, Phil Windley, Joseph Boyle, and potentially other XDI TC members will be attending and touching on the role of XDI in a personal cloud network.
***** PRESENTATIONS/DISCUSSIONS *****
--- TYPED VARIABLES
Drummond shared a new proposal he's started for Typed Variables, which he believes will be a key component of XDI forms:
He will add more detail to this proposal, which currently has a low priority.
--- XDI EVENTS
Since we have been modelling XDI link contract policies for event channels, Markus and Drummond did a informative writeup of what it would look like to fully model events (state changes outside the XDI graph) that are communicated with XDI messages.
We discussed this with Phil. He agreed that in his work on events in KRL (Kinetic Rules Language), he needs several capabilities that XDI can help with:
-
Channel negotiation between endpoints
-
Event channel portability
However Phil also noted some constraints that KRL implementations must live with, including that some KRL endpoints may not be able to do discovery. Another constraint may be how long it takes to raise an event.
We discussed several requirements around discovery for event channel endpoints. Global discovery is very advantageous, but cannot add too much latency or it could choke the system. Phil felt caching would be as essential as it is with DNS. Drummond agreed, and said we need to add caching support to the XdiDiscovery proposal.
We also discussed that KRL uses event channels that are both local and global in scope, so channel discovery needs to work at both levels. Lastly, it is also a requirement that event channels not necessarily be correlatable to a user's identity.
Les asked about whether sharing channel endpoints is an attack vector because they are network endpoints to which messages can be sent. Phil said this is why XDI link contract policies are currently being developed to provide event channel authorization.
Phil said that some channels he is working with also need to be available via other transport protocols such as SMTP. So this is also a requirement for XDI Discovery.
***** DECISION POINTS FOR THIS CALL *****
This week's decision queue is the following set of proposals:
--- ADDRESSING INNER ROOTS
A proposal for a one-line change to the XDI ABNF to enable addressing of inner roots:
We reviewed this proposal with Phil and he was fine with it.
--- EQUIVALENCE LINKS AND REFERENCE LINKS (DRUMMOND & MARKUS)
This is still on the queue as Bill wanted to consider breaking it into multiple proposals. However Bill was not on the call so we will wait one more week.
***** DECISION POINT QUEUE REVIEW *****
The decision queue stack is now shown on the following three auto-generated Category pages:
See also this list of proposals that need to be developed:
***** NEXT CALL *****
Drummond has a conflict for next week's call, but we did not have time to discuss moving it, so he will send a message to the list about moving it.