Nascar for Sharing

From IIW
Revision as of 16:06, 3 February 2011 by WikiSysop (talk | contribs) (Undo revision 3238 by Igiwydijok (Talk))

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Issue/Topic: Talking About Sharing Nascar and how to do Personal Discovery

Monday - Session 1 - D

Conference: IIW10 May 17-19, 2009 this is the complete Complete Set of Notes

Convener: Jian Shen

Notes-taker(s): Charlie Reverte SK

A. Tags for the session - technology discussed/ideas considered:

B. Discussion notes, key understandings, outstanding questions, observations, and, if appropriate to this discussion: action items, next steps:

will meyer - addthis

jim fenton - cisco

jian shen - meebo

michael hanson - mozilla

jari koivisto - cisco

jared hanson -


  • will: xauth as an XRD cache? need to switch it to be service vs. host oriented. need this to support long tail of sharing
  • jian: sharing tools want to be able to use any sharing service vs. a pre-defined set, not just oauth + xmpp
  • will: want protocol support for dynamic service discovery so users can add any site they want. this is a different use case from xauth "is a person logged in to X" use case. "services I use" is > services I'm logged in to, > services I have accounts at. how do you track services a user is interested in using. store in webfinger? local storage is browser-specific and not public. problem is how to record for a user, "what services do they care about?"
  • jim: problem is that there are too many services that people want to use. assumption was that there is a small set of services. what if a user has multiple personas or identity providers? not everything is a browser. knowing the identity provider might reveal information about a user.
  • michael: discovery is disclosure, discloses personal information
  • all: don't want to use nastyporn.com as their identity provider
  • mike: want lots of services, late binding
  • will: keep url sharing simple at first and then build on top of it
  • jian: what services make the most sense on given sites?
  • charlie: let the user specify
  • michael: MRU is frequently right. just killed css visited hack
  • will: standardize something below "what is the best service algorithms"
  • michael: explaining web powerbox. content-type-specific service discovery.
  • will: services are handling more and more types of content, more compelling to solve problem of finding out services people want to use rather than content-type based negotiation
  • michael: spec is trying to handle discovery for physical resources, very noun centric. can be extended to web? invoke flickr instead of camera? movement to move local computing resources into url domain
  • will: good, now your list of services is even larger. who registers the handlers?
  • michael: still vague, multiple options
  • will: want to be able to define "I use flickr for images" and have it be a global handler, persists across machines and be publicly knowable so others can discover it.
  • micheal: opportunity for browser to discover services and persist them to webfinger via XRDP. xprefs in the browser
  • jian: would love for service prefs to be in the browser, won't have to communicate with each service
  • michael: we need to articulate the possibilities for storing the prefs
  • jared: options are converging, xauth can cache your webfinger with browser support. re xauth, doesn't like that services can query it without your consent. powerbox lets page declare services it's interested in
  • jian: good idea to notify the user but companies are pinging each other in the background anyway. propose spec so people do it more in the open and so it's more seamless
  • will: practical issues: way for site to allow user to add that site as a preferred service. this can work with xrdp with some changes. what's the latest status of xrdp?
  • jared: thinks google will implement xrdp on their webfinger
  • will: charlie has xrdp demo, would love to start hooking in
  • michael: thinks webfinger might be the weak link for the long tail. difficult to implement
  • charlie: long tail sites can delegate their xrdp/webfinger management to others
  • jared: where is mozilla on this?
  • michael: firefox could do this as an addon. looking to see if this can make a cross device impl with weave that can persist to webfinger. can do a lighthouse impl as a tech demo
  • jared: going to come down to who implements webfinger etc.
  • michael: has webfinger support with contacts plugin. <demos it>
  • will: all of the pieces are there to put this all together. <oexchange demo>. good use case if a site can add a "save me as a preferred service" button, can use for personalization and persist to webfinger.
  • michael: public, protected and private services are separate use cases, webfinger is good for public case. powerbox is a private impl because it's local.
  • jian: trying to do openid and trust exchange. xauth providers should be audited.
  • will: use case of "send content to my mom". can't discover services she uses unless they're public, webfingerable
  • jian: likes the idea of generic contacts, describing which protocols they use. need to associate ids across various networks.
  • will: long tail niche sites can drive much higher engagement
  • michael: facebook going for universaility, question whether it can handle all use cases
  • michael: propose host-meta relation for source code link "source-of"

--going to follow up with more convo in xrdp, oexchange People did not notice me because I was over weight. My family and my girlfriends were also upset with me. The perfect weight loss solution then came into my life and everything changed, The superfood Acai Berry changed my life. And now after using Acai Berry I am the happiest person in the world. Try it and you will also be always thankful to Acai Berry.