[OpenID] OpenID Discovery for Email like identifiers - Draft 0.1

Peter Williams pwilliams at rapattoni.com
Tue Jun 2 18:51:52 UTC 2009


I venture that any id is "acceptable" - provided it delivers all the openid-featured properties of URLs.

1. you must be able to type it
2. it must enable names of user to be written in the user's native script
3. there should (only) be feature for permancy and synonyms
4. the name must discover an authoritative XRD (or XRDS) by XRI Resolution or by simple resource location


Anything on top of those 4 constraints is a bonus. E.G. clickpass, google/yahoo UI conventions, myopenid outsourcing, facebook ispassive. These latter features merely allow vendors to distinguish themselves from each other. Providing such value add does not interfere with the open systems rules (any 2+ vendors systems will cooperate at the community-agreed minimum level, without requiring vendor intervention), they are fine. The market will sort out what folks will or will not adopt.

The test of a good id  would be that, if encoded in a URL and when supported by a URL->id gateway resolver (like HXRIs), the RP conformance suite would not even know that the id might have properties in addition to those from URLs.

Folks proved that the fantastically complex XRI Resolution can be  reduced to URL resolution, using HXRI proxies. Similarly, w3c showed that sparql queries can reduce many useful rdf graph searches to a URL. So, there is little excuse to make exceptions now for jabberids, email ids, or other.

Ideally, the XRI proxy would be the gateway of name resolver gateways ; but that's unlikely to happen. Name serving is far to much of a hot-topic - given its geopolitical control properties - for there to ever be such consensus on that. For example, one can give the US a pass concerning its geo-political manipulation of the DNS, for lots of mostly positive historical reasons. This doesn't mean USG gets a pass in the future though.


________________________________
From: general-bounces at openid.net [general-bounces at openid.net] On Behalf Of Dirk Balfanz [balfanz at google.com]
Sent: Tuesday, June 02, 2009 11:03 AM
To: Santosh Rajan
Cc: general at openid.net
Subject: Re: [OpenID] OpenID Discovery for Email like identifiers - Draft 0.1

In general, I don't think you need to "adapt" webfinger. Webfinger gives you everything you need. The OpenID community just needs to decide whether the email-like identifiers falling out of webfinger are acceptable OpenIDs.

Couple of specific comments:

- re signatures: I would just use whatever the XRI TC comes up with there for XRD signing.
- I don't think you should have the openid21.provider Rel-type with the URI. Since you start with an email(-like) identifier, it is clear that you're performing discovery on an identifier for a user, not for a host. So you should only look for a Link element that has a URITemplate in it. If you want to map all users to the same endpoint, just make it a template that doesn't actually have any unbound variables in it.
- "openid21.usermeta" should should be something generic like "describedby"  - as you point out that's the user's _meta data_, not just OpenID stuff (although there was some discussion at IIW that "describedby" isn't appropriate b/c the email-like identifier isn't technically a URI, and URITemplates, and "describedby" are supposed to be for URIs - but I think that's a technicality we'll work out).
- the openid21.local_id being its own Link seems weird to me. Why not something like this:

<Link>
  <Rel>openid.provider</Rel>
  <URI>http://openidprovider.com/op</URI>
  <openid:LocalID xmlns:openid="http://...">some-local-id</openid:LocalID>
</Link>

Dirk.

On Tue, Jun 2, 2009 at 6:28 AM, Santosh Rajan <santrajan at gmail.com<mailto:santrajan at gmail.com>> wrote:

I have posted a draft spec on the wiki to get the ball rolling on this one.
http://wiki.openid.net/OpenID-discovery-for-Email-Like-identifiers OpenID
Discovery for Email like identifiers - Draft 0.1
Lets have some discussion on this in order to add this into the 2.1
proposal.

-----

Santosh Rajan
http://santrajan.blogspot.com http://santrajan.blogspot.com
--
View this message in context: http://www.nabble.com/OpenID-Discovery-for-Email-like-identifiers---Draft-0.1-tp23832524p23832524.html
Sent from the OpenID - General mailing list archive at Nabble.com.

_______________________________________________
general mailing list
general at openid.net<mailto:general at openid.net>
http://openid.net/mailman/listinfo/general




More information about the general mailing list