Properties in WebFinger
Paul E. Jones
Folks,
One of the features in WebFinger is the ability to define “properties” related to either a subject or a link.
I would expect properties defined for a link to be defined as a part of the link relation definition.
Properties that relate to the subject would be defined in separate RFCs or could be defined by other SDOs or anyone else, as properties are always fully-qualified URIs.
As an example, consider the following example:
"properties" :
{
"http://packetizer.com/ns/name" : "Paul E. Jones",
"http://packetizer.com/ns/name#zh-CN" : "保罗‧琼斯"
}
The “name” property is intended to convey the subject’s name with an optional language tag as a URI fragment. This is not defined anywhere, except here: http://packetizer.com/ns/name
A question to the group is this: do we need to define a registry for property values defined for a subject? The current WF spec does not define a registry. Would we want to define one? If so, should that go into the current draft? (I appreciate that this has gone to Last Call, so I’ll apologize for not raising this before.) If we did, what would the URIs look like? Does the IETF or IANA have a URI defined for this sort of thing?
Paul
James M Snell
I would suggest holding off and waiting to see if a registry is necessary... maybe just see how adoption goes over the next few months then decide.
_______________________________________________
webfinger mailing list
webf...@ietf.org
https://www.ietf.org/mailman/listinfo/webfinger
Melvin Carvalho
This is very similar to the @property element in HTML5
Generally what you would do is 'follow your nose' from the property URL to the description (via HTTP GET).
You could start a registry if you wanted, but it's been done a few times already, e.g. schema.org, foaf, vcard, open graph protocol ... do we want another one to maintain, and where would it be specified, who would look after it etc.
Is the XRD group still active, maybe they have something to say about JRD standardization?
--
Paul
---
You received this message because you are subscribed to the Google Groups "WebFinger" group.
To unsubscribe from this group and stop receiving emails from it, send an email to webfinger+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
Melvin Carvalho
Folks,
One of the features in WebFinger is the ability to define “properties” related to either a subject or a link.
I would expect properties defined for a link to be defined as a part of the link relation definition.
Properties that relate to the subject would be defined in separate RFCs or could be defined by other SDOs or anyone else, as properties are always fully-qualified URIs.
As an example, consider the following example:
"properties" :
{
"http://packetizer.com/ns/name" : "Paul E. Jones",
"http://packetizer.com/ns/name#zh-CN" : "保罗‧琼斯"
}
Very minor comment, would it not be better to put the properties under HTTPS?
The “name” property is intended to convey the subject’s name with an optional language tag as a URI fragment. This is not defined anywhere, except here: http://packetizer.com/ns/name
A question to the group is this: do we need to define a registry for property values defined for a subject? The current WF spec does not define a registry. Would we want to define one? If so, should that go into the current draft? (I appreciate that this has gone to Last Call, so I’ll apologize for not raising this before.) If we did, what would the URIs look like? Does the IETF or IANA have a URI defined for this sort of thing?
Paul
--
Paul E. Jones
Melvin,
A Property UI is just a “name”, so it names no difference if it were a URN an HTTP URI or HTTPS URI.
Paul
Melvin Carvalho
Melvin,
A Property UI is just a “name”, so it names no difference if it were a URN an HTTP URI or HTTPS URI.
Sure, I just meant that since WF is https oriented, it might be an idea to use https for these fields too. In the linked data world most schemas/vocabs/descriptions are http, but we are finding for some projects (particularly security oriented ones such as payments) putting things under https gives a bit more piece of mind and future proofing... :)
Kingsley Idehen
Melvin,
A Property UI is just a “name”, so it names no difference if it were a URN an HTTP URI or HTTPS URI.
If the above is true, then there's a subtlety that's being overlooked if the URI in question resolves to content the describes its referent i.e., the content can take the form of an entity relationship graph representing human- and machine-comprehensible entity relationship semantics.
Simple example: a URI denoting a property that's inversefunctional in nature e.g., the semantics of an property that associates one of more Agent URIs with an Email Address (mailto: scheme URI).
Links:
1. http://bit.ly/Y6TIfs -- how entity relationship semantics enable identity reconciliation .
Kingsley
-- Regards, Kingsley Idehen Founder & CEO OpenLink Software Company Web: http://www.openlinksw.com Personal Weblog: http://www.openlinksw.com/blog/~kidehen Twitter/Identi.ca handle: @kidehen Google+ Profile: https://plus.google.com/112399767740508618350/about LinkedIn Profile: http://www.linkedin.com/in/kidehen
Paul E. Jones
Kingsley,
A property consists of a name and value. The name of the property is a URI. The value is defined by whatever defines the property name. Presently, there are no properties defined in the WF spec. Only the syntax is defined. An example of a property is:
"http://packetizer.com/ns/name" : "Paul E. Jones",
----------------------------- -------------
| |
Property Name Property Value
What I was asking is whether we want to define a registry for property names. For example, perhaps urn:ietf:params:webfinger:<property_name> or something.
Paul
From: webfinge...@ietf.org [mailto:webfinge...@ietf.org] On Behalf Of Kingsley Idehen
Sent: Monday, March 04, 2013 3:39 PM
To: webf...@ietf.org
Subject: Re: [webfinger] Properties in WebFinger
On 3/4/13 3:00 PM, Paul E. Jones wrote:
Melvin,
A Property UI is just a “name”, so it names no difference if it were a URN an HTTP URI or HTTPS URI.
You mean a property name is just a URI? And that a URI just names the property?
If the above is true, then there's a subtlety that's being overlooked if the URI in question resolves to content the describes its referent i.e., the content can take the form of an entity relationship graph representing human- and machine-comprehensible entity relationship semantics.
Simple example: a URI denoting a property that's inversefunctional in nature e.g., the semantics of an property that associates one of more Agent URIs with an Email Address (mailto: scheme URI).
Links:
1. http://bit.ly/Y6TIfs -- how entity relationship semantics enable identity reconciliation .
Kingsley
Paul
_______________________________________________
webfinger mailing list
webf...@ietf.org
https://www.ietf.org/mailman/listinfo/webfinger
Paul E. Jones
Use of attributes in Connect might be preferred by users who don't mind sharing this information with certain RPs, but does not want to share this with the world.
Is there value in defining a URL for each of these attributes for use in WebFinger?
Paul
From: Nat Sakimura <saki...@gmail.com>
Sent: Mon Mar 11 12:09:37 EDT 2013
To: "Paul E. Jones" <pau...@packetizer.com>
Cc: webf...@ietf.org, webf...@googlegroups.com
Subject: Re: [webfinger] Properties in WebFinger
FYI, this looks very similar to Claims in OpenID Connect.
_______________________________________________
webfinger mailing list
webf...@ietf.org
https://www.ietf.org/mailman/listinfo/webfinger
Dick Hardt
Yes, that is very similar. A significant difference is that what is advertised via WebFinger is not necessarily something that one obtains through identity verification. For example, if I visit a blog and see a posting by you or if I get an email from you, I might want to know some of this information about you and I'm clearly not in a position to do an identity verification.
Use of attributes in Connect might be preferred by users who don't mind sharing this information with certain RPs, but does not want to share this with the world.
Is there value in defining a URL for each of these attributes for use in WebFinger?