OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

xdi message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: RE: [xdi] XDI contexts and cross-references


Ok, good I think I have it then. What XRI's would assert the two statements with literals?  Since we're saying an XDI graph can be represented as a set of XRIs, we need XRI's to represent the low level values.  I forgot about " char not being valid XRI syntax again, but is there a mechanism for asserting datatype property statements like those?
 
Without quotes my next choice would be something along the lines of using the data: uri scheme with base64 encoding and mime type for small literals and MIME with cid/mid for large literals:
=drummond+home/+email/$xsd$string(data:text/plain;base64,ZHNyLmV4YW1wbGVAZ21haWwuY29t)
=drummond+work/+email/$xsd$string(data:text/plain;base64,ZHJ1bW1vbmQuZXhhbXBsZUBjb3JkYW5jZS5uZXQ=)
 
I think's it's necessary to retain the ability to fully describe the XDI graph using n XRIs, and so I think we need some way to XRI encode those last two assertions.
 
What are everyone's thoughts on this?
 
Bill


From: Drummond Reed [mailto:drummond.reed@cordance.net]
Sent: Friday, March 28, 2008 3:27 PM
To: Barnhill, William [USA]; 'Markus Sabadello'
Cc: xdi@lists.oasis-open.org
Subject: RE: [xdi] XDI contexts and cross-references

Yes, Bill, those are the XRIs being asserted except the last two. The literals in the last two statements are not part of the XRIs. The XRIs are just:

 

=drummond+work/+email

=drummond+home/+email

 

This XRI is the address of the literal, but the literal itself it not part of the address. It can’t be because: a) it’s not an XRI, and b) XRIs can’t include quotes.

 

RE using the dollar sign as a suffix, yes, that would be legal XRI syntax _under the proposal for direct concatenation_ (see http://wiki.oasis-open.org/xri/XriThree/SyntaxAbnf). (Caveat, as per the XDI RDF Model doc: that is not an approved XRI 3.0 syntax yet, just the proposal we’re working with.) The proposed semantics – and nothing more than that – around using $ as a suffix on a predicate is that it means “the canonical version of a set of instances”, i.e., if +email was the predicate, +email$ means that no matter how many subtypes/instances of an email address you have, what is the one (and only one) considered canonical (which may not be any of them of course).

 

Hope this helps,

 

=Drummond

 


From: Barnhill, William [USA] [mailto:barnhill_william@bah.com]
Sent: Friday, March 28, 2008 10:41 AM
To: Drummond Reed; Markus Sabadello
Cc: xdi@lists.oasis-open.org
Subject: RE: [xdi] XDI contexts and cross-references

 

Hmm, okay, so if I was inputting the example in your email to an XDI graph I would be asserting the following XRIs, correct?

=drummond/+email//=drummond+home/+email

=drummond/+email//=drummond+work/+email

=drummond/+email+home/(=drummond+home/+email)

=drummond/+email+work/(=drummond+work/+email)

=drummond/+email$/(=drummond+work/+email) (Is this valid XRI syntax with the $, and what is the meaning?)

=drummond+home/+email/“dsr.example@gmail.com

=drummond+work/+email/“drummond.example@cordance.net

 

Bill

 


From: Drummond Reed [mailto:drummond.reed@cordance.net]
Sent: Friday, March 28, 2008 1:31 PM
To: 'Markus Sabadello'
Cc: xdi@lists.oasis-open.org
Subject: RE: [xdi] XDI contexts and cross-references

Good point, Markus – since I only had one XDI subject in these examples, the cross-reference example was relative to that XDI subject. But yes, absolutely, cross-references can be either relative or absolute.

 

To illustrate, here’s an XDI document with three subjects that uses absolute cross-references across the subjects. It also shows a way to more deeply contextualize data context, because now the actual literal email addresses for =drummond are in separate XDI subjects that represent contextualize personas of =drummond (in this case home and work).

 

=drummond

            +email

                        /

                                    =drummond+home

                                                +email

                                    =drummond+work

                                                +email

            +email+home

                        (=drummond+home/+email)

            +email+work

                        (=drummond+work/+email)

            +email$

                        (=drummond+work/+email)

=drummond+home

            +email

                        “dsr.example@gmail.com”

=drummond+work

            +email

                        “drummond.example@cordance.net”

 

=Drummond

 


From: markus.sabadello@gmail.com [mailto:markus.sabadello@gmail.com] On Behalf Of Markus Sabadello
Sent: Friday, March 28, 2008 2:56 AM
To: Drummond Reed
Cc: xdi@lists.oasis-open.org
Subject: Re: [xdi] XDI contexts and cross-references

 


Does the cross-reference always have to be "relative to the same XDI subject"? Can't it be an absolute address?

Markus

On Fri, Mar 28, 2008 at 8:25 AM, Drummond Reed <drummond.reed@cordance.net> wrote:

Another question that came up on today's call involved the meaning of the following statement (a simplified version of the example we were discussing from the XDI RDF Model doc, http://wiki.oasis-open.org/xdi/XdiRdfModel):

 

      =drummond/+email//=drummond/+email+home

 

Does that mean that =drummond/+email is an equivalent XDI statement to =drummond/+email+home?

 

No, it only means that the statement =drummond/+email is the parent context for the statement =drummond/+email+home. =drummond/+email could also be the parent context for the statement =drummond/+email+work. In X3 Simple, this would graph as:

 

=drummond

      +email

            /

                  =drummond

                        +email+home

                        +email+work

 

So if you did an XDI $get on =drummond/+email, the response would be the entire graph above. That graph in turn could be part of this larger graph:

 

=drummond

      +email

            /

                  =drummond

                        +email+home

                        +email+work

      +email+home

            "dsr.example@gmail.com"

      +email+work

            "drummond.example@cordance.net"

      +email$

            (/+email+work)

 

This larger graph includes the cross-reference "(/+email+work)". This reads as, "relative to the same XDI subject, refer to the +email+work predicate". Unlike a subcontext, this statement IS saying that one XDI statement refers directly to another one, i.e., the first XDI statement below refers to the second one:

 

      =drummond/+email$/(/+email+work)

      =drummond/+email+work

 

So, for purposes of illustration of these two concepts – XDI contexts and XDI cross-references, an XDI request/response pattern against the full graph above, translated from English into X3, might be:

 

Q: "Markus would like Drummond's email address."

 

=markus

      $get

            /

                  =drummond

                        +email

 

A: "Drummond has two email addresses, home and work."

 

=drummond

      +email

            /

                  =drummond

                        +email+home

                        +email+work

 

Q: "Then Markus would like Drummond's home email address."

 

=markus

      $get

            /

                  =drummond

                        +email+home

 

A: "Here is the value."

 

=drummond

      +email+home

            "dsr.example@gmail.com"

 

*************

A different request response pattern against the same path might be:

 

Q: "Markus would like Drummond's canonical email address."

 

=markus

      $get

            /

                  =drummond

                        +email$

 

A: "Drummond's canonical email address is his work email address."

 

=drummond

      +email$

            (/+email+work)

 

Q: "Then Markus would like Drummond's work email address."

 

=markus

      $get

            /

                  =drummond

                        +email+work

 

A: "Here is the value."

 

=drummond

      +email+work

            "drummond.example@cordance.net"

 

***********

Note that one of the last two roundtrips could be eliminated by the simple rule that the server would resolve xrefs within the same XDI context. This would result in the following request/response:

 

Q: "Markus would like Drummond's canonical email address."

 

=markus

      $get

            /

                  =drummond

                        +email$

 

A: "Drummond's canonical email address is his work email address, and here is the value."

 

=drummond

      +email$

            (/+email+work)

      +email+work

            "drummond.example@cordance.net"

 

 

 

 



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]