In the email below I didn’t mean to
imply that this proposal came from Steve – I simply wanted to credit him
with the idea of using a new element to clarify CanonicalID semantics. Steve’s
idea was in fact to use two new elements (<useCanonicalID> and
<allowUseCanonicalID>) for asserting and authorizing CanonicalIDs for
which the authority for an XRD was not authoritative.
The GlobalID+zero-or-one-CanonicalID proposal
written up at http://wiki.oasis-open.org/xri/XriCd02/CanonicalIdVerification
is a different way of achieving that same end. I discussed it in person with
Markus tonight (we’re both at the Higgins F2F meeting at Catalyst in San
Francisco) and besides solving the problem of how an authority authorizing a CanonicalID
synonym can persistently reference the authority being authorized, it’s biggest
virtue is that it makes CanonicalIDs truly canonical. In other words, making
the cardinaliy of the CanonicalID element zero-or-one gives it the same status
as the ProviderID element in an XRD, i.e., this is THE identifier that relying
parties should use as the persistent primary key for the target resource.
This simplifies the CanonicalID graph tremendously,
because at any one point it time, a resource only exists in one CanonicalID hierarchy,
even if it exists in multiple polyarchys by virtue of its Ref and Backref elements.
This in turn can simplify processing by
all applications that consume XRIs, as well as simply XRI management by XRI
authorities. Everyone wins.
Again, I encourage everyone to take the
time to read through this proposal and provide any feedback so we can finish
ED03.
=Drummond
From: =drummond.reed
[mailto:drummond.reed@cordance.net]
Sent: Monday, June 25, 2007 12:49
AM
To: xri@lists.oasis-open.org
Subject: [xri] XRD GlobalID
element proposal
In working on the Canonical ID
verification section of XRI Resolution 2.0 Working Draft 11 ED03 this weekend,
Markus’ message below combined with an idea Steve Churchill had in a long
Canonical ID verification discussion we had this past week to generate a
proposal for a small change that I believe is a significant improvement to both
XRI synonym architecture and Canonical ID verification.
The proposal is simply to:
1) Add a new XRD synonym element called
GlobalID (that “fills the gap” between LocalID and CanonicalID
synonyms).
2) Change the cardinality of the
CanonicalID element to zero-or-one (so it becomes truly
“canonical”).
Although very subtle, it is nonetheless
still an important change. I studied this it closely before writing it up in a
revised version of the Canonical ID verification wiki page:
http://wiki.oasis-open.org/xri/XriCd02/CanonicalIdVerification
See the new section 6 of this page. The problem
it solves in Canonical ID verification, together with the other benefits it
brings XRI synonym architecture, are explained in detail in the writeup.
Since we need to close on this before I
can finish the revised Canonical ID verification section, and I’d really
like to include that section in ED03, please review this proposal and post any
feedback (to the wiki or to the list) as soon as you can.
Thanks,
=Drummond
From:
markus.sabadello@gmail.com [mailto:markus.sabadello@gmail.com] On Behalf Of Markus Sabadello
Sent: Friday, June 22, 2007 8:03
AM
To: =drummond.reed
Cc: xri@lists.oasis-open.org
Subject: Re: [xri] Review of
extended CID verification (was RE: [xri] Agenda:Joint XRI & XDI TC Telecon
10AM PT Thursday 2007-06-21)
Hello,
I can't find much to comment on either.. Well there is one thing that bothers
me, but maybe I got something wrong here: The text always says that in order to
verify a CanonicalID from a different authority, you look for either a Ref or
Backref pointing to the original identifier. But wouldn't another CanonicalID
also be legal?
For example, I am trying to verify http://www.markus.at.
The XRDS I get from there has a
<CanonicalID>@id*markus</CanonicalID>. So I resolve @id*markus, and
I find no Ref and no Backref there, but a <CanonicalID> http://www.markus.at</CanonicalID>. I
know that's somehow stupid and maybe breaks some other rules, but shouldn't it
count as "verified"?
I also have two thoughts about CanonicalIDs in general, not verification:
1) I think CanonicalIDs should be declared required instead of optional. Not in
XRI in general, but in i-names. If we tell people "here is an identifier
you can use as primary key" and "you can even verify it", then
it should always be there. I-names provided by the GRS have them anyway, but a
requirement like that would be good for community i-names. Otherwise certain
people out there may set up community i-name providers that don't have these
important CanonicalIDs :) By the way, is there some authorative text that
describes i-names and i-numbers?
2) All the talking about synonyms may affect the i-services specifications,
which also mention synonyms. These suggest (even recommend) that an i-service
can either be bound to an authority, or to individual XRIs that share this
authority. I.e. you could have a single contact page for all your synonyms, or
(thanks to the the append attribute) individual contact pages for your
identifiers, even if they are synonyms. Actually I don't think anything needs
to be changed there, just wanted to remind of that.
Markus
--
@freeXRI / freexri.com / try free i-names
On 6/22/07, =drummond.reed
< drummond.reed@cordance.net>
wrote:
Gabe, agreed. Thanks for
looking this over.
Markus, if you can review this tomorrow and send any feedback to the list, I
will commence with drafting text for ED03.
Thanks,
=Drummond
-----Original Message-----
From: Gabe Wachob [mailto: gabe.wachob@amsoft.net]
Sent: Thursday, June 21, 2007 8:36 PM
To: '=drummond.reed'; xri@lists.oasis-open.org;
xdi@lists.oasis-open.org
Subject: RE: [xri] Agenda:Joint XRI & XDI TC Telecon 10AM PT Thursday
2007-06-21
After several relatively quick reviews, nothing about the new canonical ID
verification stuff jumps out to me. Its basically the 'resolve twice'
solution proposed on the OpenID lists next to the XRI stuff. Simple is
better, even if a bit inefficient..
-Gabe
> -----Original Message-----
> From: =drummond.reed [mailto:
drummond.reed@cordance.net]
> Sent: Thursday, June 21, 2007 12:38 AM
> To: xri@lists.oasis-open.org;
xdi@lists.oasis-open.org
> Subject: [xri] Agenda:Joint XRI & XDI TC Telecon 10AM PT Thursday
2007-06-
> 21
>
> Following is the agenda for a joint unofficial telecon of the XRI and XDI
> TCs at:
>
> Date: Thursday, 21 June 2007 USA
> Time: 10:00AM - 12:00PM Pacific Time
>
> Event Description:
> Weekly unofficial joint call of the XRI and XDI Technical Committees.
>
> TO ACCESS THE AUDIO CONFERENCE:
> Dial In Number: 571-434-5750
> Conference ID: 5474
>
>
> AGENDA
>
> 1) UPDATE TO XRD SCHEMA FOR XRI RESOLUTION 2.0 WORKING DRAFT 11
>
> Gabe and Bill will (hopefully) provide an update on their progress to
> produce a revised XRD schema for XRI Resolution 2.0 Working Draft 11.
>
>
> 2) EXTENDING CANONICAL ID VERIFICATION TO HTTP(S) URIS
>
> The group of XRI Resolution editors that have been working on this issue
> will report on their progress form the two special telecons held on this
> topic over the last week. An slight update has been posted to the XRI TC
> wiki on this topic at:
>
> http://wiki.oasis-open.org/xri/XriCd02/CanonicalIdVerification
>
> Also, this proposal is now referenced by the OpenID editors in the
> following
> writeup of the OpenID recycling issue on the OpenID wiki:
>
> http://openid.net/wiki/index.php/Identifier_Recycling
>
>
> 3) OTHER ISSUES IN XRI RESOLUTION 2.O WORKING DRAFT 11
>
> Drummond will give a progress report on ED03. The current draft, ED02, is
> at:
>
> http://www.oasis-open.org/committees/download.php/24286/xri-resolution-
> v2.0-
> wd-11-ed-02.doc
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>