Markus wrote:
> (Does Ootao/Wingaa/Ezibroker support this too? Or does every XRI
always have their own contact page,
> even if they are local synonyms?)
The solution we use is for the RP to use CanonicalIDs in the role as
primary key. This way all synonyms will work, including polyarchical synonyms
(such as @oatao*steven being synonymous to =steven.churchill w.r.t a given
service type.) We do CID verification “by hand” currently.
See http://dev.inames.net/wiki/XRI_CanonicalID_Verification.
~ Steve
From:
markus.sabadello@gmail.com [mailto:markus.sabadello@gmail.com] On Behalf Of Markus Sabadello
Sent: Monday, December 10, 2007
2:04 PM
To: xrdsp@inames.net; OASIS XRI TC
Subject: [xri] i-services must
perform resolution?
Think about this:
The ISS (specifications for authentication/contact/forwarding i-services)
mention that i-services can be bound either to a QXRI or to the authority
associated with it. If you bind it to the authority, that means that the
i-service ( e.g. the contact page) is the same for all XRIs that share the
authority.
So let's assume I have two i-names which are local synonyms (same CanonicalID):
=markus
=markus.sabadello
And I want only a single contact page that is shared by both i-names.
Now if the i-service provider is the same entity as the ibroker, then this is
no problem. For example when the @freeXRI contact page i-service receives a
request, I simply look in the OpenXRI store to find out if there's a contact
page associated with the authority of the XRI. If not, I check if there is one
associated with the XRI itself. When you create an i-service, there's actually
a checkbox that lets you choose whether you want it just for one i-name, or
shared by its synonyms.
(Does Ootao/Wingaa/Ezibroker support this too? Or does every XRI always have
their own contact page, even if they are local synonyms?)
But what happens when I sign up for a busy time service (using XRDSP), and again
I want it to be shared by my two i-names? The only way I see to make this work
is that the service provider must resolve the XRIs and look at the CanonicalID
to find out if they are synonyms. To be perfectly correct, it would even have
to take into account <CanonicalEquivID>s and <EquivID>s.
Does that sound correct?
Markus