XRID proposal for YADIS - Opinions anybody?

[Drummond Reed]

Just to clarify one thing (which both Ben and Lukas asked about):

In this proposal, there is NO requirement or connection to using XRIs or XRI
resolution to address the YADIS document. You can use OpenID or LID or
another other URL you want to retreive it.

The point of synergy is that all of these would point to the same type of
discovery document in the same format. This is the reason the idea of using
XRID as the service/capabilities description format came up is that when
Johannes & David first brought up YADIS with us. We immediately saw that
YADIS service discovery step was exactly like the same step in XRI
resolution that retreives an XRID. This is especially true since an XRI gets
turned into an HTTP URL during XRI resolution.

We then realized that there was no need to start with an XRI to request an
XRID -- you could request it with any URL assigned to it, so OpenID and LID
URLs would work perfectly.

And since XRIDs need the same simple, generalized service description format
as YADIS, that's when we realized the potential synergy of using the same
description format for both YADIS and XRI resolution.

The XRI TC was very receptive to this on the calls we had last week, so
that's why we're going down this road. But again, there is no requirement
whatsoever to use XRIs to retrieve an XRID/YADIS document. The synergy
really comes down to:

OpenID URL--------------->|
LID URL------------------>|
Any other identity URL--->|-------> YADIS discovery document/XRID
XRI---------------------->|
Any other discovery URI-->|

=Drummond 

-----Original Message-----
From: yadis-bounces at lists.danga.com [mailto:yadis-bounces at lists.danga.com]
On Behalf Of Benjamin Yu
Sent: Monday, November 07, 2005 12:46 PM
To: OpenID Discussion
Subject: Re: XRID proposal for YADIS - Opinions anybody?

My initial thoughts:

1. I'll need some more time to dig into the spec.

2. From what I initially see, the Resolution doc is a protocol for
transforming
an XRI's authority part into a hard URL. From which we can then examine that
final XRI Descriptor xml file for the Service References (and other meta
data).

3. If we are just saying that the XML Descriptor format is to be used as the
serialized format (view/representation) of the services, then I really don't
see any difficulties with this part. This view/representation is pretty much
as
good as any others. We could use the Accept HTTP header to get preferred
representations. YADIS could define that set of acceptable representations.

4. But if consumers are forced to support the intermediate XRI resolution
steps
of determining the final XRI Descriptor authority, then I feel that it may
be
overloading YADIS. 

4a. I personally don't want to write up code to deal with the xri:*. The
only
thing I want in my code is a GET request to my http client. I think that XRI
support in this manner needs to be left up to the XRI protocol library that
my
http client will use.

4b. And for the most part, I think that people will just be using POHU
(Plain
Old HTTP URIs) at this point.

5. If YADIS requires a more complicated service discovery mechanism, we
should
also research other specs and papers that have been publish about service
discovery. Then weigh what we have learned.

"An Architecture for a Secure Service Discovery Service"
http://iceberg.cs.berkeley.edu/papers/Czerwin-Mobicom99/

"Really Simple Discovery"
http://media-cyber.law.harvard.edu/blogs/gems/tech/rsd.html

"Service Location Protocol, Version 2"
http://www.ietf.org/rfc/rfc2608.txt

"A DNS RR for specifying the location of services (DNS SRV)"
http://www.ietf.org/rfc/rfc2782.txt

"Remote Service Discovery in the Service Location Protocol (SLP) via DNS
SRV"
http://www.ietf.org/rfc/rfc3832.txt

-Ben

--- Ernst Johannes <jernst+lists.danga.com at netmesh.us> wrote:

> Is nobody saying anything because it's so obviously good that it  
> doesn't need to be commented on, or so obviously bad that it doesn't  
> need to be commented on, or ...? ;-)
> 
>  From my perspective, it looks fine so far, and if we can "reuse" a  
> standard instead of inventing a new one, I would consider this to be  
> A Good Thing.
> 
> On Nov 6, 2005, at 11:31, David Recordon wrote:
> 
> > Drummond,
> > Thanks for getting this posted!  Hopefully people can start digging  
> > into it the next few days and we all can have another YADIS spec  
> > proposal not long after.
> >
> > --David
> 
> Johannes Ernst
> >   http://netmesh.info/jernst
> 
> 
> 
> 



Blog: http://www.badpopcorn.com/
Homepage: http://www.foofiles.com/u/byu/