Formalization of this spec?

[David Fuelling]

Hey List,

I'm wondering how the idea of webfinger is going to go from "good
idea" to actual spec. My question encompasses two relevent issues:

1.) Which Standards Body
+ Is the intention to move this through a formal standards body, and
if so, which one? OASIS? IETF? OpenID Foundation? OWF (I realize
this isn't really a standards body).
+ Might it be possible to push this work into the XRD 1.0 TC somehow?
That way, it could be a formalized mechanism, but without having to
jump through all the hoops to get this into a standards body.

2.) How will the ideas of webfinger get formalized? Is there a plan to
draft a formal-looking spec? Has one been drafted already?
I would be open to helping draft/edit a spec. I'd like to align
whatever happens here with EAUT. They're pretty close, but slightly
different enough.

Thanks for any input!

David

[Santosh Rajan]

What we need is a "Foundation for Federated Login of email like
identifiers".

[Eran Hammer-Lahav]

On 6/8/09 8:02 AM, "David Fuelling" <sapp...@gmail.com> wrote:

>
>
> Hey List,
>
> I'm wondering how the idea of webfinger is going to go from "good
> idea" to actual spec. My question encompasses two relevent issues:
>
> 1.) Which Standards Body
> + Is the intention to move this through a formal standards body, and
> if so, which one? OASIS? IETF? OpenID Foundation? OWF (I realize
> this isn't really a standards body).
> + Might it be possible to push this work into the XRD 1.0 TC somehow?
> That way, it could be a formalized mechanism, but without having to
> jump through all the hoops to get this into a standards body.

None. XRD will be an OASIS standard, /.well-known and LRDD will be IETF
documents (maybe standards), and the rest is just normal web infrastructure.
All we need is a white paper on how to use these protocols together. I don't
think this should be an actual normative specification.

> 2.) How will the ideas of webfinger get formalized? Is there a plan to
> draft a formal-looking spec? Has one been drafted already?
> I would be open to helping draft/edit a spec. I'd like to align
> whatever happens here with EAUT. They're pretty close, but slightly
> different enough.

Most of the work is making sure the above mentioned specs support the use
case. We already started simplifying them as an outcome of this work.

EHL

[John Panzer]

On Mon, Jun 8, 2009 at 9:22 AM, Eran Hammer-Lahav <er...@hueniverse.com> wrote:

On 6/8/09 8:02 AM, "David Fuelling" <sapp...@gmail.com> wrote:

>
>
> Hey List,
>
> I'm wondering how the idea of webfinger is going to go from "good
> idea" to actual spec.  My question encompasses two relevent issues:
>
> 1.) Which Standards Body
> + Is the intention to move this through a formal standards body, and
> if so, which one?  OASIS? IETF?  OpenID Foundation?  OWF (I realize
> this isn't really a standards body).
> + Might it be possible to push this work into the XRD 1.0 TC somehow?
> That way, it could be a formalized mechanism, but without having to
> jump through all the hoops to get this into a standards body.

None. XRD will be an OASIS standard, /.well-known and LRDD will be IETF
documents (maybe standards), and the rest is just normal web infrastructure.
All we need is a white paper on how to use these protocols together. I don't
think this should be an actual normative specification.

+1.  IMHO the largest webfinger's value is in pulling together the various specifications and demonstrating ("running code") the integration between them.

 

> 2.) How will the ideas of webfinger get formalized? Is there a plan to
> draft a formal-looking spec?  Has one been drafted already?
> I would be open to helping draft/edit a spec.  I'd like to align
> whatever happens here with EAUT.  They're pretty close, but slightly
> different enough.

Most of the work is making sure the above mentioned specs support the use
case. We already started simplifying them as an outcome of this work.

And this is the other part of webfinger's value add :)

[David Fuelling]

Thanks Eran, that makes sense. 

Given that webfinger may just be a whitepaper (best practice type of thing), how do you see webfinger-like functionality getting used or "formalized" by other specs?  Would you be open to other specs "normalizing" webfinger-type behavior?

I'm thinking specifically about OpenID Discovery 2.1 here.  For example, are you of the opinion that OID Discovery 2.1 should _not_ create a normalized procedure for performing discovery on an email address (e.g.)?  Or are you instead just saying you don't think webfinger itself needs to be normalized?

Thanks!

david

[Eran Hammer-Lahav]

If OpenID ends up using this exact mechanism, it needs to describe it. BUT, the whole point is that once we get the other specs done, webfinger will be REALLY SHORT.

EHL