[Atp] Re: Adding Identity/Auth to Charter

[Ted Hardie <[e-mail]>]

Hi Bryan,

I see the logic in this,  but I continue to have a concern that specifying
this by reference to DIDs as a general approach doesn't necessarily result
in interoperability for any specific system built with the documents that
are within scope this working group.  As you note, DIDs are a framework,
and two parties need to share specific DID systems to actually move forward
with anything that requires the identity information.  As a result, I think
that defining it only in terms of an abstraction boundary is as risky as
mentioning a specific set of DIDs.

We may be able to square that circle by saying that systems built using
these tools should expect to make a small set of selections from within the
space of DIDs, in order to foster interoperability.  Within the charter, we
could include a deliverable document that describes how folks building on
top of ATP make that choice (sort of meta-applicability statement).  Part
of that could be a short description of how things work in the current *sky
deployments and why those approaches were taken.

This is a bit of extra work, but I think it substantially improves the
chance that we'll get interoperable systems based on this in the end.

regards,

Ted Hardie
(No hats on)

On Wed, Dec 3, 2025 at 1:54 AM Bryan Newbold <bryan=
[e-mail]> wrote:

> We've received feedback from a number of folks that the initial charter
> scope should cover more about network identity to be meaningfully
> interoperable. I want to explore that direction a bit to see what it feels
> like.
>
> Instead of abstractly describing the requirements for a persistent account
> identifier, we could just say that W3C DIDs are the framework: DIDs are the
> identifiers, and DID documents contain the required account data (keys,
> identifiers, service locations). The working group would either directly
> specify the data required to interoperate with AT, or liaise with another
> venue if that was more appropriate.
>
> I think specifying an exact set of DID methods in a formal IETF
> specification would be hard, because most of the relevant methods have not
> (yet) been formally standardized. did:plc is open to standardization, but
> has not been submitted to the W3C or any other venue. It could take time to
> get those gears turning, on a timeline outside of this group's control. The
> ATP working group could publish a less formal BCP describing current
> recommendations ("use did:plc and did:web")? Or find some other way to
> discuss and nudge in this direction. Alternatively we could just say that
> DIDs are the abstraction boundary for the scope of the working group, and
> leave it at that for now.
>
> Handles could also be added to the initial charter. The current AT handle
> system works with either DNS TXT records or HTTPS .well-known endpoint. I
> think this work could mostly be done within the WG, though we could also
> liaise and coordinate with DNSOP or other groups as needed. I am already
> co-author on a relevant informational-track IETF draft in DNSOP about the
> AT use-case [0], and there are some other relevant drafts under discussion.
>
> --bryan
>
> [0]: https://datatracker.ietf.org/doc/draft-ietf-dnsop-integration/
> _______________________________________________
> Atp mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]
>