Call to Participate: Help improve Webfinger RFC with its lead author

@melvincarvalho made me aware of his mail exchange with Paul E. Jones, the lead author of the Webfinger RFC at IETF. Paul was about to close the maiing list, since it wasn’t used for 5 years.

When Melvin mentioned SocialHub and the many discussions here about Webfinger, Paul mentioned being open to discuss possible updates… don’t take it from my word, but read:

As the original lead author of the RFC, I’m certainly willing to work on
needed improvements. Oddly, though, nobody has ever reached out to me
directly about issues. I am aware of its use in various contexts
(OpenID and Metaverse being the two prominent). I have seen some
complaints about challenges with static sites, though most web servers
can be configured to deliver static WebFinger content. I assume what is
wanted a URI in this form:

/.well-known/webfinger/acct:paulej@packetizer.com

Anyway, if there is a desire to make improvements, I think we should
exercise caution so as to not break an otherwise trivial protocol.

Paul

And the thread at:

CC @codenamedmitri

5 Likes

@melvincarvalho getting asked by @jfietkau how to subscribe to the mailing list. It is not really clear, and IETF Search yields zero results when searching for WebFinger.

I asked Julian to respond here, if not finding a way to post to the list. And maybe you can convey the interest then?

1 Like

To be clear, I have no major issues with the general state of WebFinger and no personal need for changes to its request URI syntax.

My only problem is that RFC 7033 (or 7565?) would do well to include clear guidance on whether WebFinger usernames in acct: URIs are case-sensitive or case-insensitive. My reading of the standards involved is that they are case-sensitive in principle, but that every popular ActivityPub implementation maps WebFinger requests to actor URIs in a case-insensitive fashion (as they are free to do). However, I saw one particular implementation go out of its way to assume case-insensitivity on the client side, which struck me as a potential issue:

This seems like something the specs could clarify without causing additional client complexity.

2 Likes

@trwnh is working on the “ActivityPub and WebFinger” report. In my opinion it is pretty good, and might be relevant to this discussion:

https://swicg.github.io/activitypub-webfinger/

2 Likes