Re: [openpgp] Overhauling User IDs / Standardizing User Attributes (was: Re: Scoped trust (signatures))

[Wiktor Kwapisiewicz <[e-mail]>]

>>  From what I've seen Keybase is not interested in purely OpenPGP solution - they want to keep the data on their site [0].
> 
> So it’s not worthwhile solving this problem? Or are you saying that because they are doing it no one else should do it in a standard way?

I'm saying Keybase - as a company - will not do it in a standard way, 
that's all, no hidden meaning there.

> I think it’s really cool that Keybase lets you authenticate these other networking points. And it sounds like this is at least part of the problem.

Yeah, I like it too. But you know what's really cool? That there already 
*is* an I-D for that but using OpenPGP:
https://tools.ietf.org/html/draft-vb-openpgp-linked-ids-01

I highly recommend reading that, it's already implemented and working in 
OpenKeychain. It uses User Attributes.

Actually I asked for it to get the proper User Attribute ID assigned:

https://www.ietf.org/mail-archive/web/openpgp/current/msg08913.html

It did not happen.

>> I didn't agree with him back then, but after longer thought I changed my opinion - user attributes do not have any fallback mechanism - either most software supports that custom special attribute or it's practically impossible to work with them (yes, they are supported, but displayed as an opaque string [4]). And I say this as a person that added this packet "by hand" and use it on my key.
> 
> I don’t agree with him now. This is *precisely* the sort of thing that User Attributes were created to solve.

Maybe that's just me but what's the difference between User Attribute 
that has value "https://github.com/wiktor-k" and a User ID that has 
value "https://github.com/wiktor-k"? I see only one - you can read the 
User ID just fine using existing frontends and UAT will be "opaque". 
User Attribute will have ID, right.

> Sure, but it means that you are using a generic text field in ways that are hard to parse. Why not define it?

I'm not opposed to defining it (actually I'd very much welcome that!). 
But the difference between User ID that has URI format and User 
Attribute that has URI format does not affect the difficulty of parsing 
(you still need to handle broken data in both cases).

Kind regards,
Wiktor