[OpenID] Wiki page: Attempting to document the "Email Address as OpenId" debate.

[Claus Färber]

David Fuelling schrieb:
>> From: general-bounces at openid.net [mailto:general-bounces at openid.net] On
>> Behalf Of Claus Färber
>> I'd prefer to call them user at REALM OpenIDs. The concept of using this
>> format is not only used for email but also for other types of
>> identifiers such as RADIUS zones or Kerberos realms.
>>
>> For example, user at EXAMPLE.COM could be a RADIUS login, a Kerberos
>> principal and a Yadis ID/OpenID but not an email address.
> 
> Hmmm....I'm not sure about this one.  I intentionally rooted my definition
> of "Email Address" in RFC2822, which is where it should be rooted (if we're
> going to call them 'email addresses').  

I'm repeating myself here:

| I'd prefer to call them user at REALM OpenIDs.

> Seems like anything in the form 
> 'user at domain' would also have an identical root (RFC2822).

I don't know where this format comes from. RFC 733 gives a somewhat 
vague origin in ARPAnet mail usage.
It's certainly not RFC 2822 (or 822), which is a rather new standard.

It's also unclear whether the usage in other protocols imitated the 
email use or whether it was a parallel invention.

> Is that not the case for Kerberos realms and RADIUS zones?  In other words,
> are these *not* also email addresses? 

| For example, user at EXAMPLE.COM could be a RADIUS login, a Kerberos
| principal and a Yadis ID/OpenID but not an email address.

There's no technical reason an identifier using the user at DOMAIN format 
has to be a valid email address.

Claus