[OT] Re: User @ domain.tld as ID (Once again)

Martin Atkins mart at degeneration.co.uk
Thu Nov 3 08:01:21 PST 2005


Mike Hearn wrote:
> 
> I think people are getting mixed up here - the centralized mapping service
> was proposed so that AOL/Hotmail/whatever users can use their email
> address as their ID until such a day that those services support OpenID
> natively. 
> 
> It's mostly separate to the issue of how to implement email address
> identifiers on top of the existing OpenID protocol (which isn't hard).
> There's no centralisation involved with mapping user at host to
> http://host/openid/user
> 

I think I have a way to resolve this debate.

This assumes that we have to end up at a URL anyway. This is a valid 
assumption, since YADIS is HTTP-based. Some convention for mapping from 
an email address to a URL does not affect the underlying YADIS at all.

So my proposal is this: YADIS will go ahead and use URLs, because it has 
to as a fundamental requirement. Another spec can try to standardize 
mapping an email address onto a URL, which will then automatically 
enable YADIS to use email addresses. There's no reason why these two 
things have to be bound together into one spec.

This is good layered design, and it also allows "the market" to decide 
what identifiers people want to use. If consumers go ahead and implement 
YADIS based on URLs, then we'll know that no-one really cares about the 
email stuff. On the other hand, if the email mapping thing proves 
popular then YADIS can quite happily live at the layer below.

It also means that we can stop bickering about what kind of identifiers 
we're going to use and get some real work done.



More information about the yadis mailing list