[uf-discuss] Re: XFN for email addresses?

[Chris Messina]

I suppose this is where graph theory has to come in to some degree and
afford immediate lapses as such, for, in the example you described, if
somewhere there exists a reference, say on theryanking.com, that
points to that email address using rel-me, we'll be able to apply some
kind logic that would, in the least, be able to say that that URL has
laid an unverified claim to that email address.

Interestingly, this is, to some degree, the problem that MicroID was
trying to solve, albeit with hidden meta tags.

I agree that the lack of transitivity with email "endpoints" make them
very weak from a claims perspective, and I agree that we can do
better, but given that Technorati and the like still require a
verified email address to open an account, I don't see this behavior
going away anytime soon.

One goal of mine to develop a replacement for those "add friends from
your address book" widgets, which are so seductive and therefore so
dangerous. People are being trained to enter their email and password
on almost every new social network; in the case of Gmail, that
username/password combo access far more than just your mail (think:
Google Checkout, web history, etc). This is extremely dangerous. If we
could instead train people to type in a URL-pointer to their list of
friends, authenticate remotely, and then pull down the list of
contacts, including email addresses as necessary, we'd have a much
safer social web.

Chris


On 6/13/07, Ryan King <ryan at technorati.com> wrote:
> On Jun 13, 2007, at 4:25 PM, Chris Messina wrote:
> > On 6/13/07, Ryan King <ryan at technorati.com> wrote:
> >> On Jun 12, 2007, at 8:15 PM, Chris Messina wrote:
> >>
> >> > Clearly the biggest issue I see with this scheme is the
> >> inability to
> >> > link out *from* the email address. However, I'm not sure that this
> >> > case nullifies the utility of such links.
> >>
> >> And this is a quite large issue. It effectively removes the N in XFN.
> >> It's hard to build a network where some nodes can only by sinks[1].
> >>
> >> -ryan
> >>
> >> 1. http://en.wikipedia.org/wiki/Degree_(graph_theory)#Sink
> >
> > [Trying again....]
> >
> > And while that is a valid concern to be noted, I don't think that it
> > precludes using non-URL based indicators as person identifiers.
>
> Just be clear, I believe that by "non-URL" you mean "non-HTTP-URL".
>
> > I mean, I'm a huge proponent of OpenID and URLs for identity; at the
> > same time, that vision doesn't match with reality and I don't think
> > it's going to change immediately; therefore, XFN should not be able to
> > deal with the established convention even while things are (hopefully)
> > moving in the direction of URL-based identifiers.
> >
> > In any case, consider mailing list subscriptions -- the fact that a
> > message is sent to the identifier (aka your email address) and you
> > respond by clicking a confirmation link proving that you can receive
> > messages sent to that email address is no different than Technorati's
> > current Javascript-based URL-claiming mechanism, whereby, instead of
> > clicking a link, you insert a script that Technorati expects to find
> > at the destination URL. The same argument can be made for IM links,
> > since a bot could ping you and ask you for some data that only you
> > would know, and if you're able to respond accurately, then you've
> > similarly proven "ownership" of that destination.
>
> Actually they are different in an important way. The code we give
> people to put on their blogs includes rel="me" and when blog claims
> are successfully completed, we link back to the blog with rel="me" [1].
>
> So, these assertions are public and can be indexed by anyone.
>
>
> To step back a bit, I see nothing wrong with trying to use email
> addresses to connect people's identities. It seems to be a cowpath
> that's already paved. However, I don't think it fits very cleanly
> into XFN.
>
> For example, you can do:
>
>   <a rel="met" href="mailto:ryan at theryanking.com">ryan king</a>
>
> Which is a clear, unambigious assertion, there's no way to connect
> the URL mailto:ryan at theryanking.com to the rest of the XFN network
> via identity reconciliation and without that, we lack a lot of utility.
>
> Of course, I don't really have a better suggestion right now. :) But
> I think we can find a better way to do this.
>
> -ryan
>
> 1. We know this doesn't work for multi-author blogs, we're working
> around that.
> _______________________________________________
> microformats-discuss mailing list
> microformats-discuss at microformats.org
> http://microformats.org/mailman/listinfo/microformats-discuss
>


-- 
Chris Messina
Citizen Provocateur &
  Open Source Advocate-at-Large
Work: http://citizenagency.com
Blog: http://factoryjoe.com/blog
Cell: 412 225-1051
Skype: factoryjoe
This email is:   [ ] bloggable    [X] ask first   [ ] private