<HTML>
<HEAD>
<TITLE>Re: [OpenID] Musing on FaceBook, OpenID and the next mountain to climb</TITLE>
</HEAD>
<BODY>
<FONT FACE="Calibri, Verdana, Helvetica, Arial"><SPAN STYLE='font-size:11pt'>My personal decision whether to use Facebook or not is not based on how open their systems are. So I don’t really care if they choose to open and what they feel is in the best interest of their users. So far they have been extremely successful and so I assume they are doing something right and are offering a valuable service to their millions of active users. They also have a very sharp team. My comments were not about Facebook being open. It is a business decision that I have no special insight into.<BR>
<BR>
My comments were specific about the technical details of their platform. The fact that instead of using a community specification they decided to create yet another protocol is what I was referring to. I would give them a high score if they used OAuth but still kept their system closed and data private. They could have created the exact same product by building it on top of OAuth. They could have also made it friendlier to OpenID or even use OpenID as the basis but I think OAuth would have been an easier match and lighter on the politics.<BR>
<BR>
EHL<BR>
<BR>
<BR>
On 8/2/08 11:22 AM, "Peter Williams" <<a href="pwilliams@rapattoni.com">pwilliams@rapattoni.com</a>> wrote:<BR>
<BR>
</SPAN></FONT><BLOCKQUOTE><FONT FACE="Calibri, Verdana, Helvetica, Arial"><SPAN STYLE='font-size:11pt'>Folk adopt open systems when their customers will no longer tolerate the the boundaries that come with the feature benefits that often come with proprietary design.<BR>
<BR>
This lesson was taught to be much a senior architect at lotus, which had spent ~10 years avoiding using x509 certs in favor of a lotus format - mostly in order to help it differentiate/compete with novell and microsoft in the early lan middleware market. The troubles intelligent folk even today with x509 config (in the shib community, say) simply didn't exist in the highly tuned lotus model - and I've no doubt that their customers would have rejected the whole idea of certs (and the secure lan email tha resulted) in those days, if it had been that open and difficult to manage.<BR>
<BR>
________________________________<BR>
From: Eran Hammer-Lahav <<a href="eran@hueniverse.com">eran@hueniverse.com</a>><BR>
Sent: Friday, August 01, 2008 8:11 PM<BR>
To: OpenID <<a href="general@openid.net">general@openid.net</a>><BR>
Subject: Re: [OpenID] Musing on FaceBook, OpenID and the next mountain to climb<BR>
<BR>
A few facts:<BR>
<BR>
Facebook Connect could have been built on top of OAuth.<BR>
Facebook did not participate in the Open Web Foundation launch – Dave Morin was involved as an individual.<BR>
Facebook has been talking about their desire to open and learn more about open specs for a year now, with nothing to show for it.<BR>
Facebook has been invited and engaged in conversations with the community with nothing but a waste of time to show for it.<BR>
<BR>
---<BR>
<BR>
The fact that on the same day they announce support for the OWF, they also announce a product that is ignoring all the work done by this very same community they claim to be supportive off, is to me, a mockery. I would be happy to be proven wrong but for a year now seen nothing to make me believe it.<BR>
<BR>
EHL<BR>
<BR>
<BR>
<BR>
<BR>
From: <a href="general-bounces@openid.net">general-bounces@openid.net</a> [<a href="mailto:general-bounces@openid.net">mailto:general-bounces@openid.net</a>] On Behalf Of Dick Hardt<BR>
Sent: Friday, August 01, 2008 3:40 PM<BR>
To: Paul Trevithick<BR>
Cc: <a href="david@sixapart.com">david@sixapart.com</a>; OpenID<BR>
Subject: Re: [OpenID] Musing on FaceBook, OpenID and the next mountain to climb<BR>
<BR>
Hi Paul<BR>
<BR>
While Facebook could take the silo approach, they are interested in seeing how open standards could be used. They participated in the Open Web Foundation launch and when I was at their office earlier this week, they expressed serious interest in OpenID. See my blog post (which had to be run by them as it was an NDA meeting).<BR>
<BR>
<a href="http://identity20.com/?p=155">http://identity20.com/?p=155</a><BR>
<BR>
Given the state of OpenID tech right now, I do not think it could be used to solve what they wanted to solve in a way that would deliver the clean user experience they desired -- but I would be happily proved wrong! ( I do think they could have used OAuth though)<BR>
<BR>
As I mention in my post, this is an opportunity for the community to work with Facebook.<BR>
<BR>
Myself, I think the technology needs to be enhanced and evolved so that it has features that Facebook Connect does not have in addition to the existing features.<BR>
<BR>
If the community just sits back and says that all the bits are there -- just use them -- then this community is no different from other SSO communities that have told the creators of OpenID that they were reinventing the wheel.<BR>
<BR>
-- Dick<BR>
<BR>
On 1-Aug-08, at 2:09 PM, Paul Trevithick wrote:<BR>
<BR>
<BR>
The problem is that this isn’t a technical issue. FB currently has no business incentive to use open technologies that, among many other things, would allow users to be able to retrieve and store their own profile data and friends lists (as currently violates the FB TOS). They are still enjoying the virtuous cycle of the closed mega silos: more users begets more users. OTOH FB will open up if and when there’s a reason to do so. But for now, and for a good while, I’d say FB isn’t a good prospect for open, user-centric technologies.<BR>
<BR>
Paul<BR>
<BR>
On 8/1/08 3:28 PM, "Allen Tom" <<a href="atom@yahoo-inc.com">atom@yahoo-inc.com</a>> wrote:<BR>
<BR>
<BR>
David Recordon wrote:<BR>
> Is there really anything that Facebook did that couldn't be<BR>
> accomplished with OpenID Authentication 2.0 and OpenID Attribute<BR>
> Exchange?<BR>
Facebook Connect has a nice set of libraries/apis that RPs can just drop<BR>
in relatively easily on their site. The JS libraries implement much of<BR>
the sign in flow (displaying inline sign-in forms as well as a<BR>
permissions screen) which means that the FB Connect user experience is<BR>
consistent across all RPs.<BR>
<BR>
They also seem to have implemented Single Sign Out, because signing out<BR>
of FB seems to also sign you out of the RP.<BR>
<BR>
Additionally, FB Connect also authorizes the RP to write to the user's<BR>
FB News Feed, so there's an authorization component as well. The<BR>
authorization seems to expire when the browser session is closed, so<BR>
it's not quite like OAuth.<BR>
<BR>
And finally, FB Connect requires that the RP pre-register with FB to get<BR>
an api key which presumably allows FB to authenticate the RP, and also<BR>
gives FB the ability block the RP if necessary.<BR>
<BR>
Unlike the OpenID/OAuth/AX services currently in the wild, the FB<BR>
Connect stack is highly integrated, with built in privacy controls and a<BR>
standard UI. But as you correctly stated, I believe most, if not all, of<BR>
the stack could have been built upon open standards.<BR>
<BR>
Allen<BR>
<BR>
<BR>
<BR>
_______________________________________________<BR>
general mailing list<BR>
<a href="general@openid.net">general@openid.net</a><BR>
<a href="http://openid.net/mailman/listinfo/general">http://openid.net/mailman/listinfo/general</a><BR>
_______________________________________________<BR>
general mailing list<BR>
<a href="general@openid.net<mailto:general">general@openid.net<mailto:general</a>@openid.net><BR>
<a href="http://openid.net/mailman/listinfo/general">http://openid.net/mailman/listinfo/general</a><BR>
<BR>
<BR>
</SPAN></FONT></BLOCKQUOTE>
</BODY>
</HTML>