OpenID’s Tipping Point
25hoursaday.com
25hoursaday.com
OpenID's contribution to the Single Signon Quest has been to prove beyond a doubt that nobody wants to remember a URL for their username. Yes, we all know it's the same amount of typing as an email address, but at the end of the day, people are trained to type their email repeatedly and are thus good at it (and remember it).
So please, OpenID, hurry up go away. The thing that takes your place will probably be pretty good because of you.
e: see http://eaut.org/example/ to find out if your email provider supports OpenID resolution.
If nothing else it already has two things going for it. The first is positive press and the second is a legion of fans (some of which border on sycophants) So the question is wouldn't it be better to fix the problems in this spec rather than make another. I think so. In fact I think trying to create a new spec can be a disaster.
If you need proof of why this is ture look at all the people who continue to use RSS over Atom because of a standards fight that happened years ago.
The tough part of modifying the existing implementation will be the URLs-as-Usernames thing. They seem set on that and it just isn't flying.
I'm assuming here you can't get away with the StackOverflow approach of requiring all users to use it, so you still have to manage passwords for some users, anyway.
As an end-user, I think I'm at the point where I'd prefer to use something like OpenID for new accounts. (I wasn't using it for HN, and I lost my password about a month ago, so now I have this account linked to gmail.)
When you have tools like this, seems worthwhile to do so.
But OpenID does have its issues. Obviously trusting one provider with all of your username and password credentials isn't the best idea. I think, however, that having an OpenID URL is a good idea. But the better idea is to give new users their own unique ID on your site, and to allow them the option to link it to some other previously established online ID. Not by default.
In fact, information auditing doesn't really work without the rule of three: which (at some level) includes the separation of power from control.