A technology works in theory if you have a piece of software and you can make it do what you want it to do for the purposes of demonstrating it and letting others try it out.
It works in practice if people who utterly don’t care about your technology successfully use it because it makes life easier for them.
Plaxo and Google yesterday reported that they took 1000 random Google users who were interested in signing up at Plaxo, a combination of OpenID, OAuth, and a user interface targeted to this particular situation, and measured how many users would actually manage to sign up via OpenID instead of username and password. No prior OpenID or other knowledge was required. It turns out that 92 percent of those random users managed to sign up a Plaxo.
Does your website manage to register 92 percent of the people who want to, with the supposedly simpler username and password? Didn’t think so either.
That is very impressive. Congratulations to Joseph and Eric and their teams.
It clearly shows where the OpenID user experience needs to go. (Nothing particularly hard there, but a bit against conventional wisdom so far.)
Even more importantly, it proves in practice that the theoretical benefit of "easier sign-up for users" is actually real and can be measured, and beats the state of the art. If your competitor manages to sign up 92 percent of potential customers, and your site remains at, say, 50 percent — do you really think you can survive that for long?