So Mike Neuenschwander hung a softball out there with his latest post on becoming an OpenID power user. Dave Kearns was quick to take a swing at it with his response to Mike’s summarization: “There are no identifiers, only attributes.”
Mike’s journey to OpenID begins with a single step – getting an OpenID, which is really an exercise in picking a name. Names are important. (I am going to stop myself from going into a discussion of the gravity of names and naming. Literature is soaked in naming issues.) As Mike points out he can pick any unused name (really, any set of unused characters.) The first person in to register ian.glazer.myopenid.com can purport to being Ian Glazer. This is no different than XRI name registration or domain registration or copyright registration… you get the idea.
Dave goes from there and reminds us that identifiers have to be unique within a given namespace. He uses the example of disambiguating family members. He provides one of the most familiar examples on unique identifiers:
Your email address – every single one of them – is a unique identifier within the entire world of the internet.
What is hidden in Dave’s comments is the role of context. Given the context of family, Dave’s non-unique identifier can be disambiguated. We use the domain name in an email address to set context. I know that an email coming from mike@burton is likely to be of a professional nature and an email coming from mike@igotsmesomefreeemail is likely to not be. The context of how you use your identifier is meaningful.
Thinking out loud here… I wonder if the visual metaphors in CardSpace will help set context for both the relying party and end-user. Presenting context in a way that is meaningful to the end user could help solve a few other problems, notably phishing sites.