Digital Identity

Internet Identity

PingID: The term 'digital identity' typically refers to our electronic personal information such as name, address, phone numbers, demographics etc. However, in the new world of network computing, our digital identity promises to become much more, in fact, it strives to become the basis of our "digital-self" and the keeper of not only our most personal electronic information (credit cards, biometrics, personal health records) but the key which grants us permission to communicate, transact, and create trusted relationships with people, businesses, and devices. Changes within the computer industry are bringing about this change as we speak, and it will not be long before lacking a digital identity, we will no longer be allowed access to information, communications services or electronic transactions capabilities.

See Passport, Reputation Management, Distributed Membership System, PingID (based on Jabber?)

Some see "identity" as more about the various ways you communicate online, and how people find you. Personal domains, EMail address, Instant Messaging ID, WebLog URL, etc. I suggested that maybe a personal directory system similar to DNS, which would then provide current protocol-specific addresses for client apps. Maybe what we need is something like DNS, but for people. A distributed directory system where each user could control his data, and which would return protocol-specific addresses. Then you could assemble services from multiple vendors, and move them at will - you'd just update your identity record. You'd probably need client apps to then support this directory, so, for instance, instead of storing an EMail address your email client address book would store a personal identity address, then look up your email address when needed. Likewise, your Instant Messaging client would look up the current IM addresses for your buddy list on startup. (I wonder whether new fields can be added to a VCard, or whether standards-compliance limits you to a set list of fields?)

  • Related to this, I noted Yes, one of the nasty issues is defining unique IDs. Bob Frankston argues (in the context of DNS and name squatting) that...

    • IDs should not have any semantic content, else you get arguing over who should own a given ID (so use random numbers)

    • it's impossible to guarantee uniqueness of any usable query parameters (e.g. person name), so you need a directory that (a) includes some non-structured content (e.g. you could list a long set of keywords or freetext), and (b) allows flexible querying and (c) returns multiple records so you can pick the record that looks right. But you would in theory need to look up a given person only once (ever) and then save the appropriate ID in your address book.

    • Clay Shirky disagrees with some of this... though his context is domains rather than user names)

      • Zooko has an earlier similar essay
  • Does RDFWeb and FOAF have a role? Probably not, given that the Yahoo group only gets 10-20 msgs per month...

Edited: |

blog comments powered by Disqus