Recently, all kinds of articles about Digital Identity have been circulating. In this blog, I plan to jot down whatever comes to mind, ranging from meta-level questions such as what exactly Digital Identity requires to matters of implementation.
For this first installment, I will start with some extremely high-level requirements.
- It must be possible to use identifiers to point not only to people and things but also to their attributes, and to access metadata and data in a RESTful manner.
- Access control must be possible when doing so.
- Access control must be established as a “contract” that is legally meaningful under contract law. (This makes transactions across national borders possible as well.)
- A request for data must present a proposed contract, with the data becoming available once the contract is concluded.
- It must be possible to establish rules in advance so that contracts can be concluded automatically.
- If a contract cannot be concluded automatically, it must be possible to present the contract request to the individual.
- The contract request must be presented in a form that individuals can read and understand.
- Contracts must be applicable not only to individuals and legal entities but also to groups. (multi-party contract)
- Pseudonym-based information exchange must be possible so that individuals do not need to be identified unless absolutely necessary.
- Conversely, when real-name identity is required, the parties must be able to point to one another using persistent, unique identifiers.
- It must also be possible to access historical data associated with those identifiers.
- Third-party certification must be available for each individual, legal entity, and thing.
- Third-party audits and certification must be possible for the operational level of each identity; that information must be available and usable for automatically concluding contracts.
- A Reputation Service must be available for each identity and usable for automatically concluding contracts.
Technical Requirements
- The number of identities will be enormous. The database must therefore be distributed.
- Sharing a persistent identifier may pose a privacy risk, so it should be avoided whenever possible. A service is needed that accepts a pseudonym and verifies its existence and trustworthiness.
- Identifiers must be internationalized.
- Identifiers must be URIs.
- The data obtained must be enclosed in a consistent format that does not require schema extensions.
- For this purpose, abstract identifiers must be established as a vocabulary.
For now, that is what immediately comes to mind while I am feverish.
Incidentally, based on these ideas, as well as the slides from my presentation at Oasis Open Day (2003/11/20) and my talk at ACIMOD (2004/8), I am posting the file I prepared for a WSI study session.
wsi-xdi“XDI/XRI Primer” (materials for a WS-I study session, 2004/9/30)
Related posts

What Do Unlinkability and De-identification Mean? (1)
These matters have been debated for a long time, but recently there has again been increased activity around me aimed at clarifying the meaning of terms…
The Haraguchi 5 Principles and OpenID
The “Haraguchi 5 Principles on Identification Numbers” have been announced. Principle 1: The system must protect the rights of citizens The number is intended to protect…

Why Identity Federation Helps Improve Privacy
I received an inquiry from a certain party and sent a response, so I am sharing it here. The topic is “Why Identity Federation Helps Improve…
