Just because you can map a tuple of name, address, photograph, etc., to a randomly generated number, this by itself does not make the ID unique.

1. First you need to come out with a SUFFICIENT number of attributes to identify a human being uniquely (and be able to prove it).

2. Next you need to compare ALL of these attributes to the respective attributes of ALL humans currently in the UID data base.

3. Only IF no match is found, THEN you can assign an unused random number to the human, and add the attributes to create the next uniquely identified human.

4. Imagine the data base growing from zero population to the population of India over months / years. We are talking of hundreds of quadrillion matches. (Any average mathematical person can give you a precise number if you tell him the number of attributes and the relevant attributes of the attributes).

5. The ongoing function of new enrollments, changes and de-enrollments, which needs to continue for ever, will need matching with the entire population base. In order to do this online (in real time?), it not just a question of having a supercomputer (for none will do). It is a question of algorithm.

So let’s stop wasting time on trivialties. So far I have heard no one, from top to bottom to the talking heads on TV, even understand and acknowledge the real issues.

Comments, anyone? (I will be on long flights and ships in Sept., so may not respond to your posts.)

Identity is established by:

1. What you have (e. g., a card), the WEAKEST way to identify anyone,

2. What you know (e. g., a password), somewhat better, but not fool-proof, and/or

3. What you are (e. g., fingerprint, iris template, etc., the strongest known methods).

The strongest in the third category is the iris.

So who needs a card, smart or dumb, when there is no known service or transaction which cannot be completed with a simple iris reader connected to the national UID DB, which has the iris template as an attribute, or even as the UID itself.

Inviting comments, especially objections.

Some representative questions which need be answered before defining various field are,

1. Will UID be push or pull convenience or both? Meaning can an individual show it when she/he wants to demand some service or can a government agency seek to see it? Or both? Can it be automatically identified (interrogated) by an RF device.

2. Will it be used for government purposes only (voting, registrations, security etc) or for commercial purposes also (banking, retail, assurances).

3. Certain data may not be available (e.g. UID of dead parents) or may be difficult to get (fingerprints of old people are generally inconsistent digital signatures), so what is a sufficient overall person ID signature. Will this vary with the kind of usage or application?

4. Will the validated UID provide empowerment to individual ? Or can an official still ask for a police verification for a passport renewal or issue of a document or disbursement of individuals entitlement? If this is not intended, we can do with very little of data about the individual.

I am sure the thought process must already be in full swing.

But unless the usage schema is fully charted, this project may not provide the intended effect.

My best wishes. I am looking forward to this project being a grand success.

JN Rampal