Day: November 29, 2004

Scott Mace has posted his interview with Dick Hardt… Interesting…

Second Law of Identity

Before we get to take a walk on the Norlin side, it&#39s time for the Second Law of Identity. And it&#39s simple enough:

The Minimal Disclosure Law of Identity

The solution which discloses the least identifying information is the most stable, long-term solution.

The thesis here is that the more identifying information is released, the more a solution invites abuse by rogue (and ultimately criminal) elements. We will return to a more rigorous discussion of these dynamics later in our conversation. For now, we will just point out that we are getting many reinforcing reports about the increasing professionalism and criminalization of identity attacks.

Let&#39s now go back to the case of Eric&#39s polycomm (by the way, I saw a ‘polycom’ at Bartell&#39s today ! – but luckily it only had a single “m”. I will try to get a photograph, though of course that might be dangerous…) If you missed out on the polycomm idea – which set the terms of reference for the current scenario exploration – please take a look at this.

The second law of identity tells us:

A solution in which the polycomm had to query my mobile phone for a social security number and then use this as a key for discovering the address of an all-knowing mp3 service would be much less stable than one which allowed the polycomm to query the mobile phone directly for the address of an all-knowing mp3 service.

But this latter would still be a relatively unstable service since it would need to be able to collect and potentially divulge all my music preferences.

The most stable solution would be one where the polycomm simply asked my anonymous (i.e. next generation) mobile phone what music selection “the anonymous I” would like to hear next.

In our example, the second law of identity applies as follows… (where “<” means “divulges less identifying information”)…

momentary anonymous listener preference < well-known individual music profile < citizen SSN

To consider potential economic value in these three cases:

Information linking my Citizen SSN to other identifying information is worth a lot. Protecting it is a complex and difficult long-term goal with continuously increasing associated costs, since this information will become an ongoing target of escalating professional attacks.

My full individual music profile – if systematic in the sense that computers can guarantee – has a potential (and therefore covetable) value for those marketing music.

But my “momentary anonymous listener preference” has no value to anyone but me.

Thus “momentary anonymnous listener preference” is a solution less likely to involve negative and anti-social side-effects than the other solutions.

And our law of identity is thus analogous to any prudent advice involving risk: minimize it.

My discussion about the relationship between employees and employers seems to have moved Eric Norlin to drop his proposed amendment to the first law of identity… He still has questions, though, like “say my employer tracks my web surfing habits – do i have to explicitly hand that over? surely, my surfing tracks are an attribute of my identity”.

I hope the screenshot below will help convince him that again, the most natural (and defensible) way to organize things is for the employer to ask the employee for consent to track him. That&#39s what we do in these here parts, and it seems to work well enough. I guess if the employee doesn&#39t want to be tracked, he needs to ask for employment in a different department.

That said, I really like Eric&#39s invention of the term “identifiable walk” – though I draw different conclusions about it:

We all *do* things that can be abstracted that can be used to identify us — (this is the infamous tier 3 identity from andre&#39s article)….should we have control over ALL of those attributes? my gut says that would be impossible.

in fact, it may already be that way — we all have an identifiable “walk.” the TSA is testing programs that would use that to identify terrorist types. Should i have to give consent to that camera when i walk in the airport? or is my walking in the airport consent enough?

So now it&#39s on to discuss “Norlin&#39s Walk”.

Craig Burton has been generously heaping kudos my way for aruging that a discussion of the philosophy of identity is orthogonal to the discussion we should have about the laws of identity. And I&#39m going to hold firmly to that direction, even though I received this titillating comment – probably more relevant to Scott Lemon&#39s Axioms – from David Rollow of CSG Systems.

Take a look at W. V. O. Quine&#39s essay “Two Dogmas of Empiricism” for relevant philosophical discussion about the most basic identity issues. If you understand it you will realize that the very idea of “axioms of identity” makes no sense. The “personal” identity discussion in philosophy is about things like persistence, sameness of tokens observed at different times, self-knowledge and self-identity, problems of no interest in the age of identity theft. The Quine essay is at a more fundamental level of discussion but it applies if you apply it.

Yes, the “it applies if you apply it” view of philosophy is hard to argue with, especially if you argue with it.