May 12, 2005

Fare thee well DIDW

Digital Identity World 2005 is over. Doc Searls closed the show at noon today having summarized the themes that were presented over the previous days and earlier presiding over a Customer Facing Identity Directions panel which included Kim Cameron and Marc Canter.

As posted previously, lots of discussion and conversation about identity -- or at least the technology of digital identity -- that shows marked evolution from the themes of the inaugural event in 2002. But, although exhorted by key-notes to examine the issue from a "higher" view, the many break-out sessions (that I attended) came up short. This is not to suggest that it wasn't a good event and that the "conversation" has not advanced. In all likelihood, this stage of discussion is an essential precursor to where it needs to go. All things get to their natural end-point through a consistent set of steps, all of which need to be trod on. To wit, even if the ball is hit over the fence the homer isn't scored until the batter has had a foot on every base.

So what do I think was missing? Or, stated differently, where does it have to go? Here are some stream-of-consciousness thoughts. In no specific order, I observe that

  • The language of (digital) identity has yet to evolve and solidify to a point where it is itself not a barrier to development. There are two problems here. First, the same words are being used by different people at cross-purposes. Second, words are being used for their value-charged effect in ways that compromise -- or worse, hijack -- their implicit meaning. For example, the word "trust" remains much abused. It is being used to express a sense of confidence resulting from a defined relationship – even to the extent that the certainty is derived from a referral by a trusted third party. Thus: organization A "trusts" the credential of a person representing organization B because there is pre-arranged understanding of the authentication/authorization of the credential. This, of course, is not what the word trust implies at all in common usage. Trust is, in fact, an expression of faith in an outcome or expectation in the absence of proof in that instance. So the system is being developed as an artifice to emulate trust. Best analog: the legal system. In addition, the specific definition of the word "identity" itself remains relatively fluid.

  • The metaphors being used need critical re-examination because, extended to their full conclusion, they often do not hang for identity. The wallet "notion" is widely used both because of the physical-visual representation of separation that several credit cards in a wallet (along with other identity factors such as driver’s license, insurance card, etc.) provide as well as for the easy leap to the established credit card or Interac transaction processing models. While good to stimulate thought, this metaphor's (or analog) most significant shortcoming is that its model works well for payment transaction processing which is a defined-liability context. In these transactions, the existence, magnitude, transference, and ultimate retirement of liability is discrete and indisputable. Not so with identity. Which brings us to the necessary fuller exploration of

  • Liability remains a concept that is danced around and glibly set aside as a mere externality that will eventually take care of itself. I, personally, and many other conference attendees tend to underscore the fundamental importance of the liability issue to the ultimate value and success (or not) of any digital identity system. Unfortunately, it has yet to even develop into a proper elephant in the room -- only a cloud hanging overhead. If one thinks beyond the enterprise environment regardless of whether the transactions are B2C -- with one's own customers or those in a "circle of trust" -- or B2B/G (where the issue is resolved within the legal contracts resulting from the business deal), liability's shadow looms darker and larger -- and it has a trunk. Maybe the liability question would be more prominent we figured out,

  • "Where are the lawyers?" There must be a few lawyers with a specific expertise and practice that would be relevant to the discussion (e.g., privacy, civil rights, liability, and so on). Since it would appear -- by popular sense -- that most of the matters to be addressed now are commercial not technological (or, as Jamie Lewis said, deal with "automobiles not the chemical composition of asphalt"), it stands to reason that there ought to be a few lawyers weighing in. More cause for getting consiglieri on board is that laws are being enacted that have a direct impact on the outcome of this nascent industry's activities. Send lawyers, guns, and money . . . (That was completely unnecessary and utterly gratuitous, but I love the line.)

  • Anonymity and privacy appear to have a renewed status in the discussion. This again appears to be at least partly the domain not of technologists but of lawyers, civil rights activists, and so forth. Where is their voice (see previous)? This flip-side to the certainty we seek of a digital identity structure is critical for various reasons I think we all understand. We have but begun to scratch its surface.

  • Assuming away critical system conditions to develop the mechanics may be a necessary evil at this stage, but can't continue much longer. I specifically refer to the stance taken by technology-centric solution developers that their concern is not the integrity of the identity and initial credentialing. Rather, the solution assumes proofed inputs suitable for "trust" to develop. Thus, STS or what have you can exchange credentials and tokens satisfying the mechanical aspects of questioning, presenting, and authenticating, etc. "assertions." There is, of course, nothing wrong with this posture if one is merely contributing a service to a much larger system that is or has been sorted out. The business I work for, for instance, has a service called the Electronic PostMark, which time-stamps, encrypts, and overstamps a "post-mark" onto an electronic transaction as a statement of authenticity at a given moment. (There's more to it, but this should suffice here.) What has been so far purposefully left out is validation of sender and receiver credentials: the service accepts the digital credential presented in a transaction so long as it is provided by a subscribing credential issuer (e.g., the government of Canada).

    We are, however, at this conference and others; as "the identity gang;" as vendors and service providers; as standards bodies -- formulating a system from essentially nothing (or at least from nothing more than disjointed pieces and nifty ideas). The system has to be addressed holistically despite what is likely to evolve from unanticipated quarters. The foundational pieces like the identity creation process and its subsequent manifestations are particularly relevant to system design. Infocards as visual manifestations of role credentials for identities is an excellent idea; but they are as meaningless as a driver's license that shows an age of 23 for a boy obviously not yet 17. (Need to add here abruptly and parenthetically -- because I have no place else to put it -- that Kim Cameron's "Seven Laws" is an impressive, if not ever-so-slightly hyperbolic stake in the ground for a higher level discussion of the guiding principals for the system.) So,

  • User-control of the digital identity attributes (i.e., how much and what information to present) as self-assertion will not work because the system has not replaced the absence of trust with process and greater certainty from "projected trust." What I mean is that self-assertion (as opposed to self-presentation) of identity is meaningless. After all, if I'm skeptical that you're who you present yourself to be, then providing a credential that you've created is not much help: although Jim Rockford made hay with the self-asserted business card credential throughout the series' run. Credit cards work where cheques began to fail because a credit card (once proven valid at point-of-sale) presents the "projected trust" certainty of the issuer not the holder. Personal cheques -- too easy to forge -- ultimately diminished as such an assurance of certainty and have been largely replaced. There's also that little thing about liability acceptance for fraudulent transactions, but I digress. The driver's license as a credential works only because we trust that the state of California has proofed and authenticated the presenter. We don’t trust the holder of the license any more than before, but we have some small piece of proof (there's that word again) to increase our certainty about some particular attribute of the holder.

    One implication of the self-presentation model (which I believe in) is that there will likely to be various issuers of digital credentials attesting to different holder attribute sets so that the holder need present only as much information as is necessary and no more. The topology of such a system I explored in a series of essays ("Identity Planet") located on this page.

  • The solution does not exist. A solution and system will develop with warts and funny features. Not many things in this world have ended up as they were designed a priori, and those the did often come up short in critical ways. I believe everyone involved both realizes that and recognizes that emergence results from the intersection of critical functions, features, and developments the value and ultimate importance of any not being readily apparent until after the fact. The most heartening thing that I observed was not the "discussion" or the bun-tossing among vendors and standards bodies. Rather it was that these vendors and standards bodies are developing and implementing some active response to the challenge in the real world. They make the conditions right for more rapid advances. Oh, maybe not in the way anybody foresaw or intended, but advances they will be.

    All in all, right now it's the sardonic Gallic worldview that best captures the slow pace of fast change in this industry space: Plus ca change, plus c'est la meme chose.

    Posted by Grayson at May 12, 2005 05:33 PM