Real ID? Really?

December 15, 2009

Back in 2005, partly in response to the events of 9/11, the US Congress enacted the so-called “Real ID” legislation.  This was, essentially, an attempt to create a de facto national identification card by indirection, using drivers’ licenses because they are so common.  The legislation essentially requires state motor vehicle authorities to construct extensive data bases of identification information, including images of such supporting documents as birth certificates, and to adopt uniform anti-counterfeiting standards for drivers’ licenses. The deadline for completing all of this was set to December 31, 2009; citizens of states which were not in compliance would be disadvantaged, for example, by being subject to special security screening when attempting to travel by air.

This legislation was not well received by the states, about half of which have opposed the mandate; some have even passed laws forbidding their officials to comply.  Now the “Threat Level” blog at Wired is reporting that the implementation of RealID is likely to be delayed for at least a year.  Few, if any, states are currently in compliance, and several have not requested a waiver or extension of the deadline.

Beginning Jan.1, the law was supposed to have blocked anybody from boarding a plane using their driver’s license as ID if their resident state did not comport with the Real ID program. But the Department of Homeland Security is set to extend, for at least a year, the deadline of the Real ID program that has raised the ire of privacy advocates.

One of the motivations for the RealID legislation is said to be the fact that several of the 9/11 hijackers obtained genuine Virginia drivers’ licenses in false names.  But it has never been clear to me that, had the licenses been in their real names, it would have made any difference.

When the implementation of RealID was proposed, in early 2007, there was a public comment period.  I submitted the following comment to then-Secretary Chertoff:

I am writing to urge that the proposed standards for drivers’ licenses and identification cards be withdrawn, for the following reasons:

— The proposed rules in effect create a national identification card system, despite the fact that the REAL ID act specifically prohibits this.   That the proposed system would be implemented in a networked manner, rather than as a monolithic system, is merely a semantic end-run around the clear intent of the legislation.  Putting in place the monitoring apparatus of a police state is not good civic stewardship.

—  The proposed rules will impose an enormous unfunded liability on the states, with the certain result that getting a driver’s license — practically a necessity of life in the contemporary US — will become significantly more expensive, difficult, and time-consuming.

— Because the proposed rules create a specially privileged form of identification, data from the underlying data bases will be of enormous value to potential identity thieves and other criminals.   The proposed plan does not include any adequate privacy or security safeguards.

— Putting the proposed system in place will, in effect, create two classes of citizenship: those who have obtained an approved identification card; and those who cannot obtain or afford one.  That the second group will be subject to various forms of discrimination is a foregone conclusion.

The entire REAL ID program was part of an understandable, but ill-conceived and emotional. response to the events of Sept. 11.   But I am aware of no evidence that having a system, such as the one contemplated in the proposed rules, would have done anything to prevent that catastrophe.  We should be spending our time and resources on measures that will genuinely increase our security, such as better intelligence and police work.

I have not seen any evidence in the intervening years that has led me to change my mind.

Sign Language Efficiency

December 15, 2009

Many of you have likely seen speeches or other presentations that are signed for the hearing-impaired.  One of the minor puzzles associated with this is that, while on average it takes slightly longer to make the American Sign Language [ASL] sign(s) for a word than it does to speak the word, a sentence in sign language can be completed in about the same time as it takes to speak that sentence.  Of course, an obvious guess is that there is, somehow, less redundancy in signed language than in spoken language, but that leaves unanswered the more interesting questions of how, and why.

The “Physics ArXiv” blog at Technology Review has a report on some new work done by researchers at Princeton, to try to understand this phenomenon.  They have attempted to measure the information content of signs and speech empirically, by using samples obtained from a variety of sources:

The team has determined the entropy of American Sign Language experimentally, by measuring the frequency of handshapes on video logs for deaf people uploaded to, and as well as from video recordings of signed conversations taken on campus.

Their measurements show that the information content of handshapes (the basic elements of ASL) is just 0.5 bits per element less than the theoretical maximum, whereas the information content of speech elements (phonemes) is about 3 bits less.   From the abstract:

It is found that the handshapes, as fundamental units of ASL, are less redundant than phonemes, the equivalent fundamental units of spoken English, and that their entropy is much closer to the maximum possible information content. This explains why the slower signs can produce sentences in the same time as speaking; the low redundancy compensates for the slow rate of sign production.

(The abstract page has a link to download the original paper as a PDF.)

The researchers suggest that the difference exists because the communications channels used are different.  Redundancy in speech helps permit understanding even in the presence of noise.  Sign language is communicated via a presumably less noisy visual channel.

The goal of all this is to better understand the information content and characteristics of sign language, to enable it to be transmitted and stored more effectively via digital media.

%d bloggers like this: