Thursday, April 06, 2006

Scaling and Scoping an ID Card Protest

Tom of Blairwatch has been doing some sums on how badly the ID card registration process is going to fuck up. You may recall we did something similar for the biometrics. Using the Home Office's own figures, in which the best identifier (iris scanning) had a success rate of 96%, and assuming that each one of 44 million cardholders is checked once a year, there will be 4% of 44 million, or 1,760,000 misidentifications/nonidentifications. Even assuming that the identification can reach the reliability standards of the public-switched telephone network, 99.999% reliability on 99.999% uptime, there would still be 44 botches, any or all of which could result in wrongful arrests, fines etc.

Tom concludes that over the scheme's first ten years it will have to process people at the rate of 1 happy citizen-unit per 72 seconds. Charlie Stross argues that's optimistic for various reasons. But hold it for a second (or 1/72 of a citizen). As Tom points out, 1 per 72 seconds is the rate at which the queue grows when the thing spreads itself over the landscape in a pool of its own credibility.

Now, let's consider a hypothetical registration centre. Probably a good example might be the UKPS London passport office behind Victoria Station. There are (I think) two floors of processing, each with a queueing capacity of about 200. There is also a further queue to pass through security screening and a muppet check of documents, with perhaps space for 50. I.e. 450 people can queue at any given moment; 450/72=32,400=540 minutes' queue capacity=9 hours, or effectively a working day (actually, a queueing day - you don't have to give people breaks from queueing). That's not the whole story, though, as it's futile to queue more people than the centre can clear within the remaining working hours. Assuming (as Tom's original study does) that the shivil shervants work 8 hours a day, the maximum queue is some 4 hours=200 citizens.

To put it another way, the backlog increases whenever a centre - any centre - is disrupted for 4 hours. I would suspect that the natural burstiness of people arriving at the centres will mean that some fraction of the queueing capacity will always be in use - if it's an average of 50%, an effective 2-hour demo would queue-out that centre, with people spilling into the streets, backing up traffic and creating a risk to health and safety which (of course) neither they nor the employees in the centre can be exposed to.

A simple exploit would be simply for more people to turn up and join the queue. After all, only a few people at each site disruption would need to do anything arrest-worthy - the rest are just queuing up to get their ID cards. (Assumes J.G. Ballard voice) It would seem that the best way to defeat this tyranny is simply to yield to it.. And queueing as a means of revolution is certainly rooted in the organic context of British society.

No comments:

kostenloser Counter