* Posts by Dr Dan H.

4 publicly visible posts • joined 11 Apr 2007

Reverse-engineering artist busts face detection tech

Dr Dan H.
Pirate

Using the technology against its self

Most CCTV cameras are highly sensitive to infrared light, which is helpful since you can illuminate scenes covertly with infrared lamps. However, it shouldn't be difficult to develop face paints which look like normal skintones to us, but which reflect or absorb strongly in the near-infrared. It is also quite possible to put infrared LEDs onto spectacles, which emit strongly enough to interfere with the face-recognition systems..

Home Secretary swats away calls for Mosquito ban

Dr Dan H.
Jobs Halo

There is a better way, you know

The purpose of the Mosquito is to emit a sound which is tooth-grindingly annoying to children, but which has no effect on adults. So, instead of a near-ultrasound tone generator, all you need is The Best of Frank Sinatra played on a random, endlessly repeating loop via a number of small, well-armoured speakers all around the area to be protected. Kids tend to absolutely loath any music which is not their current fashionable favourite, and old Frank seems to top the list of Lovecraftian horror as far they're concerned, whilst being quite pleasant for older folk.

To add to the horror for kids, open a tea shop in the vicinity which is heavily subsidised for pensioners; this ensures that not only is the area a sort of black hole in Youth Musical Coolness, it also has hordes of people whom kids really don't like associating with; this should shift the youths to somewhere else and garner quite a lot of approval from the Complaining Oldies sector of society, at not a lot of extra cost.

Plague: The new Black Death

Dr Dan H.

But where's the vector?

The medieval plague was most emphatically NOT Y. pestis. What people always forget is that the modern rat Rattus norvegicus has not been in the country for all that long; R. norvegicus was not present in medieval Britain. The black rat, Rattus rattus was present, but only in the ports; it seems to have been at the northern extent of its range and didn't breed outside warm buildings.

Out in the countryside, there is no evidence for rats being present. In modern times, wherever there are rats and owls present, you always find the bones of rats in owl pellets, and these owl pellets persist quite well in the environment. In archaeological layers you also find owl pellets, complete with almost all the expected prey that modern owls eat; voles, mice, frogs, toads and so on.

What you do not find in the owl pellets is the remains of rats, any rats.

Rats were simply not present in the countryside of medieval Britain. So, the major vector for Y. pestis was not present in medieval Britain, which pretty much rules out Y. pestis for being the plague.

The minor counter-arguments are epidemiological. Y. pestis breaks out sporadically whenever the rat population booms and gets infected; the plague then runs riot through the rat population and kills them off. The surviving homeless rat fleas then transmit plague to humans. What this causes is unconnected outbreaks of plague, always preceded by an increase in rats and a crash in their population.

The medieval plague, by contrast, seemed to break out along the roads and rivers first and isolated country spots much later; this is consistent with a long incubation period viral disease.

Finally, the population of Western Europe seems to be especially immunologically resistant to viral diseases, being the descendents of a widespread viral plague (the vulnerable members of the population died out then.

DNA database 'will span most of the UK population'

Dr Dan H.

Three cheers for the Police Paternity Database!

OK, so the police are slowly building up a database of the DNA of everyone who comes into contact with them, and presumably have a fairly complete database on the familial relationships of everyone in the UK. So, they know who everyone's brothers, sisters, mothers, fathers and uncles are, right?

Err no, they don't.

As far as is known (and the police could probably comment more authoritatively on this) between 2% and 15% of children are not sired by the person who is officially their father. The rate is lowest where the husband is high status (i.e. really rich and powerful) and highest where the husband is low-status (i.e. on benefits, poor, a regular "customer" of the police).

This has two implications. Firstly, there will be an error rate for the statistics of around 10% or thereabouts, which will force the police to go back to basic police work to secure convictions, and may free the odd criminal on appeal.

Secondly and more sinisterly, the database will also be proof that certain people were not fathered by their official daddy, which is proof absolute of infidelity.

This might just be a show-stopper; if the data ever gets out, and given this Government's record on data security this is pretty near certain. When investigating a crime, an outsider could even, by a process of deduction, work out that a family member who _should_ have been arrested but was not arrested could not be sired by the same father hence was someone else's bastard.

This sort of information leaking is not what the database was designed for, and should probably limit its use to straight identification, unless the police wish to dice with the moral and ethical conundrums that come with sensitive medical data. My guess is that they simply have not thought it out and would rather not face the problem until it bites them, which is a very silly attitude to take.