I work for a company that develops neural network software which is used for face recognition on a number of airports. The problem we've had over and over again is that government officials and airport security personnel have great difficulty understanding some elementary statistics.

Let me give you an example. One version of the software offers 99.99% accuracy (symmetrical true positive and true negative), a number that always seems very impressive to various officials.

What they don't understand and what we have to remind them all the time is that they need to take into account the large number of faces that are scanned by the software and that a 0.01% false positive rate isn't something you can ignore.

For instance in a large airport that has say a million people getting scanned yearly it means that 100 people will be incorrectly flagged by the system. The prior probability that a traveler is a 'person of interest' is less than 1/100,000. Plugging the number into Bayes' theorem you get that when the system flags a passenger, the probability that the passenger was actually a person of interest is around 9%.

The officials typically only listen to the 99.99% figure and ignore the reality of the relatively large numbers of false positives when dealing with huge numbers of people. Subsequently they treat the people the systems flag much worse than they would if they realized that the probability of a 'catch' being correct was less than 10%. We've done our best to try to educate them but usually they don't want to listen as it's an uncomfortable truth and it's much more convenient to say that the system has an accuracy of 99.99%.

On the face of it this may be silly as EU law obviously doesn't apply to US companies. That however would be misreading the whole thing. The EU is controlled by two entities the European Parliament and the Commission. The latter writes the laws and proposals and the former votes for or against them.

Members of European Parliament (MEPs) are democratically elected. Their primary problem is that nobody in Europe cares what they do or what they say. The EU decisions are in practice always complex compromises. The UK may vote for privacy laws the Netherlands wants in exchange for increased fishing quotas and the Netherlands wants it because the Dutch government can use it as a political tool for some other purpose. In short political ideology does not exist in the EU. This is a big problem for MEPs as they can't get reelected unless they get enough publicity and look as if they are doing something the voters care about. The system works against them and so on occasion they make loud noises about any issue they think will be of interest to the voters. Given the complex nature of compromises in the EU they seldom have the opportunity to do this. In this case the opportunity they saw was in the word "twitter". They know that voters recognize it and have scrambled to make themselves look like they are doing something decisive in the public interest. It's not real, it's just collecting brownie points from the public and getting their name in the papers. So you can forget about it. It has nothing to do with EU privacy laws or the US or twitter - it's strictly a PR thing.

The other branch of the EU executive and legislative power is the Commission. It has two functions. One is to act in the interest of the entity that controls it - the EU's civil service and the other is to provide a mechanism for national level politicians to get unpopular decisions through. The EU is run and controlled by the EU bureaucracy - it's civil servants. The Commissioners represent primarily the interest of their departments. The interest of the civil service is entirely self serving. They are for sending SWIFT data to the US as it will mean many fact finding trips to the US and other countries for the people in the departments. They are strongly for the introduction of checks and balances for sending the data as it creates more work for the civil service and ultimately increases their budget. The politicians on the national level have no problem with this as their use of the Commission is to get through unpopular legislation. When something popular is introduced it's always handled at the national level and the local politicians take credit for it. When it's something unpopular they simply say "we hate it too, but it's EU legislation, we can't do anything about it".

That's how the system works and it's not easy for the MEPs as they are not civil servants, they are politicians and need publicity and votes while they are not really meant to have any significant political power. That's why there was such an outrage at the EP rejecting a gay bashing candidate for the post of the Commissioner for Justice a few years back. Things like that are not supposed to happen and as a rule they don't. So when you hear that the EP is making an inquiry or that MEPs are making noises about something, you can safely ignore it. It doesn't have to make sense as nothing will ever come of it - they are just trying to get themselves noticed in order to get reelected.

Since little information is available in English, here are some points of clarification:

-The main charge was "aiding copyright violation". The decision of the court is mainly based on the fact that TPB did nothing to prevent it and that they in every way advertised that you could download copyrighted stuff on their site. The fact that this can be done with Google or any other search engine is beside the point according to the court. Google cooperates at least to a limited extent with copyright holders while TPB made a point of pissing them off.

- According to the court indifference to the possibility of the copyright violations occurring is not enough as an argument to let them off the hook. This is not so much a controversial point in the guilty verdict but a very controversial one when it comes to sentencing.

-According to Swedish law you can be found guilty of aiding even if the perpetrators of the main crime (i.e copyright violations) is unknown and the full extent of the crime is unknown as well.

-According to the court information provider neutrality as defined in among other things the EU's e-commerce law does not apply to TPB. Their main argument is that TPB was not a general service provider but a search service largely aimed at facilitating downloading copyrighted material.

-The most controversial point is the sentencing. The basic question is if the three specific persons could really be sentenced for crimes that they did not and could not have had information about (each individual download). The court's answer is yes and the reasoning behind it is fairly vague and general in nature. When it comes to the damages the reasoning is rather strange: Basically they say the following: The industry claims X million Euros in directly lost profits. This is clearly absurd as not all who download would have actually bought the product in question. So we'll split the difference and put the damages to X/2. X/2 turned out to be 46 million Swedish crowns. (€5 million)

Apart from the questionable reasoning one should put into context that a premeditated murder will in Sweden cost you on average 5 years in prison and 100,000 (~€10.7k) crowns in damages to the relatives. So although the guilty verdict of the court may be reasonable, the sentencing is very extreme by Swedish standards. As a rule damages are never in the millions and the idea is that the guilty party should have a chance to actually pay them. The sentence of 46 million crowns in damages is simply outside any Swedish legal practice.

Perhaps not all that unreasonable. The Geforce 3 was launched in 2001 and was the first real GPU that could do floating point operations.

2001: Geforce 3, 2.6 MFLOPs
2009: Geforce GX295: 1.78 GFLOPs

That's a factor of 680 in six years. Ok, the 295 is a dual GPU solution, but Nvidia is set to release the nv300 series this year that will have twice the number of cores. So he's actually predicting that the increase in computing power will be smaller than in the previous seven year period.