Thanks, Ben. Yes, it's perplexing, and worrying from the legal aspect.

I'd like to do some more research into it. One of the cases cited in the Guardian was about a Toyota Camry acceleration fault, where no one believed the people saying 'the car just accelerated out of control'. Because the electronic 'brain' of the car registered nothing untoward.

But I remember translating papers for an insurance case in Spain about a BMW that had crashed on a slight bend in a highway, where again the driver said the car just went mad and wouldn't steer.

In that, each side had to put up their expert witness to analyse the data from the memory.

I think maybe the position is different in countries based on Roman law - certainly the idea of probatio diabolica is often cited in legal cases here, that if evidence is contested, the burden of proof is reversed where one side holds all the cards.

I think 'discovery' is mean to perform the same function in the US legal system, but is maybe more about presenting the data, rather than having to prove their value, in terms of computer systems.