Comment The Royal Mail's postcode database is to be privatised and campaigners for "open data" are furious.
But they only have themselves to blame; the open data campaign has been conducted with staggering utopian naivety. I strongly support the idea of opening up public datasets and I'm going to explain how open data can succeed, rather than fail like this.
Before we start, what do we mean by public data? It's information collected by the state and public authorities at the public's expense. Most of this has no value, it's simply noise. But some of it, most obviously anonymised medical data, has immense commercial value. In other cases the commercial opportunities will be much more modest, but they're still there.
Either way, and whatever the data, we won't know who will tap into it and how they'll begin to realise its commercial value. Hence the strong case for experimentation - that is, for opening it all up. Drugs companies, however, will pay a fortune for useful results. We know, because they've told us; the cash collected could help the UK build schools and help hospitals lower debt repayments.
So why insist on giving all that information away, for free, online? Why handicap ourselves?
Today's open data campaigners suffer from some crippling philosophical handicaps. Raised on utopian US West Coast ideas where value and property rights on digital things are considered to be evil, they simply can't see a halfway house between closed and free. Yet open doesn't have to be free. The problem isn't open data per se, but the no-strings-attached data giveaway of it.
Giving data away with no strings attached is potentially the worst use of taxpayer assets since the giveaway of state industries in the 1980s - a scandal that could cost billions. At least the Conservatives argued that former basket-case nationalised industries, once privatised, rapidly became profitable and boosted corporate tax revenues.
But this doesn't work with Google and its ilk quite so much: the beneficiaries of the open data giveaway really don't like to pay tax here. They work very hard to reduce their exposure to tax. And in any case, we want to maintain the direct link between data and value.
Some campaigners argue that "the data has already been paid for" - but this is another typically child-like argument. It refers to past costs, not potential future value. The great taxpayer rip-off lies in the future, not the past, and can only proceed because we refuse to put permissions and stipulations on the data.
Here's a much better idea
Instead of using the UK's "New Open Government License" (details here) to make data available to the world, let's open up the information with a licence that removes barriers to experimentation, so the proverbial kid in a bedroom can crunch it or come up with a clever idea of tapping into its value.
The licence merely needs to stipulate that a cut of any commercial value realised from this is returned to the owner - the UK taxpayer. It means that when Google harvests and sells the data onto Big Pharma, we get a cut. That could be billions of pounds.
It's actually a tried-and-tested idea in the software world. It's commonplace to have a license that permits experimentation on software components and libraries. It allows experimentation, but restrictions only kick in when the intellectual property is used in a commercial product. Take, for example, the Unreal Engine from Epic Games. It's a highly regarded 3D library that's free for non-commercial and educational use, and even commercial users can sign up and grab the source code right away.
But once money changes hands, you pay Unreal a cut. After all, the 3D engine is a crucial part of your commercial product, so it's only fair. Here's the licensing information. The agreement has permitted all kinds of innovative uses beyond games, such as in training, visualisation, animation and simulation. Similarly, you can't exploit data without the data itself. It's only fair to pay once that becomes commercial.
The beauty of such a licence - let's call it a "Fair Value Licence" - is that it appeals across the political spectrum. I've spoken to Labour MPs who like the idea of of major US corporations paying their bit, and naturally see the benefit of revenue to the Exchequer, too. And when I ran the idea past Matthew Sinclair of the Taxpayers' Alliance, he saw a benefit I hadn't thought of.
"Government departments would be incentivised to open up more data, which they aren't today, if they could keep a percentage of the percentage," he pointed out.
Today, though, the hardest thing is getting the data "opened" in the first place.
There would probably need to be some kind of market test to ensure companies don't remain in permanent "beta", depriving taxpayers of their revenue, or inhibiting more innovative uses by companies using their market muscle to just give away the data. Google in particular is notorious for this, and once the advertising giant reaches an established market position it stops innovating; Maps being a prime example of this.
But what's not to like?
The UK open data movement has failed its first big test. The Treasury was convinced the postcode database had some value, and including that in the sell-off of Royal Mail made the whole package more attractive. You can hardly blame them for that, even though the worst of all worlds is now the most likely outcome. A private monopoly will exploit the data and the taxpayer won't see a penny. Trebles all round!
It really isn't such a big change to get open data back on track, although a Fair Value Licence may be a tricky one for non-technical people to understand if they don't know the software licensing world. The more dogmatic Freejadists might have trouble accepting it as well.
But if you believe in open data, you should think about it. And back it.
It's only fair. ®