How to save UK's open data: Meet the 'Fair Value Licence'

If the Postcode database is sold off ...

... will I be free to set up my own rival system? I plan to make a fortune from the sale of vanity postcodes, for example to a community who'd all like to live in SIN.

Stop using postcodes

Bloody Conservatives selling data that I have personally funded with other tax-paying members of the public.

I have no problem with the a share of income going to support the postcode administrators, but I strongly object to the database not being made available for non-commercial use. I would suggest:

1. Boycotting the use of postcodes

2. Creating an open source alternative based on longitude/latitude, with a simple country or town code. How difficult can it be?

Not a good idea. "not for commercial" use stipulations are hard if not impossible to define. If you want to build an aggregate data set involving 5 or 10 datasets, with all of them having these kind of restrictive conditions,then the aggregate data set is bound by all of them.

The idea of giving a cut is also difficult; you integrate 20 datasets, everyone wants 10%, well it doesn't add up. Or you can negotiation on a per use basis, but then the negotiation costs more than you make.

If you want a different, but similar example, there are lots of TV programmes from the 80's that people *would* pay for, but are hard to release, because the back ground music is copyright be someone else, and the original license was for broadcast. In the end, many just don't get released.

Now, the postcode database is critical because it is the basis for many other proxy results. For instance, post codes are a good proxy for social class -- vital for many medical analyses. The postcode is particularly useful, because it's such a standard resource; if you have 20 analyses and they all use postcodes, they become comparable. If one uses postcode, one long/lat they aren't.

Of course, it is true that many who will benefit don't pay taxes here; but that it a completely different issue. Indeed, not long ago the Reg was arguing that Google, Apple and so forth *shouldn't* pay taxes. Still, the value to the country of having an open postcode database probably exceeds the value from the sell-off; unfortunately, the actual value is harder to calculate because it is a diffuse good; but this is exactly the sort of thing Governments should take into account, and corporations rarely do.

A sad position to be in, but not one that is at all surprising.

I cannot think of any of the privatised industries that have become better since other than probably North Sea oil (Which paid for our WW2 debt).

Royal Mail used to be really good 2 deliveries a day and always on time now it is 4 days if you send something first class.

BT - broadband network sucks and they still get subsidies all the time.

Water/Gas/Electricity - We get the absolute minimum amount of maintenance it is ok at the moment because it was originally done pretty well.

8/10 - further work required

Generally I agree but Phil Lord's point is also accurate. Ideally you'd like value identification/creation to ramp up without being bogged down in the negotiations that his scenarios would create if a cut were required. Much of the public interest benefit is non - financial as Phil Lord also notes

The current licensing is naive in a number of other ways too.

It has no 'off' switch - if any use of the data is considered against the public interest - eg combining with other public data sets or FOI requests to break privacy it cannot be withdrawn. For relatively static or stable data (postcodes!) that's an issue with political backlash.

It doesn't encourage feedback - there's no benefit in providing data correction nor service improvement information back into the public sector - often there is not even a feed back route.

The other side of the equation is that public sector is not incentivised to properly document the data providence, accuracy and collection methods that define its meaning nor improve the data quality and quantity (eg sampling rates, accuracy more data fields, frequency of release).

I'd suggest a revocable free licence (possibly underpinned by regulation in case of abuse to take it from civil to criminal) whilst allowing the creation of a more privileged/engaged user group that returns some value - payment in cash or in kind, couldhave guaranteed access (so 'approved') and sharing the costs of data improvement.

So long as the licence cost doesn't increase

We pay £7K a year for PAF and Postzon. It was a massive cost at first for a small company but over the years we've adjusted and grown so the cost isn't an issue anymore. If the new private owners increase the cost to get their money back on their investment then it will become an issue for us and possibly devastating to other small companies who depend heavily on it.

As it stands it's impossible for a penniless creative thinker or programmer to use Royal Mail data to create something innovative or useful without securing funding just to pay the licence fee. This is what Tim is talking about. So should it be free? Not if it makes it difficult for the owners to finance the resources required to keep it up to date. This is no easy task and lots of people are employed to manage the data in terms of people calling/emailing requesting changes to data, adding new addresses, and more people to manage data quality and supply.

The current format of PAF is very 1970's and it's horrible to work with. The RM have had many opportunities to update it but fail every time to act. There isn't even a unique identifier for each record. This is supplied on request only as a separate file where you need to cross reference the two datasets in order to match up the correct unique identifier. So maybe a new private owner will choose to modernise the whole database and offer it in modern formats. There are other quirks in how the data works and can only be realised by reading the manual cover to cover.

Returning value to the taxpayer

Errr? Isn't that what VAT and Corporation Tax will do to anyone who creates derived products for sale from Open Data? ... without all the paperwork bureaucracy and licence police your scheme would involve. Before you say Google, Vodaphone etc don't pay enough tax, sure, but that is a problem you need to solve anyway and they won't care about the tiny crumbs they would have to pay when exploiting your closed version of Open Data.

If your attitude had prevailed the www and the internet would be crippled by electronic toll booths and micro-payments making up most of the traffic. Some of those with vested interests would still like to cripple the internet that way. Yet, as a low cost of access relatively level playing field, open public infrastructure the internet has revolutionised communications and the economy. Open data, such as the Postcode Address File is just such an infrastructure.

Anyone fancy privatising the DNS system and charging for DNS look-ups? In the same way that that would cripple the internet, charging for postcode data cripples its use in the 'real' world.

What is really being sold is the right to toll any internet transaction that involves a postcode look-up as well as other postcode use which isn't growing. That's a Tax! Selling tax raising powers is simply wrong.

Open PAF now!

Open PAF now!

Errrr? Isn't that what VAT and corporation tax is for? Those will return the value of Open Data to the taxpayer without the need for costly and unnecessary bureaucracy and that is why the 'fair value' licence shouldn't apply to public data. Before you say what about the Googles and Vodaphones who don't pay tax, that's a completely separate problem that needs to be solved anyway. They wouldn't be bothered by the few crumbs of tax on re-using Public Data. At the moment Google are voting with their cheque book and refusing to licence PAF data for use on their maps, where it would be valuable for their users, but wouldn't increase traffic enough to justify paying.

If your attitude had held sway as the internet and www developed they would be crippled by tollgates and micropayments making up much of the traffic. It is the open environment that reaps dividends and creates value and that environment is now an essential infrastructure even while vested interests try to undermine the principle. Public core reference data such as postcodes and addresses is a similar infrastructure whose use needs to be maximised not tied up in complex licences and expensive fee collecting mechanisms.

Anyone fancy paying for DNS lookups on the internet? Making people pay for postcode or address look ups on the internet or in the 'real' is similarly daft!

That's why we need to open the Postcode Address File now and not cripple its use by excessive upstream charging or cumbersome licensing.

I'll going to drill a hole in my head and then become a politician...

Postcode Address File (PAF).For Free?...How is it that politricks is never short of ideas on how to tax the little guy i.e. me. Whereas politicians are always giving things away for free to mega-corps ... In the US, Mega-corps can even become people and have special rights too.... Mmmh I wonder if that works the other way around?. Cos I'd sure like to reinvent myself as MeMegaCorpI.com and get more things for free from stupid politicians!

This sounds reasonable, but how will it be enforced, given the iffy data quality?

I agree that non-commercial uses should probably not be charged for lookups; however how will you track the real quantity of commercial use, to avoid an innovation stifling flat license fee.

The raw data provided is also of iffy quality and structure, because I am aware of someone who tried to make consistent use of it and ended up having to reprocess the data to make it bearable to use, even just for basic location Postcode look-up!

Data quality is critical for any as-is quantity charges, so I doubt this approach will work.

There are also internet based services now which are a lot cheaper and even free, so this is too little too late!

The problem is how you define "commercial usage". The likes of Google will quite happily use the data as part of a "free" service that doesn't generate any revenue directly, thus doesn't require any payment, whilst siphoning of as much cash as possible through other streams such as ads on the site or selling on aggregated data from service usage.

It's less of a problem with things like the unreal engine, because the only people making significant enough amounts of money with things like that are selling an actual product, where the revenue streams are easier to identify and thus charge for.