It's worth plugging all the ways in which UK citizens can contribute to independent reverse engineering this info in a clean room manner:

If you've got an iPhone, Android phone or other GPS device: http://www.freethepostcode.org/

If your house was built before 1940 then locate it on an out-of-copyright map: http://www.npemap.org.uk/

Just type in your streetname and postcode: http://postcodedb.sourceforge.net/index.php?p=submit

Identify the location of nearby postboxes: http://www.dracos.co.uk/play/locating-postboxes/

You can see the results here: http://old-dev.openstreetmap.org/~random/postcodes/

Click on the names of the three different sources in the top left to compare or use the + sign in the top right to switch the underlying map as well.

I wonder if there's a 'canary' post code in this data that doesn't actually exist but could be used to prove where you got the data from. That's the sort of thing I'd do.

I'm fascinated by Mountweazels: http://en.wikipedia.org/wiki/Fictitious_entry

Anybody have any good examples they've seen on the computer side?

NAVTEQ's map of San Francisco has a street called "Submarine Bells Way" that doesn't exist in real life. Tele Atlas's map of Oakland, CA, includes a made-up street name "Del Norte Drive"

someone who has a legal source could do a diff against this and see if that's the case

The 'legal source' would have to have the canary address - so you could tell if it was used.

Unless somebody thinks this was a deliberate leak to poison the free data - in which case it would have it's own canary address

Why would there be a 'canary postal code' in the 'legal source?' At the very least, if it was there just in case the source was 'stolen' then they would have to tell people that legally purchased it to ignore that address. In that case, the information about the 'canary postal code' would be 'out there' anyways.

To stop me buying a legal copy and then immediately selling it as my own data. You don't have to tell the users that the data is there, simply generate a postcode that doesn't exist - then when a new postcode mapping product comes on the market, check for that postcode.

> then when a new postcode mapping product comes on the market, check for that postcode.

I realize that, but won't adding a 'bad' postal code possibly interfere with customers' usage of the data in the first place?

No because the customer would never use that bad address, none of their customers would ever use it as a home address for instance.

If they decided to post a letter to every address in the database then one of them would be undelivered - but that is probably not a problem given the efficency of the post service!

It's well-known that there are canaries in Ordnance Survey map data.

This was the standard method of "copywriting" Maps - put in an artificial or geological feature that isn't there. You know who's been stealing your data then - The "There be Dragons" hack

Related: the Free the Postcode [1] campaign is a public-sourced database of postcode data.

[1] http://www.freethepostcode.org/

This isn't the full database as far as I understand which maps postcodes to actual addresses. This leak only deals with coordinates.

I think this is actually better. This is the secret source which is referred to in order to create, but doesn't include, the PAF file which maps postcodes to addresses. The PAF file is however easily available, though under a licence, and therefore is regularly leaked.

http://www.royalmail.com/portal/rm/jump2?mediaId=400085&...

The available columns are

Postcode, IntroductionDate, GridRefEast, GridRefNorth, County, District, Ward, UserType, GridStatus, Country, WardStatus, NHS_Code, NHS_Region, Long, Lat, OSRef, Update

You can already get the co-ordinates from a postcode using google maps. So I can't think of anything massively exciting you could do with this information. Maybe some sort of offline navigation app.

how about "what postcode am I in"? when out and about.

also "click on the map, what's the postcode?"

quite handy for getting the area code i.e N1

If those are the columns this isn't 'precise' postal code data. The log/lat is just the 'centroid' of the postal code. At least in the US, most -- if not all -- postal codes are irregularly shaped, so it's not like you can pull up precise GPS data on the perimeter of the postal code (i.e. in order to draw it on a map).

UK postcodes are equivalent to the 5+4 US zip codes. They list part of a street, typically 10-20 houses.

still close enough to search for things.

% postcode

NG15

% ff http://www.google.com/webhp?q=ng15+curry

bah google changed their uri scheme

such a shame...

Just FYI - this won't be much use unless they update it regularly - we have a churn of about 6000 postcodes a month in the UK (business specific addresses, new property etc).

I completely fail to see what is so important about this. Don't you by definition need to know the postcode of whomever you need to send mail to? If so, what benefit do you get from this stuff?

The UK doesn't have the same kind of Freedom of Information Act as the US, or the same conceptual approach towards citizens being owners of the government.

A lot of stuff that is paid for by taxes in the UK is not automatically public domain, but copyright of the government (or technically, Her Majesty's Stationary Office, the govt. printing bureau) and only available on payment of a stiff fee. Basically the state qua legal entity is much more 'selfish' than in the US, where pretty much anything it puts out is public domain by definition, and it's deeply annoying if you're the sort of UK resident with a presumption of that information wants to be free.

And is very expensive and restrictive in how you can use it.

We licensed the Postal Address File (the list of official correct addresses for each house) and we had to add encryption to our product to stop people being able to extract the raw data from the code.

The same applies to Ordnance Survey map data - there are a lot of apps that aren't possible/easy in the UK because of the cost and licensing of the maps that the taxpayer paid for.

'Useful' information could be the ability to do a 'what postal code am I in now' query based on your current GPS coordinates. Or to pull together your own 'postal code map.'

Usually postal code data isn't that accurate though. The Long/Lat information is usually just a point in the 'center' of the postal code and most postal codes are irregularly shaped.

You also need it to generate a "addresses within X miles of Y" type list for a website.

Yea, but you need to combine it with a db of address->long/lat pairs.

If you want to display a heat map of incidents/sales/etc where you have an address then something like this is essential. In the US, for example, the Census Bureau's Tiger/Line database is free and maps all the way to street addresses. This UK database isn't nearly as thorough as Tiger/Line but it's much better than nothing.