Recent BBC news coverage stated that UK postcode data will be made freely available under an open licence from April 2010.
Colleagues at EDINA pointed out that some of the coverage assumes that the open data will be the same as that contained in the Royal Mail’s Postcode Address File – but this is uncertain.
Since 2000, UK postcode data has been managed by a consortium called “Gridlink” which comprises Ordnance Survey, Royal Mail, the Office of National Statistics and the General Register of Scotland.
Ordnance Survey collects the data from the other consortium members, and they all have the right to re-sell the collected data.
We can see the contractual setup of the Gridlink consortium in this response to an FOI request regarding the National Statistics Postcode Directory.
In short; it’s complicated. So it’s strange but not surprising to be sent a link to a confused opinion piece in the Financial Times discussing the role of “the Royal Mail, whose intellectual property the postcode datasets are”.
There are many different “postcode datasets” produced and licensed by different members of the GridLink consortium.
ONS and Royal Mail both sell “data products” which add lots of contextual data to the basic elements – which are the fact that a postcode exists, and the fact that it exists at a location. PAF records all the delivery addresses which share each postcode. The NSPD is meant for demographic statistics – it includes references to codes for census areas, health authorities, local government, etc covering the location of the postcode.
The simplest bit of raw data is the association of the postcode with the national grid reference at 1 meter resolution. The rest is added value, created by looking at the spatial relationships with other data sets. Imagine pushing a pin through a loose stack of paper shapes, giving them a shake, and seeing which ones stay on the pin. Then note which shapes stayed on the pin, and arrange the notes in a table form with an entry per postcode.
If the promise of data.gov.uk is realised, then anyone should be able to derive datasets that look a lot like the NSPD (but not the PAF). Raw postcode data provided under an open license does not stop Royal Mail from creating their own added-value products.
Open access just to the mapping between postcode and location would re-enable the “civic mashup” services that were dependent on ernestmarples.com. It’s enough to answer questions like “Who are my local councillors?”, “Are we in this school’s catchment area?” and “What new development work is being planned within a mile of me?”. Open postcodes will be fundamental to unlocking data.gov.uk and turning it into knowledge.
The data resulting from the activity of Gridlink is owned by the Crown, and is Crown Copyright. Postcodes are facts, they can be observed and independently recorded. NPEMaps and Free the Postcode are, or were, two efforts to reconstruct the facts from open sources – from old maps, or from GPS coordinates, and peoples’ knowledge about their own postcode.
I could go on in more obsessive detail, but you probably get the point. I’m unnerved by the idea that emotive media coverage of the Royal Mail’s future, as well as OS’s, will colour the consultation on opening state-collected geographic information in the UK. I would like to see more facts set out straight.
Thankfully, one worrying mis-statement in the FT article as already been corrected – the consultation on how to provide open access to the Ordnance Survey’s data was released today as expected. It runs til March 10th, and the likely last possible date that the current government can act on this is April 22nd. More on this later.
I’m told that since the postcode numbering system is devised and controlled by the royal mail then it falls under copyright law (ie there is originality in the scheme). This is just like the dewey decimal system that oclc vigourously defends under copyright law.
Does that complicate matters at all?
“It’s enough to answer questions like “Who are my local councillors?”, “Are we in this school’s catchment area?” “
Not sure this is entirely the case.. Both local authority areas and LE catchment areas are strange beasts that require polygon data from the boundary line dataset? PAF & GIS extract certainly are enough to know who my “Closest” schools and elected officials are tho, so it’s certainly moving in the right direction.
Ian I – right, I was making that statement having in mind an assumption that we will see the BoundaryLine data released under an open license from April 2010 as part of the “Making Public Data Public” programme.
That would include “political” administrative boundaries as listed at http://www.ordnancesurvey.co.uk/oswebsite/products/boundaryline/
School catchment areas, I am now realising, are another matter – I had no idea there was such uncertainty about how they are defined and managed, as the comments on this catchment area project proposal for Show Us A Better Way suggest:
http://www.showusabetterway.co.uk/call/2008/07/catchment-areas.html
I’m amazed to learn that there is a small body of academic research on geometry models for catchment area definition http://eprints.soton.ac.uk/54904/
Not to mention the real estate marketing value of catchment area information, I’m sure I’ll learn to appreciate this as my toddler turns school-age…
Any info on where the source data on fuzzy boundaries for school catchments can be acquired, would satisfy my curiosity.
Ian D – a difficult point. Again it would be useful to see a written reference for the claim.
The 2003 OCLC lawsuit over re-use of Dewey Decimal classification system is a new data point for me.
I found an archived discussion about whether DD codes can be attached to items in an (open access) institutional repository of academic work. Here are some words from an OCLC representative:
“There are absolutely no restrictions on using the Dewey numbers. You can assign those numbers to your works and then use them to organize your works. It would be nice if you said something on your site about Dewey being copyrighted. But otherwise, numbers are numbers and you can use them to your heart’s content.
The problem comes when you try to assign meaning to those numbers; then you’re using the work of the Dewey Editors. The text associated with those numbers IS copyright. But, if you can restrict the usage to just browsing up and down the numbers, you’re good.”
http://old.nabble.com/Re%3A-subject-hierachy-using-Dewey-Decimal-would-allow-moreuser-friendly-browsing-p21076114.html
Is the situation analogous? The part of the Dewey classification system that gives meaning, provides a description, that looks like a “creative work”. The part of the core postcode data that gives meaning, the location reference, that is a “fact” which can be reconstructed independently.
Postcodes are strings, locations make them meaningful.
It could be possible to argue that the laying out of the postcodes over locations in the UK is a creative work as a whole. Crown Copyright is claimed – but then refer back to the discussions about the dubious applicability of copyright to geographic information in the first place, and the reasoning behind the old Talis Community License and the Open Database License.
At this point I suspect you understand the law much better than I. I don’t understand how OCLC can claim copyright on a work that was made in 1873 – even if there have been additions and refinements as new bodies of work arise and need to be classified, there must be some portion of the DDC which is “out of copyright”. The time period may vary according to jurisdiction, but there still is an eventual expiry time – that is part of the design of copyright – right?
Is it possible to keep making small additions and refinements to a complete work and thus keep it in “copyright” indefinitely?
There is constant change around the margins of the postcodelocation mappings, as buildings are built or abandoned. In theory the original 1959 set of postcode locations should be emerging from Crown Copyright … (checks watch)… but this feels like hairsplitting.
We currently have this prospect of a really full policy recognition of the social and economic value of open geodata. I’m still marvelling at it. Let’s hang on to the upside?
I have learned lots this evening, thanks both!
@Jo: regarding your last point about indefinite copyright from constant updating the basic answer is no, you don’t get an indefinite copyright this way. If you possess a an 1873 copy that would be out of copyright today no matter what updating had happened in the meantime. However, obviously if all you have is the copy today that is in copyright and without a way to extract only the 1873 content you’d be stuck with only the copyright version.