Systeme D

September 29, 2009

The neogeographers are coming for your children

Paleo or neo – which are you?

I have a degree in Anglo-Saxon Norse & Celtic, I play the church organ, I work for a printed magazine about an 18th century technology, and I have several shelves of Ordnance Survey maps (excluding the unfolded ones, which are elsewhere):

barn_shelves

Which I guess makes me a paleotard. Glad we’ve cleared that one up.

This was originally going to be a nice, friendly, fuzzy post about how ‘paleotard’, ‘neogeographer’, ‘freetard’ and the like don’t actually mean anything. How it’s quite possible to love the Ordnance Survey’s beautiful maps yet still think that maybe, maybe, their licensing isn’t 100% perfect. How it’s possible to be an OpenStreetMap activist yet still happily use Landrangers and Explorers, the AA Close-Up Atlas, Multimap, etc. etc. etc.

Then I saw Gary Gale’s excellent rant from the AGI Geocommunity event (where paleo vs neo appears to have been the main topic of conversation). He makes the same point. We are, to misquote Moby, all made of tards.

we_are_all_made_of_tards

Also, I figured out something more interesting to say.

Let’s talk about scale for a moment. That’s a lovely old-fashioned concept, isn’t it? When was the last time you heard a freetard talk about a 1:50,000 map, or better still, one inch to the mile? I must be a paleo after all.

But scale is really, really interesting.

Ordnance Survey makes most of its money from very large scale (i.e. very detailed) mapping. That’s what government, utilities, property developers and insurance companies pay for. OS’s accounts are frustratingly vague by comparison to, say, British Waterways’, and when you’re holding up BW as an example of How To Do Transparency Right then you know something’s amiss. But a simple read-through of the space allocated to small-scale and large-scale products in the 08-09 Annual Report demonstrates that large-scale is where the money is.

Yet the majority of ‘neo’ stuff only needs small-scale data, and that’s especially true of public interest projects from CycleStreets to MySociety – the sort of thing Governments of any stripe should be encouraging. Even OpenStreetMap, which has an ever-advancing army of buildingtards doing bonkers micro-mapping stuff (hey, London guys, you should sort out your road numbers first. They suck), acknowledges this by only mapping road centrelines. OS MasterMap it ain’t.

Ordnance Survey clearly knows this but won’t admit it. It does claim to be adopting a “hybrid” model, but at heart, there’s no movement from: our data, our terms, you use it only as we see fit, we own everything.

So the notion that the data should be let out without these terms keeps gaining ground, from the Guardian’s Free Our Data campaign to the very learned Power of Information Review report.

Interestingly, though, OpenStreetMap contributors aren’t pushing for it (which rather kiboshes an otherwise entertaining rant from the AGI). That, I think, is because they’re interested in the small-scale data, too. OSM will have a complete small-scale map of the UK in a few years’ time, and as the backlash against data imports shows, they’re having fun making it. There are now hundreds of UK towns and villages where OSM is more complete and up-to-date than OS Landranger and Streetview. Even the most trivial analysis of OSM’s growth rate shows a quick trajectory towards national coverage.

So if the large-scale stuff doesn’t matter, and OSM is advancing towards small-scale completion, why am I bothering to write any of this?

Because, I guess, I really am a paleotard – well, a little. Like lots of people, I love Ordnance Survey maps: that’s why I collected a whole edition of them. I don’t want to see the Landranger relegated to being an expensive niche product used only by those who don’t have satnavs. I don’t want to see a generation of map users believing that there’s nothing better than rounded-end Internet cartography.

But that’s exactly what’s happening. I lost count of the number of head-in-hands moments during the Ordnance Survey presentations at the Society of Cartographers’ Summer School. “Extremely popular sites nick all the tiles” was the most quotable. Geovation was the most tragic – a whole initiative devoted to encouraging ideas, when there’s no shortage of ideas or even of coders, only of data. The bar chat about OpenSpace usage limits, and Landrangers-for-iPhone that are more expensive than the paper maps – really – demolished any pleas of “listen to us, we’re trying to work with you” from the presenters. It was one long exposition of the fact that the OS still doesn’t get it.

brock_os_disabled_kml

I was reminded of a dyed-in-the-wool BW section engineer telling his boss he couldn’t cope with the new (1990s) way of doing things. “I’ve spent all my life trying to keep people away from the canals. Now you’re telling me you want to get as many people onto them as possible.” Ordnance Survey is still at the “trying to keep people away from the data” stage.

So a whole generation of people is growing up without the Ordnance Survey. I would love to see statistics on the average number of OS maps bought by 25-year olds today, compared to 10 and 20 years ago. The (yes) epic fail that is OpenSpace is the most obvious example: no-one can ever name more than one site made with it, and even that site (Where’s The Path) is essentially a comparison tool for Google aerial imagery.

This is why OS must open up its small-scale data. Not because of any doctrine that facts should be free. Nor because we need the data – we do, but OSM is filling that niche. But because today’s maps and tools are being built without OS data. Today we’ve had the announcement that Flickr photos can now be linked directly to OSM data. If OS had opened its data three years ago, maybe we’d have seen Flickr linkage to OS IDs instead.

The Ordnance Survey, simply put, is heading into irrelevance.

All it can do is release small-scale data, unencumbered by anything but attribution, and hope it’s not too late. Right now it probably isn’t. In two years’ time it will be. Off the top of my head:

  • Street names and geometries
  • Public rights of way
  • Postcode->co-ordinate lookup
  • Administrative boundary data, including electoral divisions, Access Land etc.
  • Major natural feature names and geometries (peaks, main rivers/water areas)

Yes, it would result in some initial loss of income for OS. I’ve already explained the long-term benefits for OS in ensuring its continued relevance. Many will also make the case that the benefits to the wider economy will more than make up for it.

There are two more political considerations.

One is obvious. Releasing small-scale data would remove the need to stop wasting money on nonsense such as OpenSpace, Geovation, lawyers’ letters to Google (and others) about derived data, and today’s five-minute wonder, OS VectorMap Local which has thus far failed to set the world on fire.

The second is less obvious. Ordnance Survey is regularly talked about as a candidate for privatisation, helping to fill that modest hole in the public finances. (It’s usually somewhere in the list between BW and the Royal Mint.) Let’s say the small-scale business dies and OS is no longer a household name, though still earning much the same from its large-scale data sales to utilities and local authorities. All of a sudden, the ‘national interest’ argument for keeping it in the public sector is much smaller – and the political fallout from selling it much less. And I don’t get the impression that OS wants to be privatised.

But hey, I guess the Stasi is pretty much immune from privatisation.


4 Responses to “The neogeographers are coming for your children”

  1. I’m not sure where you draw the line between large scale and small scale, but OSM seems to be at least trending towards larger scale stuff, even if it’s happening slowly (and with a bit of friction, as evidenced by the ’should bridges be independent of their ways’ debate).

    I agree, though, that the OS could quickly jump ahead if it opened up access to some of its data – everything from streets to buildings to paths to natural features.

    One of the big problems that the OS can never overcome though, is that the internet is global, and requires maps and map data that span the world (or at least the places where internet-using people live). Ordnance Survey is always going to be UK-only, and so that limits a lot of its use.

  2. I’m glad someone’s keeping their old 3″ disks – I got rid of all mine.

    Anyway, for all the great stuff that OSM is doing, it’s still stuck with being a *street* map as far as I can see. It lacks essential details like contours, terrain, etc. And getting all of *that* data for free without the OS is a *hard* problem. So I guess I’m a paleo-geographer too.

  3. Martin says:

    @David Cantrell:

    OSM doesn’t have contours, but that can be added by using the SRTM dataset from NASA; e.g. see an example use at:
    http://www.opencyclemap.org/
    or
    http://croydon.cyclestreets.net/
    to give a very immediate example (and pan about a bit).

    (Whether that gives sufficient resolution for the uses that OS data is put to, I can’t say.)

  4. [...] Survey. I wrote here previously that OS’s best chance of surviving was to open up street name/geometries, boundaries, postcodes, [...]

Leave a Reply