The folks over at theTagzania blog just announced this:
During the course of history, we have used several types of tests to tell if something is what it seems to be, or to check if it is up to expectations.
Now we use industrial stress tests, quality assurance checklists and software metrics. We put dummies inside of cars and crash them into concrete walls at full speed, to check if the dummy gets out in one piece or not; then we decide whether the car is safe enough. Centuries ago, the church used the trial by water ordeal: tie a rock to a woman’s leg and throw her into a lake; if she drowns, she’s not a witch; if she comes back to the surface, she is a witch and has to die by burning.
Free software also has tests, which are scientifically more accurate than the trial by water, as one has to tell apart free-as-in-free-beer licenses from free-as-in-free-speech ones. During the development of free software licenses (such as the GPL) in the 90s, there are some tests which are well-known amongst free software advocates: the “desert island test”, the “dissident test” and the “tentacles of evil test”. These test were an integral part of the Debian Free Software Guidelines (or DFSG for short). The DFSG are just a definition, and are not easy to explain to a layman: it’s better to check against some use cases.
For example, the “desert island test” assumes a castaway in a desert island with a solar-powered laptop, and some software in it. For this software to be free, the castaway has to not be legally forced to distribute any changes made to that software, as he just can’t. Technically, this test checks that the license asks for source code redistribution only when the binaries are distributed. This tests makes it easier to understand whether a license complies with the DFSG, rather than checking against the DFSG themselves.
And what about geographical information? Is there any test that lets us know whether or not the information from National Mapping Agencies (NMAs) or Spatial Data Infrastructures (SDIs) is really available under conditions that allow the citizenship to make the most out of them?
In order to know if a given set of geographical information can be considered free (as in free speech) or open or “libre”, it can be checked against the Open Knowledge Definition (OKF), just as free software can be checked against the DFSG. However, the OKF can be a bit dense for laymen.
There is no easy test to know if a set of geodata is free/open/libre or not.
I hereby propose the Cake Test.
What is the Cake Test? Easy: A set of geodata, or a map, is libre only if somebody can give you a cake with that map on top, as a present.
Cakes are empirical proof that most the data in most SDIs cannot be used freely, because of the licensing terms of the SDIs. And they are an empirical proof that attendants to the latest spanish SDI conference could taste themselves.
Even if maps, or geodata, are published in a web site for free (at no cost), it doesn’t mean that a cake can be made with them. Some examples of technical or legal obstacles for the “cakefication” of geodata are:
- Not being able to download the geodata to a computer.
- Not being able to copy the geodata to a different medium, or not being able to redistribute it. In order to make the cake, a bakery needs a CD with the images, or an e-mail with them.
- Not being able to use the maps for profit (AKA “commercial use”), even if the person giving the cake away is not making any profit (AKA “indirect commercial use”). The one giving the cake does not make money, but the bakery does.
- The obligation to sign a license (or such) for commercial use. Do you really expect people to go to a bakery and say “hey, I’d like a cake, but you need to sign this commercial-use-of-geodata-license thing first”?
- The obligation to notify any usage of the geodata prior to using it. If we have to tell we’re making a cake, it wouldn’t be a surprise, would it?
A set of geodata must comply with lots of conditions in order to bake a cake with it, and may seem complex when applied to a geodata license (and the Cake Test is just a neccesary condition, not a sufficient one), but the goal of the Cake Test is very simple:
If a layman can’t decide whether he can or can’t decive if he is able to give away a cake as a present, or plainly isn’t able to, then the geodata cannot be used freely, is not free, and is not libre.
On the other hand, if some day someone gives a NMA or SDI a cake as a present, then that NMA or SDI is on the right track for information reuse. And the day that happens, they’ll probably throw a party, as they’d already have the cake.
The Cake Test also illustrates the concept of long tail.
Usually SDIs and geoportals are built in order to the big geodata consumers to generate more profit, or lower their costs. This is the green section of the graph: quite few data consumers, but consuming a lot of data involving a lot of money. They are government agencies, and big corporations and projects. On the other end of the graph, in the yellow section, would be them cakes.
Obviously, cartography applied on top of pastry products is just a marginal contribution to a country’s Gross Income. But it’s not a null contribution and, most probably, no one thought about it before.
The long tail is full of cakes and other marvellous things that haven’t been invented yet. How many new uses for geodata are there to be discovered? How long is cartography’s long tail? The only certain thing is that, in order to know that, free use of geodata has to be encouraged.
(This is a translation of an article originally in Spanish, available here)
These great commercials have got Verizon in a bit of trouble
Read the slashdot story.
I was pleased to meet so many folks working at the city of Toronto and at the province of Ontario who showed so much interest in Open Data.Â There were many great conversations going on, from the exhibition floor at the city hall rotunda to the mixer at a local pub later.Â All of these are great signs of a new open-awareness at the city and I see it as overwhelmingly positive.
Being new to the world of Open, the city wanted some feedback regarding for what applications people would use this newly available data.Â As Toronto Transit Commission data, addressing data and road centrelines were all released I thought immediately of the travel planner for London from mySociety.
I had that chance to talk to many folks about OpenStreetMap through the course of the day and I was pleased to share my enthusiasm for a travel planner like this using the Toronto data.
The data we have now is imperfect but rather than critiquing the quality of the dance steps of this bear, let’s marvel that Toronto released open data at all.Â Most of the data sets grew up in separate silos in Toronto departments.Â The folks at the city are as new to these data sets from other departments as we are.Â They’ll get used to working with each other in an open environment and that will move them to more of the open tools, standards and practices that we take for granted.Â I’m sure we’ll see a bug tracker soon.Â We’ll see increased use of open formats rather than proprietary lowest-common-denominators.
Bravo, Mayor Miller, for recognizing the benefits of Open.Â Bravo, Mark Surman for challenging Toronto to become a city that thinks like the web,Â This is an important step along that way.
The Geographers, GIS buffs and spatial analysts at UCL have always been big supporters of OpenStreetMap.Â Over the last few years more and more of their attention has been turning to the study of crowd sourcing as a way of creating geographic information.Â One great example of applying academic research to OSM is Muki Haklay’s 2008 OSM Quality Evaluation work, in which Muki compared OSM data to data sets produced by the UK Ordnance Survey (OS) – the UK Government body charged with mapping the UK.Â The OS has a fierce reputation for for producing some of the most accurate and most detailed maps in the world, so its impressive to hear that: “The positional accuracy [of OSM] is about 6 metres, which is expected for the data collection methods that are used in OSM. The comparison of motorways shows about 80% overlap between OSM and OS…”.Â Of course OSM moves quickly – there are now more than 100,000 OSMers around the world compared with 35,000 this time last year – so when I spoke to Muki earlier this year I was excited to hear that there is more OSM data analysis on the way.Â Watch this space and in the meantime, take a look at this recent presentation.
A team of UCL Geographers are taking their interests in crowd sourcing on the road, speaking at the American Association of Geographers annual meeting in Las Vegas next week (22nd – 27th March).Â Papers include: “Neogeography: Crowdsourcing and Mapping for Masses” – something that should be of interest to any OSMers in the area.Â Another interesting looking title comes from TeleAtlas Chief Scientist Don Cooke talking about “Neogeography and Crowdsourcing: the View from a Walled Garden“.Â I got the opportunity to talk to Don after giving a presentation about OSM and crowdsourcing at last year’s Where2.0 conference.Â He was a big fan of OSM and the crowd sourcing model.Â One thing’s for sure – the often distainfuly labelled “paeleo” generation are not going to roll over and die.Â Guys like Don Cooke or the UCL geographers are veterans of an industry that has created vast data sets, empowered millions of people to make better decisions, as well as creating companies with multi-billion dollar price tags.
How will OpenStreetMap react when the “walled garden” approach to crowd sourcing puts the power to edit and create maps in the hands of everyone with a mobile phone or a sat-nav?Â To make sure that the open approach to crowd sourcing keeps on producing data sets that can be favourably compared with those of the walled gardens we need to keep one step ahead.Â Learn from the acheivements and mistakes of companies like Tom-Tom, who are embracing crowd sourcing.Â Openess alone isn’t going to build a free world map.Â We need to expand the reach of OpenStreetMap – attract 900,000 more mappers in more places of the world and help them produce better maps.
One of OpenStreetMap’s greatest advantages is that we don’t just give you a beautiful draggable map – we give you the data, so you can do what you like with it. Well, this weekend, that just got a whole lot easier.
OpenStreetMap now has an ‘Export’ tab, joining ‘View’ and ‘Edit’ at the top of the screen. It gives you an instant way to get the map data in a format you want.
And this is just the start – our mailing lists are already buzzing with possibilities for new formats, such as Adobe Illustrator for cartographers, or shapefiles for GIS professionals.
With this new feature, the difference between OpenStreetMap and the “corporate” mapping sites becomes a whole load clearer. Other mapping sites’ agreements with their data providers (such as Navteq, TeleAtlas or national mapping agencies) simply wouldn’t allow them to give the data out like this. With OSM, we actively encourage it!
The work behind this was done by Tom Hughes, winning OSM’s coveted Lolcat of Awesomeness developer award for the fifth time.
The State of the Map 2008 international OpenStreetMap conference website and blog is now live and the call for participation/papers has gone out. Speaker slots are limited and like last year are bound to fill up quickly, so send us your abstracts!
The conference is being held at the Kilmurry Lodge Hotel in Limerick, Republic of Ireland on 12th and 13th July 2008.
Critics of OpenStreetMap have in the past used various arguments for why they believe the project can never be a success. Often quoted is the concern about data accuracy and completeness and another is about how changes in our world are noted and managed. Well it’s clear we needn’t worry too much as some great work by OSM contributor Dair Grant clearly shows.
Dair made a comprehensive and detailed comparison of his completed map for Haywards Heath, Sussex with that of the TeleAtlas derived Google Map and found 89 apparent differences, that’s an astounding number for a small town.