Political philosophy in Open Street Map

Thea Clay made the killer point in Chris Osborne‘s “What’s wrong with OpenStreetMap” session (video here). It was even better than Mikel Maron‘s observation that people should agree with founder Steve Coast just a little bit less!

Foundation Board member Henk Hoff (a very nice-sounding chap) was describing the classic technocratic argument that in a “do-ocracy”, those who get on with doing things make decisions by default. Steve must have loved it, you just get on with useful stuff and avoid getting bogged down in pointless debates. Right?

Thea pointed out that in a community of tens of thousands, only a few hundred can “do” stuff like making amazing tools and creating useful maps from the tags they’ve invented. I’m in the bigger mass of people who want to contribute data, see lots of good uses and ways in which OpenStreetMap can improve, but lack the skills to “just get on and do” much that I’d like to with the data. I’ve been at it for five years.

I spend a lot of time involved in local groups trying to improve their town centre, grow food on their housing estate, represent tenants or establish a community-owned renewable energy company. They’ve come up with some great ideas for using OpenStreetMap, but none of us have the skills or money to “just do it”. I’ve tried, for two years, to get others to help with this with very little success. So is OpenStreetMap just not for us? Do I leave them to use Google Maps and miss out on the potential of such an amazing open platform?

Henk’s suggestion that the decisions of the “do-ers” are generally accepted is also just plain wrong. Nobody has ever been able to agree on the appropriate difference between marking a feature as a footway or a path, after years of debate. The same goes for wood/forest and countless other features.

That’s a pretty sad inditement as it makes for even more confusing map detail (the map key is absurdly long, and why should the average map user care if it’s a footway or a path?) that also makes applications like routing needlessly complicated.

Anarchy gives OpenStreetMap real energy and the space for innovation by the do-ocracy elite. But it leads to very inconsistent data and the exclusion of all but a technically hyper-literate few.

Most mappers come from countries run by democratic governance; they use software created by communities with clear and functional democratic structures like this, this and this; there’s a great study on Debian that drives the point home. Many mappers have engaged with the pseudo-democratic process of voting on new tagging proposals. I suspect that most mappers – if you canvassed beyond the extremely vocal elite – would support a more democratic approach to basic decisions about core tags, use of funding, etc.

That won’t stop people “just doing” what they want to do, but it will make the data more useful and the project more accessible for the vast majority of people outside of the technical elite.

Finally, just as we let people who prove their technical skill loose on the technology, why not let people with proven organisational and political skills loose on these problems? An audience member suggested OpenStreetMap was going through puberty. Well as much as teenagers hate to admit it, sometimes authority borne of experience and training is worth listening to.

For these reasons and more we need to hear much more from the likes of Thea and Mikel, and a lot less from techno-junkies like Steve.


  1. SteveC said:

    Hi Tom

    Not sure what planet you’re currently orbiting, but the whole foundation is democratic and as you surely know just had elections too.

    The problem with the save the trees brigade that you want to hear more from is that they don’t actually have any concrete suggestions (or more importantly, actions) to improve things. Like you they make vague statements about ‘more democracy’ or ‘get the government involved’.

    OpenStreetMap is a do-ocracy far beyond merely the technical aspects. If you want to help the foundation, you can do so, come join the calls. If you want to set up a tag system, do so. If you want to write better documentation, do so.

    The problem of course is that would all require actual work, whereas whinging about it is much easier. If you don’t want to do the work, then you can pay people to do it. If you want a feature added to potlatch, or a more shiny piece of documentation then you can find someone in the community or on odesk.com etc.

    Also, go read the first few chapters of Cryptonomicon for my views on you trying to bring down the ‘techno-elite’. We all got here by working hard and learning, I apologise we didn’t have the good grace to be mediocre and create a crappy system that nobody would use, and thus be eager for your help.

    16th July 2010
    • Tom Chance said:

      Steve, the Foundation’s remit is extremely limited.

      The funny thing is that when people get on with doing things related to governance, community outreach, etc. including those activities I describe in my post, you dismiss that work as “not concrete”.

      You arrogantly carry on as though your cobbled-together brand of anarchism is The One True Way and rather than engaging in an intelligent debate about the points I raised you posted a long general attack full of fallacious arguments.

      As you say, have fun.

      16th July 2010
  2. SteveC said:

    As someone who did community outreach for the entire lifespan of the project and invented mapping parties to say I undervalue it is ridiculous. Where did I undervalue it? As the person who set up OSMF with an all-voted board (and didn’t set myself up as grand dictator a-la wikipedia), where did I say governance was ‘not concrete’?

    I dare you to pick one thing you find wrong in OSM, state it publicly and then fix it. I double dare you.

    Or you could just write another post complaining that the technical people in OSM (who are volunteers) don’t do exactly what you tell them to.

    16th July 2010
  3. Jens Meilson said:

    Hearing the foundation’s chairman to say general discussion of the project’s destiny as ‘silly’ is pathetic. Only having value for individual actions is fine for a toy project, but naive for a project of such a size.

    My publicly stated problem is that I’m tired of seeing sensible discussion being shouted down as by a petulant child. In the long view I think Steve’s actions will harm the project more than most of the other criticisms people have of the site.

    So will you help me fix it by stepping down from being chairman?

    17th July 2010
    • Jonathan Bennett said:


      You have the option of standing against Steve at the next OSMF elections if you think you’ll make a better chairman. The membership will then get to decide, democratically, whether they agree with you.

      21st July 2010
  4. SteveC said:

    What actions are there but ones made by individuals?

    I’m still waiting for a suggestion guys. If you were chairman tomorrow what would you do? There have been vague suggestions that the software should be easier to use, but as we’ve seen on the mailing list the people who write it are volunteers and hard to change course. ‘More democracy’ is kind of hard in OSMF, what could you make more democratic when every single phone call is open and everything is held to a vote?

    17th July 2010
    • Jens Meilson said:

      “What actions are there but ones made by individuals?”

      So this thing ‘teamwork’ is really new to you?

      An example. I think some of the more personel anarchy tagging is wrong, and I think people should work more closely to a better schema. All the blah_blah=yes makes me crazy. Sure I can make a bot to fix it. I will maybe get two beers. But I will also get many many crazy people.

      So JFDI and individual action is not good enough. It needs people to talk about it. Is OSM going to be good data in a good model understandable by machine? Is it lots of personal tagging styles that are never quite the same? Talk you call ‘ridiculous’ and ‘kind of silly’.

      If I was chairman tomorrow, I would stop saying it was defined by JFDI. I am thinking this project is too big and too important for the chairman to be only shouting about the design of the homepage and details of that level.

      Or maybe it will stay being a sort of hobby hacker project and will be a different project or fork that changes the world.

      Sorry Steve, you were a great guy to get the project started. Now it is involving so many people, it needs a different type of leader.

      17th July 2010
  5. Language aside, this thread raises a very good point about what I call a “societe a deux vitesses” two speed society (2SS) of ‘haves’ and ‘have nots’:
    – those who have both time and tech ability to create great work like OSM (and its the very time and effort distributed among many that makes OSM)
    – those whose commitments allow neither of the above, thus mainly consumers, but who also wish to have a say even tho they don’t build it
    [I myself am weaving a treacherous path in the graveyard of geo-startups]

    Is a consumer not allowed to offer critique (observation not criticism)? Consumer are not whinging {favorite persona}! Rather: are those not vested in the technology, not the most likley to offer suggestions and improvements? I see the widespread acceptance of GI tech just crossing Moore’s chasm: consumers on the upside after it, and techs on the downside before it.

    And speaking for myself at least, the longer I take to cross that chasm to join consumers, the longer I perpetuate that two-speed society… And doing that is illusory, as consumers will rule the day sooner rather than later. So I admire OSM’s work, and imagine the frustration in getting constructive feedback. But let’s use all means available – blogs, crowdsourcing, meetups, unconferences etc. that we already use in fact – to move ahead.

    The real issue IMHO is “where’s the money” in neogrography? what is the business model that will allow creators and consumers to build a single sustainable path, not a conforntational two speed society?

    17th July 2010
  6. I agree with you that the whole tag=value thing can be crazy-making when there are two “tag” names for the same concept, e.g. “path” and “trail” and “footway”. I used t’ rail against it, just as you have. But I finally came to a realization: it doesn’t matter exactly how people tag AS LONG AS their meaning is documented in the wiki. Voting isn’t going to solve anything; you’ve got to get people actually using the tags that you voted on … and how does that happen? They’re in the wiki. So why bother voting; just put your meaning into the wiki. If it conflicts with someone else’s meaning, then figure out how to resolve the conflict (and document it in the wiki), or else pick a different tag and document THAT.

    18th July 2010
  7. Frank said:

    Hello Tom Chance and others!

    I’ve wittnessed those fights about tags and sometimes I’ve been swearing about that a particular tag doesn’t work ( either on the online maps or my Garmin GPS device ). So I feels a kinda of urge to make something of my own, Maybe it would be possible to make a “Mapsuite” that contains many functions.

    I can just guess that lots of thge code can be reused towards different applications like an editor, slippymap tile renderer, Garmin IMG exporter, reverse geolocation and so on.

    But I don’t really know in what end to start – I’m kinda bad of organisation…

    I did not watch the whole video, the sound was to bad. Does it exist a transcript of the speech or an other video?

    20th July 2010
  8. I think you’re fundamentally misunderstanding both “do-ocracy” and the nature of a (pseudo-)charitable, volunteer-driven project.

    You don’t need to be a “techno-junkie” to “do”. Andy Robinson invented the entire tagging system; he’s a tunnel engineer with no programming knowledge. Steve Chilton produces one of the best-looking webmaps around; he’s a traditionally trained cartographer, not a techno-junkie in any way. And so on.

    But to make a difference, you always need to “do”.

    Apache and Debian are not significantly different. From the Wikipedia pages you reference: “Together, the Developers may make binding general decisions”; “Membership [of] the foundation is granted only to volunteers who have actively contributed to Apache projects”.

    They share the same fundamental tenet with OSM: control is exercised by those who “do” (here, “Developers” and “volunteers who have actively contributed”). That’s because they’re all volunteer organisations, and volunteers will only work on what they want to work on. That’s what “volunteer” means.

    So you might have a great idea for how your local renewable energy company could use a webmap, but neither you or the energy company are capable of doing it. You have two options: either convince a volunteer that they _want_ to do it, or pay a hired hand. But you can’t force a volunteer to work on a project that doesn’t interest them. The “countries run by democratic governance” that you cite largely work on the hired-hand model: civil servants don’t work for free.

    OSM has a particular challenge in that there are two types of “do”-ers: the few who develop (in the broadest sense), and the many who map. Happily, most of the developers are also mappers, so the code largely evolves in a way that keeps the mappers happy. (It’s also worth noting that the developers are, very often, altruistic. When I say “volunteers only work on what they want to work on”, that doesn’t mean “what they want to work on” is necessarily something of interest to them and them only. I want to, and do, write a user-friendly editor, even though I’m perfectly capable of using a user-hostile one for my own mapping.)

    OSM has many problems. There aren’t enough developers; we need more. And yes, perhaps if we get them, they will choose to organise themselves through a democratic process as the Debian and Apache developers have done. Secondly, no-one from the mapper community has yet taken on the urgently needed task of documentation, which would greatly help to reduce tag confusion (and more).

    But neither of these will be solved by imposing a new structure on the existing community.

    (Tagging is largely a separate issue and not one I’ll go into here.)

    21st July 2010
    • Tom Chance said:

      Hi Richard,

      It’s very easy to look at people such as Andy, Steve and myself and conclude that there are plenty of people deeply involved with the project who aren’t capable of doing the core development work. But that’s not really the point. Just being able to join the mailing lists and use JOSM without a lot of hands-on help puts us in a very small minority of techie people.

      Anyone who has tried outreach to non-technical people or communities faces the question: how can we make OSM more usable for them? Documentation, as you point out, is a big part of the problem, but not the end of it.

      For Apache and Debian, it is OK to be more led by developer consensus because their products are aimed at very technically capable people. For KDE, the only free software project I have been deeply involved in for any period of time to be able to speak about, there is a very different attitude. A lot of care and discussion time is put into considering the next market segment the community might conquer, and the various unmet needs that the developers might want to fulfill. That is facilitated by the governance body and companies who invest in KDE working collaboratively to identify and meet those needs, paying for bounties and employees where concerted efforts to give direction to volunteers isn’t enough.

      Can the OSMF function in this way, attracing and pooling significant funds and using a transparent and meaningful process to identify priorities for potential growth segments?

      Questions about tagging and quality control have similar analogues in mature free software projects, typically discussed and determined by democratic bodies. It would take a relatively small effort to identify some low hanging fruit that would make OSM more attractive and usable to big segments, but the governance and culture of OSM militates against that at the moment.

      I’m not saying that I want to be able to force a volunteer to work on my pet project, but a basic feature of a do-ocracy is that it will ultimately only satisfy the needs of those who can do.

      In the case of OSM, that means that the slippy map will satisfy most who find it; the editing process will satisfy time-rich and skills-rich people, or those supported by people who put effort into outreach; the evolution of tagging and the like will satisfy the extremely time rich or the casually individualistic; and the direction of development will satisfy the extremely techy.

      I think OSM has the potential to be much more than this, but a great many companies, government bodies, voluntary organisations and interested but skills/time-poor individuals will pass us by unless we have a grown up discussion about how governance can help shape OSM’s future.

      To echo my post’s conclusion, a little less of Steve’s petulant histrionics (as demonstrated in his comments on this post) and a little more space given to seriously discussing these issues and acknowledging the implications of different models would be a good thing.

      Trying, as Steve as done, to dismiss my arguments with an ad hominem “get on and do it” (despite having contributed a huge amount of data over five years, organised mapping parties and technical workshops, introduced a range of organisations and government officials to OSM and more) is just avoiding a debate about the pros and cons of the do-ocracy.

      21st July 2010
      • Tom Hughes said:

        I think the point is that our products are aimed at technical people – we provide the data and tools etc that allow other people to build the kind of end user tools and web sites that will enable the kind of people you are dealing with to achieve what they want.

        22nd July 2010
      • Andy Allan said:

        Tom, there’s a big thing you keep missing here: there are many, many developers who know fine well what the issues are, and we’re working on them. We have lists of these user interaction bugs and we fix them.

        You, and many others like you, don’t know about these tasks and you keep assuming that simply because you don’t personally know about them then they can’t possibly exist. You’re wrong, you just aren’t aware of them. Have some humility and assume that the developers aren’t ignoramuses.

        You say that a “do-ocracy” only satisfies the needs of those who can do – which is a) insulting and b) clearly wrong. If this was the case, all of the editors could only be used by programmers, but we have made them usable by others and will keep improving them regardless of what else happens or which aspersions you cast.

        Finally, you’ll find continuing hostility if you continue to use phrases like “grown-up conversation”. We’re not kids, we’re extremely dedicated professionals – some of the best in the world – and to infer that we’re otherwise is completely insulting.

        23rd July 2010
  9. SteveC said:

    What’s the alternative Tom?

    It’s really simple – if it’s not a do-ocracy, then who is going to do the work?

    Nobody is dismissing your mapping efforts, it’s just these vague points about the documentation should better or JOSM should be easier are well known. We all know it.

    They won’t fix themselves, so who do you propose fixes it if the coders are slow to do it and you don’t want to do it? Do you want us to force people to work on it? Do you want to pay people? Where would the money come from?

    21st July 2010
    • Tom Chance said:


      I think you’ve misunderstood a few things.

      First of all, there are many “issues” that aren’t to do with development, or that can’t be wholly fixed with code.

      Second, I know very well how dedicated and conscientious many of the developers are, not least from having sat in the same room as you on a few occasions. I know you are seeking to address many technical issues around usability and am grateful for all the hard work on code, map data, documentation and more put in by others in the community.

      Third, I didn’t make so bald a statement as OSM “only satisfies the needs of those who can do”, but pointed out that those who can “do” generally lead the direction of the project and define what needs to be done.

      That is really the crux. Does OSM remain a project generally led by a technical and time-rich elite with a belief in the public good, as Tom Hughes suggests? Or does OSM grow out to be led by a wider group of enthisiasts, finding ways for people with other skills and from other worlds besides open source geotech become involved? If the latter, then I believe it can also become more useful to other communities, rather than leaving it to others to fork or attempt to bolt on to OSM’s anarchic processes.

      Finally, I think the comments on this post make clear that every contributor here is quite capable of a grown up online conversation with the exception of Steve Coast.

      23rd July 2010
  10. Harry Wood said:

    Just on a specific point. In your original blog post…

    “Henk’s suggestion that the decisions of the ‘do-ers’ are generally accepted is also just plain wrong. Nobody has ever been able to agree on the appropriate difference between marking a feature as a footway or a path, after years of debate.”

    Here and in the discussions, I think we’re making a mistake by conflating organisation/steering of OpenStreetMap with organisation/decisions/documentation about tags.

    Open tagging is a frustrating challenge causing a lot of problems for new editors and for downstream data users, but it’s also a strength of the project (arguably). We know the arguments around that. Suffice to say it’s a big problem, perhaps the biggest for OpenStreetMap, and I don’t believe it has a simple solution. We try to solve it with committees or democratic processes or whatever…

    But let’s debate that later/elsewhere. I think we can treat tag decisions as a very separate from other OpenStreetMap steering decision making. Again, for OpenStreetMap steering we could have committees or democratic processes, do-ocracy or whatever. They don’t need to be the same organisational structures as those for tagging.

    You’ll notice that the people steering OpenStreetMap (the key “do-er”s) already have this separation clear in their heads. It’s not so much that their decision on footway or a path has not been respected. They’re mostly staying out of it.

    22nd July 2010
  11. David Martin said:

    Tagging is an interesting area. As someone who is very keen on OSM and data reuse, the current tagging structure seems to be ill thought out.

    I will qualify that. At present the current tags seem to be a blend of keyword/value or keyword/flag pairs that are semantically poorly related. Looking at the work of the semantic web community and (my personal experience) the biological annotation community, there is a very strong case for reengineering the tag structure as a series of ontologies. Each tag then has a specific name, a specific definition, and more importantly a specific relationship to other tags through the standard ‘part of’ and ‘kind of’ relationships.

    Tags would then be computable as properties would inherit from parental nodes where not explicitly stated. Suggestions for new nodes in the hierarchy could be easily accommodated, the number of tags needed for any way would reduce (as a lot of properties would be implicit) and life would be easier for those writing eg. routing software.

    22nd July 2010

Comments are closed.