Today in our ongoing interview series with Open Geo thought leaders we had the chance to chat with Tom MacWright, creator of mapschool.io (amongst much else). 

1. Who are you and what do you do? What got you into the open geo world?

I’m Tom MacWright: on the internet, I mostly write software and prose. A lot of my work has something to do with maps, since I work at Mapbox, a company that builds modern geo technology. Maps are a fun subject since they touch a lot of other areas, like geometry, art, and information theory. But I’m not a GIS guy or a lifelong lover of atlases: I arrived here haphazardly and mostly just like to learn.

2. You created mapschool.io What was the motivation and what has been the response been?

I really like teaching and explaining: before mapschool I wrote things about using math for geometry, explaining tricky concepts like k-means clustering or raytracing with simple English and room to tinker.

mapschool came about when I was looking at existing GIS education and thinking about which concepts are important for modern GIS work. Though my experience with it is minor - a few classes and a lot of reading - traditional GIS education has been a disappointment. It’s inappropriately focused on pedantic details of technology and teaches a lot of recently-developed and possibly-incorrect practices as if they were eternal natural laws. I wanted people to explore maps from a creative, conceptual standpoint, so I wrote what I wanted to read.

The response has been great - I know a few people are using it in classes, and a lot of people have contributed improvements and expansions of certain sections based on their deep experience and knowledge. Translations have been incredible: it’s now in English, Italian, Spanish, Swedish, and Japanese. I still really want to develop it into a printed zine and redraw the illustrations in black and white, but that’ll take a little while.

3. You’re very prolific on github. Some of open geo software projects take off, others never seem to hit critical mass. What’s the deciding factor, and what advice do you have for aspiring geo hackers or those currently sitting on the opensource sidelines?

Before discussing factors for success, I’ll just put this out there: ever since GitHub introduced Stars, Forks, and Followers, people including myself have worried about why a number wasn’t high enough or why some person’s stupid jQuery plugin is doing so much better than the thing I wrote. Just like every other aggregation, like Twitter followers or Klout scores, it’s an arbitrary and ultimately meaningless thing to care about.

That said. I’ve poured a lot of time into writing projects with insanely varied rates of success, and have a few observations.

  • Package and explain. After completing a big project, what it does and how seems obvious and even boring. Don’t fall into the trap of assuming context other people don’t have: explain your project from the ground up.
  • Contribute. Submitting a pull request to an established project with a great maintainer will usually net you great feedback and a nascent connection with other people.
  • Publish early. Starting off, your code won’t be good. Luckily, nerds relish the opportunity to tell people they’re wrong: let them, and try to internalize as much as possible about what you can do better next time.
  • Learn to read. Experienced developers don’t trust documentation, they always read the source if something goes wrong. Whatever you’re using or developing or extending on, have a source checkout of it, and use that to understand.

4. You live in Washington DC. As an outside observer it feels like the open geo scene there is exploding with more and more interesting collaborations between government, NGOs, companies. Tell us a bit about the scene.

There are some cool things happening, like the National Park Service’s collaboration with OSM) and the FCC’s Broadband Map. I used to be more plugged into the NGO scene when Mapbox was Development Seed, but since then it’s less on the radar. I’m really optimistic about 18f’s attempt to modernize government technology thinking.

5. OpenStreetMap recently turned 10 years old, what’s your guess as to where the project will be in 10 years?

10 years is a long time.

In terms of technology, OSM outgrew their infrastructure years ago and are running on borrowed time. There’ll be a crunch point soon where they have to grapple with a database that doesn’t fit on any single computer, and it’ll have to make the jump, or fail.

On a ‘map’ level, the big question will be how OSM mixes with other data sources. The first wave of imports, mostly from governments, had some big successes but also a lot of hiccups. The second wave, sensor data and satellite-imagery-derived data, will be different. New data sources will be updated faster, but more importantly, will require contributors & technology to grapple with the mixing of their homegrown organic traces with data created as an ambient byproduct or as an algorithm output.

But the bigger factor is culture: OSM is more about people than it is about technology. People are finally starting to realize that online communities have culture, whether it’s Reddit’s remorseless aggression  or Wikipedia’s complex social and identity structure. As Anil Dash eloquently wrote, If Your Website is Full of Assholes, It’s Your Fault.

Right now OSM falls in the middle: there are great people, but also a lot of disruptive and antagonist behavior that goes unchecked. The project desperately needs new people, as contributors, as the beginning of diversity, and to build the technology core. I don’t think that it’s doing nearly enough to accommodate newcomers and I think the culture repels them. It has certainly repelled me at times. More than anything else, I want to see OSM grow culturally over 10 years - I want OSM to be friendly.

Many thanks Tom, for taking the time and all the great software. Totally agree with you on the need to open OSM culturally, a point I tried to flesh out in a recent summary of this summer’s SOTM-EU conference. There are some encouraging signs of progress but of course a long way still to go. 

You can see all the Open Geo interviews here.