A Critique of OpenStreetMap
Justin O’Bierne’s critiques of OpenStreetMap’s base map of North America — essentially, the first thing first-time visitors to OSM would encounter: the Mapnik layer — has apparently been stirring up a bit of controversy in the OSM community. Part one looks at the Mapnik UI, labels, boundaries, and roads; part two looks at city labels; and there is more to come. It’s a useful process, but it seems to have stirred up more than a few angry bees.
Some of this stuff can be dealt with by any OSM contributor, but other stuff is intrinsic to the Mapnik layer. Many of the problems Justin identifies are a result of applying British standards to North American roads and cities. (The definitions of “town” and “city” are quite different in Canada, for example, where cities can have fewer than 10,000 people; and figuring out which roads are trunk, primary and secondary is, from my own experience, a bit of a challenge.)
Open-source projects don’t generally do user-interface stuff very well, as anyone familiar with Linux-on-the-desktop efforts can tell you: too many people conflate user-interface questions with “making it pretty,” when in fact it’s all about ease of use. Whether OSM is there largely to provide basemap data for third-party projects (e.g., MapQuest Open) or whether you actually expect people to use Mapnik for their online mapping needs, I don’t think it’s heresy to say that much more work is needed here. Even map contributors need a decent UI, and if you want more people to contribute to OSM (let me give you a hint: yes you do) then an accessible — and yes, attractive — map will help draw more people in.
And, considering how many third-party apps use Mapnik tiles directly — I’m thinking of all the apps on my iPad that use OpenStreetMap — there really is a strong case for making Mapnik better.