The early days of OpenStreetMap
We didn’t realise it at the time, but OpenStreetMap came from a unique moment in internet history.
It was still the time of the Old Internet. The time when you, or anyone, could upload a few pages to your own “webspace” or “homepage”, writing about what took your fancy. You didn’t have to worry about Facebook and Google as gatekeepers, or HTTPS, or your server being compromised within five minutes of turning on, or GDPR, or front-end frameworks, or any of the curses of the modern web.
And it was the brief flowering of Web 2.0… before it all went corporate. Sites like Flickr were young and independent. Every month saw the launch of some new crowdsourced knowledge base.
The origin myth of OSM is that of crusty GIS types, national mapping agencies, and cartographers sneering “you can’t map the world with volunteers”. But of course we could do it. We never doubted that we could. In 2004 it was obvious that you could build OSM.
It was so obvious that several people had the same idea at broadly the same time. Steve Coast started OpenStreetMap with Matt Amos and Tom Carden. Jo Walsh and Schuyler Erle started London Free Map (also 2004?). Nick Whitelegg started freemap (Oct 2004). I started Geowiki after hatching the idea with a bunch of friends from university (Sep 2002). We all coalesced around OSM because Steve went out and evangelised for it, speaking at endless LUGs and Dorkbots and hack weekends, whereas the rest of us essentially wanted to sit at home, hack on code or draw maps.
2004 was when the homegrown web met the participatory web, before it all got turned to shit by Facebook and Google. You wouldn’t start OSM in 2024: the nascent project would be squashed by Google, or founder under Reddit/HN users’ expectations of “why not good yet”, or flame out in some controversy over privacy or disputed borders. But we got through all of that in our early years, when the world was simpler.
What’s remarkable about OSM is not that it started, but that it thrived. Other OSM-adjacent projects from the same period didn’t. OpenGuides could have been the Time Out or Rough Guides to OSM’s A-Z, but petered out. OpenFlights too. Wikitravel got nobbled by a private buyer, re-emerged as Wikivoyage, but never really got traction.
Why did it thrive?
It didn’t, at first. In 2004 and 2005 OSM was still not much more than an idea. Events were still mainly Steve going out evangelising. The tech was an endless succession of false starts. The server was unreliable. There were 1000 users by December 2005, but they didn’t map much. It took until January 2006 before Britain’s (fairly few) motorways were mapped.
But a community was slowly building around the idea. People were talking and a common purpose was being forged. From that common purpose, a handful of people identified pieces in the jigsaw they could fit. Imi Scholz wrote JOSM (December 2005), Artem Pavlenko wrote Mapnik (winter 2006), I wrote Potlatch (March 2007). Glue code like the openstreetmap.org website itself (rewritten in Rails in May 2007), and mod_tile/renderd (December 2007), held it all together. Add Yahoo imagery (December 2006), and OSM was finally at the stage where you could draw a road and it would appear on a map. We had a conference in 2007… and something to talk about.
Since then, it’s been incremental. OSM today looks not much like OSM did in 2004, quite a lot like OSM in 2007. 9th August 2004 is OpenStreetMap’s official birthday, but 0-year olds don’t do very much. It took a few years before we learned how to walk.
Posted on Sunday 11 August 2024. Link.