Skip to content

Instantly share code, notes, and snippets.

@tmcw
Last active May 28, 2018 21:19
Show Gist options
  • Save tmcw/e23284b2319fbbaf92707c03fe88ae4f to your computer and use it in GitHub Desktop.
Save tmcw/e23284b2319fbbaf92707c03fe88ae4f to your computer and use it in GitHub Desktop.

Geojson.net

Geojson.net will be a replacement for geojson.io, the simple editor for map data. In most ways, it has the same intent, goals, limitations, and ideas of geojson.io - they're both projects of mine.

I created geojson.io as a side project in 2013, and it thrived for a few years, as it simply solved the problem of previewing, modifying, and creating map data. I think it benefited from simplicity and unity of thought: it wasn't a product, it didn't have overarching design goals or any sort of leadership. There was the core functionality, and a bunch of functionality that pretty neatly layered on top of that without making the whole thing too intimidating. I'm pretty happy with how it went.

Over the last two or so years, though, geojson.io hasn't changed much, and the web has. Which means that, in a few ways, it's just straight-up broken: GitHub integration is broken, it was never updated to accommodate for the deprecation of anonymous gists, and things like geolocation is also broken. It's also still on HTTP, which means that Chrome will soon mark it as 'insecure'. All bad news.

I don't want GeoJSON editing to be hard again, and I don't see great alternatives to geojson.io sprouting up. I also fear what it'd become if the intent wasn't simple, if there was some second or third-level goal of generating money or converting users. I also don't think it has much potential as a product - quick GeoJSON editing isn't something that people are itching to pay for.

So, I'm building it again, as a side project, just like the last time. I'll merge your PRs, and treat the occasional kudos as all the payment I need.

This kind of project is all about self-management: you have to say no to yourself, you have to be aware of fun things you could implement but would corrupt the focus. That said, I am aiming for a few minor things on top of the original mission:

  • It'll be much easier to run your own installation, and much easier to develop. This means complete stack installs - the frontend and authentication layer. Possibly something like one-command to deploy on Zeit or Heroku, too.
  • It'll be much easier to contribute. geojson.io was built before React had such huge mindshare, so it was created with the same d3 architecture I used for the iD project. This app will be much more vanilla, and much easier to contribute to in little chunks.
  • Security and privacy will be elevated, significantly. All HTTPS, no tracking, better control over GitHub permissions. User safety will be prioritized over any sort of analytics or development ease.

WHAT'S NEXT

  • Development at this stage is fast and chaotic - if you want to join in, I'll be working in the master branch of geojson.net.
  • If you are interested in using the thing, cool your jets, it'll be a little while.
  • The issue tracker's open, if you have thoughts or suggestions.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment