This project's funding goal was not reached on May 27, 2012.
About this project
Have you ever tried to display a lot of markers on a map? Have you seen this kind of map (left side of the image above) with a lot of markers where you can't see the map anymore and where you can't click anywhere?
You're probably going to say, "What! I already have markerclusterer for my google map, why do I need something else?".
You need it for many reasons:
- MarkerClusterer is a bit outdated, hard to use and hard to customize.
- It only works on GoogleMap. But today, we have many other mapping systems like BingMap, or the more and more popular OpenStreetMap. Many (like FourSquare or Apple) are switching to OpenStreetMap. Those systems deserve a reliable and efficient clustering library.
- You'll be able to change mapping systems, without being stuck because of a third-party library.
- The new library will be easy to use and to extend and fully themable.
In the past, Sébastien has worked on different open-source projects, the first, Prototype Window Class (2006) has been popular for years.
He has the track record and skill to build an efficient client-side clustering library.
Christopher O'Connor is project manager at navia.com. Together with the rest of the crew, he has overseen the launch of several successful apps.
- Google Maps (V3)
- Bing Maps (V7)
- OpenStreetMap throughout leaflet.js. (supporting Leaflet will open this library to many mapping systems)
The build system will make distrib files for each mapping system.
Look and feel will be themable.
We'll write code in CoffeeScript but will also make a task to build JS files (full source, minified, gzipped) and will run this task before any commit to github. You won't have to run this locally, you'll be able to get distrib files from github.
We will make detailed documentation using docco and github pages.
Everything will be well tested using the mocha framework.
As soon as we begin to work on the project, we'll open a public github repository so everyone will be able to track progress and even be able to participate in early stages of development.
Everything will be managed using github (issues, pull requests...)
We believe and love open-source. we use many open-source projects every day like Ruby, Rails, jQuery, ... But we think that open-source doesn't necessarily mean to work for free. Most popular open-source projects today are funded.
This is why we're trying kickstarter.com today. we believe in the developer community to make it happen.
I immediately wonder: what will be the difference between your server side clustering engine and this library?
maptimize, our server-side clustering engine, is for site that have a LOT of data. More than 10 000. If you have hundreds of thousands points to display on a map, don't even think of using a client-side library. Because,
- you'll have to send all points to the browser. You could send megabytes of data.
- you'll depend on client browser/computer. Imagine that on an old computer with IE6 :)
But with up to few thousands points, client-side clustering works great. Many website don't have hundreds of thousands data to display on map.
- is easy to install
- doesn't require any code in your server application
- As I said before, works great with small amount of point.
If we get more than expected or if you go faster than our estimate, the money will be used to add support for more mapping system. Of course with leaflet or Mapstraction you can use any mapping system, but I know people who prefer to use "native" API instead of that kind of wrapper. This is why we want to support both.
We also plan to make other representation like graphic or heatmap. For example, imagine a map of election results, we could display pie chart depending on zoom level instead of a simple cluster with a number of people.
- (30 days)