Almost 5 years after the kickstarter made Blockbuilder real, it’s time to make a hard choice. By the end of the year Blockbuilder will become a searchable archive of blocks.
2020 is a lot different from 2015… to say the least. In-browser IDEs and code sharing tools are way better and Observable is taking its rightful place as the standard tool for the #d3js community.
Back in 2015 there was no easy way to quickly share or tweak an existing d3 example, so almost 200 people backed the kickstarter, and in the 5 years since more than 5,000 people have used blockbuilder to create, fork and share d3 examples. In that same time period 400,000 people have visited Blockbuilder over 2 million times. The vast majority of those visits are to those very examples created by the community.
We’ve indexed more than 40,000 d3 examples from the community on blockbuilder.org/search and moving forward, this will be what the project focuses on. The search page will remain available at blockbuilder.org/search, but it will not be updated as often. Think of it more as a searchable archive of d3 blocks, which are still incredibly valuable for many people (Blockbuilder still has more than 10,000 unique visitors every month). We acknowledge that many fewer new blocks are being created, and will still try to include them by updating the index from time to time.
The editor and the ability to create, edit and fork blocks will go away, redirecting folks to the search page most likely. When someone visits a user page (like blockbuilder.org/enjalot) or an individual block (like blockbuilder.org/enjalot/d5ee7ee1bb8dcbbdbfb0ecd1480aa640) they will be redirected to the corresponding bl.ocks.org page.
Keeping the IDE code, server backend and the continuous search infrastructure maintained has become too expensive, especially in regards to time: GitHub is deprecating one of the ways we authenticate API calls, the elasticsearch index is brittle and the UI codebase was established as the very first project I learned React with. shout out to @micahstubbs who has been keeping the lights on and shepherding some nice improvements from @hydrosquall. Dedicating time and attention to refactor and re-architect fundamental aspects of the project is something I can’t bring myself to do given the amazing alternatives available today.
Observable is what I’ve been using for all my open source examples & d3 prototyping over the last year. It’s what I wish I could have built and now it’s here.
There is also @currankelleher’s VizHub which feels like an improved (and better maintained) implementation of blockbuilder’s core concept.
On top of that, all the blocks anybody ever made in blockbuilder are still available on bl.ocks.org (the original goal was to help people make more blocks!)
All that said, Blockbuilder has always been open source and a sufficiently motivated person can certainly host their own copy of the IDE.
https://github.com/enjalot/blockbuilder
https://github.com/enjalot/blockbuilder-search
https://github.com/enjalot/blockbuilder-search-index
If you are interested in being involved with any aspect of the plan, or interested in hosting your own copy of the app, please reach out to me, @micahstubbs or join the #blockbuilder channel on the d3js slack
I can’t express the depths of my gratitude to the #d3js community for the opportunity to participate in so many people’s creative processes. I was deeply moved by the trust granted to me by the kickstarter backers and the support I got from friends and internet acquaintences alike. I’d like to give a personal shout out to a few folks who have strongly supported this project through the years: Erik Hazzard, Micah Stubbs, Shirley Wu and Victor Powell. Of course none of this would be possible without Mike Bostock’s tireless commitment to data visualization tools.