Facebook Details Comprehensive List of Projects Opened to the Community

by Ostatic Staff - Jul. 21, 2016

Facebook, like Google, has shown itself to be a strong contributor to the open source community. In recent months, we've covered numerous open tools that were hardened at Facebook and released to the community. These include Haxl, a library that eases access to remote data, and a tool for optimizing the performance of open source distributed SQL query engine Presto, as noted in this blog post. Facebook has also open sourced its machine learning system designed for artificial intelligence (AI) computing at a large scale. The platform is based on Nvidia hardware.

Now, in a comprehensive post, the company has rounded up 54 of the open source tools that it has open sourced in the past six months, some of which could give birth to businesses if steered forward by the right entrepreneurs.

"With more than 1.65 billion people on Facebook and more coming online every day, our engineers are hard at work making sure that our services work smoothly for everyone," write the company's enginners. "Part of this work involves streamlining our processes so we can keep moving fast as we continue to scale: We build tools that enable engineers to work more easily across platforms, automate testing to catch problems sooner, and help improve the overall performance of our products. Many of these tools are built with the intention to open source them."

In the first six months of 2016 Facebook reports that it added 54 new projects to its open source portfolio. Many of the launches gained immediate traction within the community, according to the company, including more than a dozen new projects that have garnered more than 500 followers each.

Some notable examples include:

Draft.js: Facebook says "we open-sourced this React-based rich text editor framework at React.js Conf earlier this year." With more than 6.7K stars — 1,000 of which it accumulated within the first few hours of release — this project has introduced an easier way to customize rich text that easily integrates into React applications for the web.

ReDex: ReDex is a bytecode optimizer that makes Android apps smaller and faster. Facebook open sourced ReDex at F8 this year, and showed that it reduced the size of the bytecode in its own app by 20 percent. ReDex also arranges the bytecode more efficiently on disk, which helps engineers avoid user-visible slowness, especially on memory-constrained devices. To date, ReDex has received 2.8K stars and 240 forks.

Memory Bundle: Also released at F8, this suite of tools that includes FBRetainCycleDetector, FBAllocationTracker, and FBMemoryProfiler lets iOS developers profile an app's memory usage at runtime. It automatically finds and fixes instances where memory allocations may cause a crash, which helps improve overall app performance. Together, the memory bundle tools have more than 3.8K stars and 280 forks.

On the  artificial intelligence front, Facebook open sourced Torch implementations and training scripts of deep neural nets for image recognition, and Torchnet, a modular collection of boilerplate code, key abstractions, and reference implementations that builds on top of the Torch framework. You can also read more about its Big Sur machine learning project in our previous post.