Without open source, Facebook would not be possible. It is built entirely on it. We use, contribute to, and publish a lot of open source software here as engineers, including components of our essential infrastructure like HipHop and Thrift.
However, in our C++ services code, one obvious barrier to publishing further work has been the requirement for any open-sourced project to remove dependencies on unpublished internal library code. Folly, a collection of reusable C++ library artefacts created and utilised by Facebook, was today open sourced as a first step in helping to address that issue. This declaration was made during our C++ conference held at Facebook in Menlo Park, California.
Folly is fast
Folly contains a number of tools that we frequently use in business; this is code that works every day on thousands of servers on behalf of 900 million people. Although there are some tenuous connections between these services, great performance at scale unifies them all. Some of them will have a pretty narrow concentration, such as lowering contention or cramming items into limited memory. Others have a wider range, like our string-formatting library or our in-memory JSON manipulation package. However, regardless of the situation, our goal was to create parts that were quicker and more effective than the ones we had previously.
Folly is easy to use
Although speed and ease of use are also very crucial, as we have several hundred engineers that work largely in C++, performance is of utmost importance. We heavily rely on current C++ as a result, which has allowed us to maintain a lot of programming comfort without using a lot of our hardware resources. In terms of programmer convenience, we think several of these packages even represent a step forward.
These are few of the main benefits FOLLY Library.