The React license for founders and CTOs – James Ide – Medium
Decent explanation of _why_ Facebook came up with the BSD+Patents license: “Facebook’s patent grant is about sharing its code while preserving its ability to defend itself against patent lawsuits.”
The difficulty of open sourcing code at Facebook, including React in 2013, was one of the reasons the company’s open-source contributions used to be a fraction of what they are today. It didn’t use to have a strong reputation as an open-source contributor to front-end technologies. Facebook wanted to open source code, though; when it grew communities for projects like React, core contributors emerged to help out and interview candidates often cited React and other Facebook open source as one of the reasons they were interested in applying. People at Facebook wanted to make it easier to open source code and not worry as much about patents. Facebook’s solution was the Facebook BSD+Patents license.
(tags: facebook bsd licenses licensing asf patents swpats react license software-patents open-source rocksdb)
HN thread on the new Network Load Balancer AWS product
looks like @colmmacc works on it. Lots and lots of good details here
(tags: nlb aws load-balancing ops architecture lbs tcp ip)
Java Flame Graphs Introduction: Fire For Everyone!
lots of good detail on flame graph usage in Java, and the Honest Profiler (honest because it’s safepoint-free)
(tags: profiling java safepoints jvm flame-graphs perf measurement benchmarking testing)
Teaching Students to Code – What Works
Lynn Langit describing her work as part of Microsoft Digigirlz and TKP to teach thousands of kids worldwide to code. Describes a curriculum from “K” (4-6-year olds) learning computational thinking with a block-based programming environment like Scratch, up to University level, solving problems with public clouds like AWS’ free tier.
(tags: education learning coding teaching tkp lynn-langit scratch kids)
So much for that Voynich manuscript “solution”
boo.
The idea that the book is a medical treatise on women’s health, however, might turn out to be correct. But that wasn’t Gibbs’ discovery. Many scholars and amateur sleuths had already reached that conclusion, using the same evidence that Gibbs did. Essentially, Gibbs rolled together a bunch of already-existing scholarship and did a highly speculative translation, without even consulting the librarians at the institute where the book resides. Gibbs said in the TLS article that he did his research for an unnamed “television network.” Given that Gibbs’ main claim to fame before this article was a series of books about how to write and sell television screenplays, it seems that his goal in this research was probably to sell a television screenplay of his own. In 2015, Gibbs did an interview where he said that in five years, “I would like to think I could have a returnable series up and running.” Considering the dubious accuracy of many History Channel “documentaries,” he might just get his wish.
(tags: crypto history voynich-manuscript historians tls)
How to Optimize Garbage Collection in Go
In this post, we’ll share a few powerful optimizations that mitigate many of the performance problems common to Go’s garbage collection (we will cover “fun with deadlocks” in a follow-up). In particular, we’ll share how embedding structs, using sync.Pool, and reusing backing arrays can minimize memory allocations and reduce garbage collection overhead.