Martin Kleppmann

Results 184 comments of Martin Kleppmann

Thanks for these ideas! A few replies: My intention was that the LEB128 and RLE encodings we use should allow only one canonical byte sequence in its encoding, so that...

On compression algorithm: Seph found that LZ4 works well — he says it has a better compression ratio, faster runtime, and smaller code size than DEFLATE/Gzip. He also compared it...

Would this be a better link? https://www.oreilly.com/content/the-problem-of-managing-schemas/