Friday Blast #28
Horia Coman
Posted on January 20, 2019
A focus on probabilistic data structures with some Bitcoin thrown in this week.
Probabilistic data structures for web analytics and data mining (2012) - an article I’ve wanted to read for a long time. It covers Bloom Filteres, HyperLogLog counters, Count-Min Sketches and all that nice stuff and it also has a bunch of examples and real-world numbers for applications. There’s a bunch of formulas, but not proofs or intuitions behind why they happen. That’s a tall taks in general, since there’s a lot of non-intuitive material to cover.
Sketching data structures (2010) - just a small introduction to Bloom filters and count min sketches, with some analogies.
Probabilistic data structure showdown: cuckoo filters vs Bloom filters (2016) - a comparison of the two and an introduction to Cuckoo filters. The Bloom filters used are counting ones, which allow for deletion. It’s a bigger data structure than a regular one, but it makes for a fair comparison to Cuckoo filters.
On-call at any size (2017) - presents on-call practices for companies of 10, 11-100, 101-1000, 1001-10000 and 10000+ engineers. It’s really interesting how on-call and the nature of keeping systems up changes from small to huge. On one end of the spectrum there’s a lot of manual shepharding, while at the other there’s basically a statistical understanding of what the system should be doing. And failure, loads and loads of failure.
Updating navifation for StackOverflow, Enterprise, and Stack Exchange Sites (2018) - an overview of recent UI changes to SO and the process behind them. Some details about the upcoming channels feature.
How the Bitcoin protocol actually works (2013) - Even though I knew about Bitcoin since forever I didn’t take the time to properly understand it. Which is a shame, because even cursory glances suggested that it’s seriously cool tech. Apart from all the hype and mania surrounding it. The author is Michael Nielsen, who’s something of a polymath, operating in quantum computing, deep learning, networks theory and apparently distributed systems. The article doesn’t disappoint - it’s a very lucid presentation of what one might want out of a digital currency and how to go about doing that - ending up with Bitcoin at the end.
Posted on January 20, 2019
Join Our Newsletter. No Spam, Only the good stuff.
Sign up to receive the latest update from our blog.