Here are some blog posts.

key: ■ = theory, ■ = technique, ■ = musings, ■ = journal

• 30 Mar 2016 ■ How bad is the Windows command line really?
• 20 Aug 2015 ■ Porting Mesos to Windows
• 24 Mar 2015 ■ FoundationDB proves your primary datastore is the worst place in your stack to bet on new tech
• 09 Feb 2015 ■ 84% of a single-threaded 1KB write in Redis is spent in the kernel
• 26 Jan 2015 ■ Half of IBM's profits come from mainframes
• 09 Jan 2015 MSFT open sources production serialization system written partially in Haskell
• 01 Aug 2014 ■ The time I found a bug in the .NET framework and fixed it by hand-altering the DLL.
• 09 Jul 2014 ■ Beginner's guide to OCaml beginner's guides.
• 07 Jul 2014 ■ Should I Machine Learn?
• 24 Jun 2014 ■ Recovering deleted files using only grep
• 14 Jun 2014 ■ I needed to use a regex on an airplane without Internet. So I reverse engineered the API.
• 12 Apr 2014 ■ How does Akamai's "secure heap" patch to OpenSSL work?
• 25 Feb 2014 ■ Day 266: The unfamiliar world of OS X syscalls
• 23 Feb 2014 ■ Day 265: Semaphores and mutexes. Mutices. Muticii.
• 22 Feb 2014 ■ Typo in Apple's SSL implementation causes uniform failure to validate key exchanges
• 31 Jan 2014 ■ Swisher's AOL.com, 15 years later
• 04 Jan 2014 ■ What "viable search engine competition" really looks like.
• 02 Jan 2014 ■ Ticket to ride: reflecting on the value of my CS degree
• 01 Jan 2014 ■ Day 212: More about counting things with fancy math.
• 28 Dec 2013 ■ Day 208: telnet, turning you into a human server
• 27 Dec 2013 ■ Day 207: How to count things.
• 26 Dec 2013 ■ Day 206: What is an IP address, anyway?
• 16 Dec 2013 ■ General lessons from scaling large systems at Microsoft
• 06 Oct 2013 ■ Twitter's IPO in simple terms
• 19 Aug 2013 ■ Adding Haskell's `where` keyword to OCaml
• 09 Aug 2013 ■ Turning the Apple //e into a lisp machine, part 1
• 26 Jul 2013 ■ ■ The obvious Python parser
• 08 Mar 2013 ■ ■ ■ What exactly is the difference between machine learning and statistics?
• 04 Mar 2013 ■ Motivating the Bayesian prior with de Finetti's theorem
• 23 Feb 2013 ■ Why psuedorandomness is important in frequency moment estimation