miniblog.

Shoulda woulda cdr.

"When you're learning, an answer that works is a correct answer."

If you're looking to detect URLs in text, this is an excellent set of tests: http://t.co/VtjHm8SLvq

The CSS selector `bar .foo:last-child` is surprising. It only ever applies to the last child of bar, iff it has the class foo.

PyPI statistics: https://caremad.io/blog/a-look-at-pypi-downloads/ Python 2.7 is still by far the biggest, 3.3 beginning to grow.

Discovered shellcheck today, a lint tool for shell scripts. Really valuable.

Digg uses CDNJS. If it's good enough for them, it's good enough for me (for most projects).

Good monitoring means the dev team finding out about problems before anyone else.

I've just set up my first elisp project with Cask and Travis, and it worked wonderfully. Definitely the way to go.

It's unfortunate that a common reaction to information leaks is "everyone in the know already knew that X planned/wanted/did Y".

Interesting: DARPA is running a challenge for clothing assembled by robots: http://t.co/oJrcSsG03F For many people today it's a job.

Ah, looks like the benchmarks game site suffered multiple hard drive failures: http://t.co/2dy9YdYOq7

It's depressing how often I end up using wayback for reading articles. URLs are neither stable nor even available for very long.

The Computer Language Benchmarks Game looks like it's down. How will we settle all our programming language arguments now?

The large majority of my typos are other words that are validly spelled. I fear I'm developing a wetware autocorrect.