henderson reed hummel





Tools I like to use (and thoughts thereon)

I live on the command line and in the browser:


I’m a huge fan of the Unix as IDE perspective. If you’re unfamiliar, I really recommend you read the series that started it all. For those of you without the time, I’ll summarize it here: essentially, it’s a workflow that emphasizes the use of small, modular, configurable tools (like those available on most *nixes) over the use of one monlithic tool (like your average IDE.) So, when I’m coding, I use many different tools, each of which are tuned and configured for only a few tasks. The advantage here is that I can more easily understand and extend each one.

A prime example of my attitudes in action is psg, a static site generator I built. It’s 77 lines of vanilla Python, with no dependencies aside from Pandoc, which is available as a package in most Linux distribution’s default repositories. It’s straightforward to extend, it has very few side-effects or expectations (meaning that it integrates with other tools well), and it’s easy to install (you can just curl it from GitHub.)

On building tools

I am a huge proponent of building your own small tools. I find that it keeps your mind sharp for programming, and you can often come up with a tool that satisfies your needs more than available solutions. Plus, you get to program and learn on the way!

I find that Bash is especially powerful for creating these simple tools. Other languages provide all kinds of benefits (types, safety, debugging tools, etc) but so far I haven’t come across anything that is as quick and effective for quickly scripting behavior.

For example, see the notetaking system I use:

mdless() {
      pandoc -s -f markdown -t man $1.md | groff -T utf8 -man | less
umedit() {
    vim "$UMDIR"/$1.md;
um() { mdless "$UMDIR"/"$1"; }
umls() { ls "$UMDIR" | \
    grep ".*\.md" | \
    grep -v "[0-9]*-[0-9]*-0-9]*" | \
    cut -d "." -f 1 }
# completion functions for um
_um() { compadd $(umls) }
compdef _um um
compdef _um umedit