Hacker School, Wednesday, July 30th, 2014

Goals:

  • Explore the behaviour of inotify.
  • Work on the Protagonist rewrite.

Actual:

  • I explored inotify in a few different ways. I was so surprised about the behaviour of inotify, and the way it interacts with my desired use case, that I have decided to make a short presentation and post about it, so I won't elaborate here.

  • I made some good progress through implementing the new Protagonist design.

    However, the design is really complex, and I can't fail to notice that I rarely want to work on it, and when I did yesterday, it took a lot of effort to stay motivated. I had to keep the new design document open at all times to remember the desired invariants; it was hard to keep it all in my mental model at once. That seemed like a bad sign.

    Maybe this is normal for a project of sufficient complexity, but I started thinking about why the complexity had taken such a hit. It has arisen out of the attempt to accommodate immutable files.

    Remembering this, I was inspired to watch a talk that Zach Allaun had earlier recommended, when I talked about addresses and contents: Rich Hickey's The Value of Values.

These three things together — the behaviour of inotify under linking and copy-on-write-like applications, the complexity of Protagonist when trying to accommodate immutable files, and Rich Hickey's talk — compelled me to decide to re-re-think the Protagonist design in its approach to mutability.

Comments

Comments powered by Disqus
Share