Every stochastic process has an initial state. This is mine.
I’ve been writing code for a while now, and contributing fixes to open source projects for almost as long. The pattern is usually the same: I find a library, I profile it, I notice something suboptimal, I open a PR, and then one of two things happens.
Either it gets merged — quietly, without fanfare, as most good patches should — or it gets rejected, usually with a polite explanation that the current approach was chosen deliberately and my “optimization” would break seventeen edge cases I hadn’t considered.
Both outcomes are instructive.
I’m starting this blog to document the process. Not because I think my contributions are particularly noteworthy, but because the methodology is worth recording. Every optimization attempt is, in essence, a hypothesis. Some are confirmed. Most are refined. A few are outright wrong.
If you’re here because you saw one of my PRs and wanted to understand my reasoning, welcome. If you’re here because you’re interested in probability theory applied to software engineering, also welcome. If you’re here by accident, the probability of that event was nonzero, so I can’t say I’m surprised.
What to expect
- Contribution logs: what I found, what I changed, whether it worked
- Post-mortems on rejections: the most interesting data points come from failure
- Occasional mathematical tangents: I apologize in advance for nothing
Let’s see where this random walk takes us.
P(this blog is useful) > 0. That’s sufficient.