Dr. Brian Robert Callahan
academic, developer, with an eye towards a brighter techno-social life
My previous post generated a bit of confusion: some comments on Hacker News, Lobste.rs, and elsewhere seem to think my semi-retirement says something negative about OpenBSD or that I had a falling out with Theo or something. The text does not invite such a reading. And I haven't spoken to Theo in years; not for any personal reason just that no real opportunity or need came up.
I don't really want to clarify myself, but I asked Claude about it, and it understood me exactly, so I am just going to copy and paste the prompt:
Explain what the section titles are doing in this blog post: https://briancallahan.net/blog/20260322.html
And the response:
The section titles are doing two jobs at once, and the second one is easy to miss on a first read.
## The surface job
Each of the seven `####` headings honestly labels the section beneath it. Walking down the post:
1. **"To make something the best"** sits over the opening reflection on wanting OpenBSD/the BSDs to be the best they can be — the pursuit that started technical and became personal.
2. **"You have to examine it"** introduces Todd Mortimer's 2019 AsiaBSDCon paper on anti-ROP compiler mitigations and the grad students' wish to test the techniques' efficacy. The section is the *setup* of an examination.
3. **"Even, or especially, when it falls short"** covers the testing method and the discovery that the results were "not so clean-cut" — modest at best, partly flawed at worst.
4. **"Facing reality"** is the data dump: the actual percentages, the paired t-test, the paradoxical combined-mitigation result.
5. **"In order to make it better"** is the post-mortem — possible explanations for the disparity, and the reviewers' point that unique-gadget count isn't a meaningful metric.
6. **"You might need to step back"** is where he confronts the conflict of critiquing OpenBSD from the inside and decides to semi-retire.
7. **"So that you can launch into the future"** is the resolution: the research lab, more BSD conferences, fostering the next generation of developers.
So far, unremarkable. Descriptive headings.
## The hidden job
Read the headings in order and they aren't seven independent labels — they're fragments of **one continuous sentence**. String six of them together and you get:
> "To make something the best, you have to examine it, even — or especially — when it falls short. In order to make it better, you might need to step back, so that you can launch into the future."
That sentence is a complete, grammatical, self-contained aphorism. And it isn't a random nice-sounding line: it's the thesis of the entire post, stated obliquely. The literal narrative — start loving a thing, test its claims, find the claims wanting, reckon with that, decide how to improve, retreat from a role, reposition for bigger impact — is a point-by-point *enactment* of that hidden sentence. The headings are simultaneously a table of contents and a compressed restatement of the argument.
Notice the grammatical mood, too: it's second person and gnomic — "you have to," "you might need to" — phrased as universal advice. The post itself is intensely personal (his commits, his students, his retirement), but the heading-sentence launders that private experience into a general lesson anyone could apply. That's a deliberate rhetorical move: "here is what happened to me" reframed as "here is what one does."
## The one heading that breaks the pattern — and why
You'll have noticed I quietly skipped **"Facing reality"** when assembling the sentence. It's the fourth of seven, sitting dead center, and it doesn't fit the flowing syntax: drop it in between "when it falls short" and "in order to make it better" and the sentence stumbles.
That, I'd argue, is the whole point — and it's the most interesting thing the titles are doing. Six headings glide along as a smooth, almost lyrical piece of wisdom. Right in the middle, the elegant rhetorical machinery cuts out and gives way to two blunt, plain words: *facing reality*. The form of the headings reproduces the content of the story. His expectations going in were smooth and confident — these mitigations are "entirely free," powerful, everyone should adopt them. Then the data interrupted that smooth story with something flat and undeniable. The headings stage that same interruption typographically: the aphorism is broken, mid-stride, by reality, exactly where the post's argument is broken mid-stride by reality. The blunt heading is the hinge the polished sentence bends around.
You can read it two ways, and both land in the same place. Either "Facing reality" is *excluded* from the sentence — the one moment he drops the artifice and just states the fact — or it's a stark standalone beat *planted inside* the sentence, a caesura: setup… **reality.** …resolution. Either way, the plainest heading marks the exact pivot of the essay, the collision between the romantic ("make something the best") and the empirical (the numbers).
## A couple of finer touches
The closing pair carries a deliberate paradox. **"You might need to step back / so that you can launch into the future"** says that *retreating is how you advance* — which is precisely how he frames the semi-retirement. He's at pains throughout not to call this quitting; the post's own title is "Semi-retirement, or, really, *changing my relationship* with the BSDs." The heading-sentence encodes that reframing: stepping back from the developer role is the move that lets him launch the research lab and reach more projects and people than he could from the inside.
And the bookends mirror each other. The first heading and the last describe the same commitment from opposite ends — wanting to *make something the best* at the start, *launching into the future* at the close — which matches his explicit final line that his responsibility to make the BSDs the best hasn't changed, only the way he goes about it. The headings open and close on the same value, with the examination, the disappointment, and the reckoning sandwiched between.
So, in sum: the titles are a covert single sentence that doubles as the post's thesis and as accurate section labels, written in universalizing second-person wisdom-speak, deliberately ruptured at its midpoint by the one literal heading — and that rupture is itself an argument, letting the *shape* of the headings act out the post's core experience of having a clean hypothesis broken by messy reality.
It is a message about what one does about commitment when circumstances change. You're free to disagree with it, but it is important to disagree with what I actually said (and believe).
More OpenBSD mitigation evaluations coming to an ACM/IEEE conference proceedings book near you...
