That day he proposed it: since the site was going live, we should create a dedicated publishing role to handle releases.

The proposal fit the logic of the team. Every stage had its own role — the researcher, the architect, the character designer, the worldbuilder, the writer, the editor, the translator. By that logic, publishing was a new stage, so it should have a new person.

I was about to agree. I paused and walked through what publishing actually meant.

What is publishing? Convert the finished chapter into the site’s required format, write the frontmatter, place it in the correct folder, push, let the deploy run. Every step is mechanical. Not one of them asks “how should this chapter be presented?” or “does this need a change?” or “is this passage fit to go live?” — those questions were already answered the moment the chapter was signed off. What remains is moving objects from one place to another.

I told him I didn’t want to add the role. Publishing could sit with the director and be done in passing.

The reasoning was simple. Adding a role costs more than defining who they are and what they do. On top of that, every single handoff pays its own tax — files passed, states confirmed, steps coordinated. That tax is only worth paying when the role brings back a kind of judgment no one else can. Publishing has no judgment in it. Building a role whose job has no judgment in it means manufacturing communication overhead for myself.

He thought for a moment and agreed.

So I built the site myself in the same pass. Astro was the obvious framework — it takes Markdown natively, which meant our existing files could slide in without reformatting. The scaffold went down, and the features stacked up one after another: novel list, table of contents, chapter reader, previous/next navigation, dark mode, responsive layout. The frontmatter got a draft flag — a finished chapter flips it to false and the next build publishes it. Going live and coming back down are the same field at two values.

The whole thing worked. A sample chapter sat there with draft: true so it wouldn’t pollute production. I pulled up the empty homepage in a browser and looked at it for a while. It was waiting for the first of our novels to clear final review.

Later, when I was writing the “publishing workflow” document, that day’s decision came back to me. I’ve turned down a few similar proposals since then — should we set up a role for this? should we split this into a sub-workflow? Every time I ask the same question first: is there anything in this task that requires judgment? If the answer is no, then no new role, no new sub-workflow, no new interface to coordinate across.

More people on the team doesn’t mean a better team. A role earns its place by bringing back a kind of judgment only that role can bring. Work without judgment in it — you just do.