Three More Posts from My Publishing Plan

I ran the publishing plan again today, this time for three posts. The queue was empty, so the selection came from the weighted draw. The result took me from an API lesson to Italian crime fiction and then back to software teams. That is a pretty good sample of what happens when a network has sites with very different interests.

The first stop was API Course, with a practical introduction to curl. The lesson starts a small server on the reader's own computer, then uses curl to fetch a response, inspect headers and send JSON. The examples were run against a local server. I like having a course article that leaves the reader with something to try, even if the little server itself is just a teaching example.

Next came Yellow Fiction, for a look at how Mondadori's yellow-covered crime books gave Italian mysteries a name that later travelled into cinema. The piece follows that connection to Mario Bava and the visual language of giallo films. It is the sort of publishing history I enjoy: a design decision on a book cover becomes a word people use for an entire kind of story.

The third article went to Agile Software Development. It asks what a team should examine when a coding agent opens a pull request: the behavior promised by the change, the failure cases, the tests and the person who will own the result. One example involves retrying a payment after a lost response, where a neat patch can still create a serious problem if the operation is not safe to repeat. A green test suite is useful, of course. Someone still has to understand what was tested.

There was a mistake in the selection process. ChatGPT left WordPress sites out of the first draw, so API Course was picked from an incomplete pool. The calculation was corrected before the other two draws. All three sites had an overdue score of 1.86 under the plan's quick-date calculation, and the site plans and recent posts were checked before writing. The three articles were committed to their repositories and entered in the network log. Their live deployments and Hugo builds were not checked during that run.

The plan is supposed to help me keep the whole network moving, which makes the missed WordPress sites worth recording. The next draw needs the whole pool.