The calendar, published before the writing is.
Nothing here has been published yet. What follows is what we intend to write and the argument each piece has to make — titles and outlines, no drafts, no bylines and no dates we would have to invent. A plan you can hold us to beats an archive you cannot check.
Everything else that is queued
Filter by tag or search the titles. Selecting one opens its outline — that is all there is, and the page says so rather than linking you to a draft that does not exist.
Eight planned pieces is not a large calendar. Try a broader tag, or clear the filters.
Three rules for anything published here
Show the alternative we rejected
A design post without the discarded options is an advertisement. Every decision piece names what was tried first and why it failed, which is why several of the outlines above end on the thing the design does not solve.
Postmortems are blameless and specific
When there is one to write: named commitments we missed, real timestamps, and the diff that fixed it. If the interesting failure is organisational rather than technical, we say that instead of hiding behind a race condition. There are none yet, and we are not writing fiction to fill the section.
No launch posts
Product changes go in the changelog and the docs. This is for reasoning that outlives the release it came from — if a post would be worthless in a year, it does not get written.
One email when the first one lands
New posts only, and there are none yet — so this list is silent until there are. Incident notices live on the status page and have their own list, so subscribing here will never get you a product announcement.
Disagree in public.
Every outline above started as an argument inside the team, and the useful ones get settled by someone outside it. The fastest way to change our mind — or the shape of a piece before it is written — is a specific counter-example.