Our underworld web site has always carried two kinds of writing. Some of it is news — a release, a workshop, a paper accepted. The rest is the sort of thing a reader later wants to point at: how a solver behaves, what a benchmark measured and what it demonstrates, why a method works the way it does and where it stops working. The two have been mixed together, and there has been no way to cite the second kind.
So we now have technical notes that are written like a short paper or perhaps a preprint. Each one carries a version, an archival PDF built from the same source as the page, and a DOI that resolves to that PDF. Deposit is deliberate rather than automatic: a note is published here first, and only gets an identifier when we decide it is worth one.
The point is for us to write down what we have created with enough content to be useful and not a lot more, and then refer to it — from a paper, from an issue, from the next note; instead of describing it again. Not quite documentation and not yet a research publication.
The first two are up: Moving the Mesh Without Remaking It, on redistributing mesh nodes to follow the physics without remeshing, and Running Underworld in a Browser, on launching any repository at any version without installing anything.
Everything written before is still here, at the same addresses. The fifty posts that already carry DOIs keep them, and those DOIs still resolve.
If you work with Underworld and have something worth writing down, the notes are open to you — how to submit one. You can read all of them, or browse by topic.