0x0000bench
Build a target set.
Choose the formats and languages you want to test. The bench gives you the list of pinned targets, a script that runs them, a manifest for your artifact, and a table for your paper. Every selection gets an ID derived from its targets, so anyone can rebuild the same set.
A plain shell version of parsarium run-selection, for when you want to edit the steps. It runs the targets one after another and writes every receipt under one directory.
Run it directly with the parsarium command, or keep it with your experiment artifact: it pins the target IDs, releases, controls, feedback kinds, and the Parsarium commit.
A starting point for the evaluation section of a paper or report. Fill in the setup lines with your own budget and hardware.