xact VS tinyjam

https://github.com/wtpayne/xact - Model Based Systems/Software Engineering tool with support for machine learning and synthetic data.

MBSE is a vastly underappreciated technique, and one that deserves much more exposure in the mainstream tech community.

Sadly, I have found myself exceedingly disappointed at the tooling available to support MBSE, having mainly used Simulink, and also experimented with a few others like Rhapsody, Capella and EA.

Constant mouse usage with Simulink gave me really bad RSI, and merging models was a pain, due to the way that layout and structural information were mixed together in the xml-based .mdl file format.

So (naturally) I made a text-based alternative. It's a bit like TensorFlow in that you create a compute graph (computational model), and then run it.

The model itself can be generated dynamically in a script, using JQuery-like syntax to add or change nodes, or alternatively it can be stored and version controlled in one or more text files, using YAML, XML, JSON or TOML (or some mixture of those) to serialise the structure in an easy-to-diff-and-merge form.

https://github.com/wtpayne/xact

I have a ton of unfinished open source projects, but here's a recent small one I believe can take off:

https://github.com/mourner/tinyjam — a bare-bones, zero-configuration static site generator that deliberately has no features, an experiment in radical simplicity.

Essentially a tiny, elegant glue between EJS templates and Markdown with freeform structure (enabling incremental adoption) and convenient defaults, written in under 120 lines of JavaScript.

I also made a modern EJS implementation specifically for this: https://github.com/mourner/yeahjs