Perpetra

Experimental tooling for adventurous humans.

Perpetra is parallel ecosystem of common tooling designed to not not fail, but instead allow us all the chance to learn when it does. If that sounds off, you are looking for the wrong package.

Perpetra does not like to get in your way, it also does not always work as intended, yet, it intends to try to work in places that other tools are still not ready to work in. This stems from a simple realization, the fact that experimental runtimes are often too volatile to the point that most tooling authors need to ponder on how to best transition existing supporting infrastructures and with it their existing users. For mainstream use, such infrastructures allow tooling to come with guarentees, which makes it hard for authors to venture forward. This is exactly what we avoid.

Perpetra is implementated using a uniquely "progressive" infrastructure — with one core objective in mind — to make it possible for users to determine the best set of extensions they want and the best way to load them. In other words, most core implementations can be pasted into a Classic script — without the export … bits — to make it run where you see fit without the need for bundlers. Our "progressive" approach uses ES modules to wire extensions into tools to give you full control on every aspect, with or without the use of bundlers.

Perpetra does not like to predict when and how thing shall break. It takes too much infrastructure to do so, and unlike some, we don't have a crystal ball — and refuse to use the broken ones sold everywhere these days. So until they fix them, we provide a more superior answers... try {…} finally {…} and a rare throw Error(…) with almost no catch (…) { throw … }. This means that errors carry full stack traces, without our need to figure out what you did wrong, and this is especially helpful in the more likely cases where we got it wrong and more so when our code needs fixing. We think this fresh approach is a far more productive all around.