Nowhere have I suggested that the contributor should not make a reasonable attempt at producing a bug free patch. And Jolt provides plenty of design docs actually: https://jolt-lang.net/docs/rfc/README.html
I also strive to have a clean and modular codebase, including building tools specifically aimed at helping ensure code quality in addition to manual reviews https://github.com/jlt-commons/jolt-metrics
So, if you'd like to point out specific gaps in documentation or problems in the code architecture, then feel free to share them and we can discuss that.
> What I'm against is making the project difficult for human understanding because human understanding is limited or whatever.
I'm against that as well, so not sure what your point of contention is really. The README simply points out that there are limits to what we can reasonably understand, and that the mitigation for that lies in the process such as having good specifications, benchmarks, regression tests, and so on.
Again, if you can point to specific aspects of the project that you claim make it difficult for humans to understand, then please be specific regarding what they are instead of making vague accusations. Also, it's quite interesting that you chose to take a few bits out of context while ignoring the following which is directly relevant to the accusations you are levelling here:
> The core skill in using these tools is the same as the skill needed when writing code manually. One has to understand the given requirement, the shape of the solution, and how to verify that the code produced actually matches the intent.
> What these tools do is automate the mechanical aspect of producing the code. Once it is written, you can read it, understand it, and change it as you would with any other code. The intention and the ideas come from the human using the tool, and new ways to automate the work of typing out code do not change the role or the responsibility of the developer.
> Focusing on how the code was produced rather than on the code itself dilutes the responsibility since blame should never be deflected to LLMs. The ownership of the code rests with the person submitting the code, and they must take responsibility for its quality and correctness. Either the code is clean and lives up to the standards of the project, or it will be rejected and the author will be asked to make changes to address the problems.
Can you point to where in your mind I'm suggesting that developers shouldn't try to ensure code quality or understanding?
I'm using jolt in a new project and yesterday submitted my first PR to jolt, to improve some testing functionality. I'm pretty confident in it's quality and I'm super excited about jolt. It has opened very interesting frontiers for me as a clojure developer.