Nice generator! Parsons Puzzles are a well researched way of scaffolded learning in the CS community.
Last year when I was teaching UC Berkeley CS61A, we tried out a variant called Faded Parsons Puzzles, where the draggable code lines also have blanks and get verified via Python doctests. I made a browser-based runner for that style, for anyone who wants to easily create new Faded Parsons Puzzles.
Woah — I spent a couple of weeks last year doing a quantitative comparison of the outcomes of students using FPPs vs ones using traditional PPs. And I’m pretty sure I emailed someone at Berkeley (last name Weinman, I think?) asking some questions about their paper on FPPs. The Internet really is a small place!
I think you should be able to use OP's site or the js-parsons library to make JS puzzles, if you're okay with "line-based feedback". To use Faded Parsons with JavaScript and test-based feedback, you'd need to decide the format of the tests, since JS doesn't have a built-in doctest module like Python. Once you did that, it should be straightforward enough to use it for JavaScript. Happy to help if you end up branching off my repo.
Oh nice, I didn't realize you had that capability. It looks like you're using js-parsons, which is also what we branched off. I ended up deleting the line-based grader logic in our js-parsons fork, since we only need doctest-based grading and I wanted to clean up the code, but in theory, both kinds of graders can co-exist.
Last year when I was teaching UC Berkeley CS61A, we tried out a variant called Faded Parsons Puzzles, where the draggable code lines also have blanks and get verified via Python doctests. I made a browser-based runner for that style, for anyone who wants to easily create new Faded Parsons Puzzles.
See demo here: https://pamelafox.github.io/faded-parsons-static/index.html More background here: http://blog.pamelafox.org/2022/08/a-browser-based-runner-for...
We'll be presenting a poster about our experiment with FPPs at this year's SIGCSE.