Proposal: Introduce eval steps in edit mode about primer

[dhess]

There are some good ideas here, so long as we're very careful that students don't conflate actions they perform on their programs with evaluation steps that the evaluator performs (even under the student's direction) on evaluated expressions.

When we have some time to return to frontend work, let's have a dev+design meeting about this.

[georgefst]

so long as we're very careful that students don't conflate actions they perform on their programs with evaluation steps that the evaluator performs

As we've discussed, there's some tension here since we also want them to recognise that all evaluation states are valid programs which they could construct in edit mode.

[dhess]

Should this PR be retitled something like, "Add eval-like refactoring actions"? It's not clear from the title whether you're proposing that students should be able to do evaluation in/on expressions directly in edit mode, rather than needing to go into a special eval mode; or if you're proposing that we add actions like, "Function application -> let binding."

I think those are two very different conversations, so let's make sure we're all on the same page.

[georgefst]

Should this PR be retitled something like, "Add eval-like refactoring actions"? It's not clear from the title whether you're proposing that students should be able to do evaluation in/on expressions directly in edit mode, rather than needing to go into a special eval mode; or if you're proposing that we add actions like, "Function application -> let binding."

I want a way for students to get used to (at least some of) these actions before being launched in to a separate mode where they have to get used to them all at once. It can be a first step towards understanding the actual semantics of the program. That's the main thing I'm getting at here.

On the other hand, I have been thinking about whether Eval can be more of an overlay than a separate mode (a kind of "undo/redo on steroids"), and that has fed in to this a little.

I think those are two very different conversations, so let's make sure we're all on the same page.

So yes, maybe we should focus on the former here, and save the latter for a separate proposal (once I've given it more thought), but they are closely linked.