I've had plenty of open book exams where the prof knew you would fail if you grabbed the book for more than a second. It's pretty much the same as the exams where you get to write your own cheatsheet: if you need it too much you are screwed.
I had used Illustrator to lay it all out. Lots of well type set diagrams and graphs alongside the equations.
Of course after spending the better part of 2 days making it I barely had to refer to it during the test!
The sheets were horrifically over made. They also benefited from my being fairly near sighted. You can fit a lot more in when using a 4 point font!
Did that for a few I did not care about. Passed. Ofc I’m not as much of an expert in those compared to the ones I “broke my teeth” on.
Read the course syllabus, now divide it into three lists:
- What you know you know
- What you know you don't know
- What's left is what you don't know you don't know
Maybe a simplified example might be a question that forces you to consider different data structures and choose the right one? A student may not have the experience to know off the top of their head but they have a reference they can skim to check. The trick would be setting it out such that a student that didn’t know the principles would completely miss this and not know what to look for. Like they would do nested loops instead of populating a hashmap, perhaps.
80% of my mates didn’t solve it. It was right there in any graph algo book.
It does not help you if you don’t understand the material. If the exam is done right at least.
So in my book (pun intended :P), allowing and actually encouraging a "cheat sheet" is a good thing. Open book is worse, as it's usually way too large and badly indexed. And who's gonna use an actual book in their actual job anyway?