This feature is very useful, and so far it hasn't stopped me from doing anything interesting. ![]()
To be honest, I expected more friction!
I also thought it would reveal more bugs in existing code. It's caught only a few. It does seem that by and large, if you get the arity wrong you tend to break things badly enough to notice something is broken... it can just be hard to debug by eyeballing it.
Today there was an example in @gchiu's PatientDB I found, where I hadn't updated it for UNTIL being made arity-2
until [insert years n, n: me - 1, n = 2018] ; oops, should be INSIST
months: ["December" "November" "October"...]
So what was happening here was that the UNTIL was using the first block as its condition, and the months: [...] block as its body.
Humorously enough, this still worked. The data block of strings being executed was a no-op, so it just kept running the condition.
But with the protection, it errors that "December" is being discarded when the months block is evaluated.
Certainly it's a good sanity check, and better to know what your code is doing than have it doing no-op body evaluations of data!