Peirce Quincunxial notebook broken?

Actually, it looks like you’re right! Using the API described here, compare the entries in nodes with id=3 of version 357 (the current version of the notebook) and version 356. Earlier versions seem not to be public.

Interestingly, for both versions, the field publish_time returned by the API gives the same time on Dec 17, 2017 so it can’t tell us when the update was made. However, it’s now reasonable to reinterpret the question mark in the fork I linked in my previous post. It’s not that you were accidentally looking at the wrong notebook, but rather that that notebook contains the question mark since it was a fork of version 356! That fork was created Mar 7, 2018, so this suggests that the change to the original notebook occurred after then.

It’s possible that the require statement used to secretly allow a question mark and that this was later removed. If you dig through the commits to notebook-stdlib and d3-require you might find the answer…