> This needs a good deal more discussion before sweeping through the code and change a long standing Python idiom that has stood the test of time.

I agree this should be discussed, which is why I created this issue about it, as place for that discussion to take place. This is also being discussed on python-dev.

The examples I've mentioned are both rather new, and they have implemented sentinel values using three different methods, with those in dataclasses even being inconsistent among themselves. I fail to see how this can be considered "chang[ing] a long standing Python idiom that has stood the test of time".

> You "recently noticing this" doesn't mean that it is wrong.

I find this wording makes things unnecessarily personal and negative. Let's discuss things based on technical merit.