On 5/12/26 08:52, Tim Daneliuk via Python-list wrote:
I stipulate that there is some utility to a sentinel feature. But it
does so at a
cost. That cost is bloating the language definition with yet another
feature that
has fairly limited application beyond what is already possible.
I've seen this happen over and over again in a variety of languages.
"Elegant"
solutions are proposed to legitimate problems but these solutions make the
language bigger, harder to master, and harder to maintain.
I think it depends on how you squint through the lens. To me, it makes a
lot of sense for the language to adapt to help solve problems that have
proven to be tricky to handle gracefully *without* help from the
language, and I don't see that as bloat, just as normal adaptation as it
matures and is put to new uses. It's the bigger things that usually
give me headaches (looking at you, async, cough cough). A single
sentinel() call doesn't really make the language harder to learn, does
it? You code away. If at some point you recognize you've run into a
scenario where you need a sentinel value and nothing is quite right, you
go searching for help - and oh, there is a feature to help with that.
I know the decision makers consider these factors very carefully - PEPs
often go through a fair bit of requested change with some rejected or
deferred indefinitely until proof of need (vs. cost) can actually be
demonstrated, and how to convey a language change to the user community
is a required component.
--
https://mail.python.org/mailman3//lists/python-list.python.org
