Olivier Leroydefuneste@fosstodon.org
Jun 4, 2026, 1:54 PM@wronglang @adam_wysokinski @nxskok
here isn't it because R is full on "copy on write" ?
Replies
Krzysztof Sakrejda 🏳️⚧️ 🏴☠wronglang@bayes.clubJun 4, 2026, 1:56 PM@defuneste @adam_wysokinski @nxskok all I'm saying is they could've saved people a lot of grief by not making brackets a syntactically correct constructor with weirdly specified arguments.
- Olivier Leroydefuneste@fosstodon.orgJun 4, 2026, 2:16 PM
@wronglang @adam_wysokinski @nxskok
Fair point, I agree that the [ ] def make it looks like some kind of setting (and like mentioned in the thread will throw you out of bonds error with [[ ]])
Jun 4, 2026, 2:30 PM@wronglang @defuneste @adam_wysokinski @nxskok
R's "YOLO, let's do it!" attitude (not to mention what passes for a type system) drives me nuts, but I understand _this_ behavior at least.
>it was designed by statisticians who happened to do programming
I understand making this comment in a hand-wavy fashion, but I think this underpins the valid technical reason here: In general, data you get from the real world (especially the kind of data that people who use R a lot get) SUCKS. It's full of holes and gaps and spurious and invalid values. If R choked by default on every one of these, no one would ever get anything done.