If you have a decent IDE, it'll offer you the ability to swap between the "old" and newer way of doing things when you encounter code written in one of the styles.
I can't say I've had any issues getting code using the new syntax through code review though. C# 14 has been out long enough that the team is familiar with much of it, and the IDE is helpful at reminding you to consider adopting new syntax. That aside though, the collection expression syntax is pretty familiar for anyone who's ever written e.g. JavaScript.
We're at an interesting time for programming language development. Typically, newer features made the developer experience better. Either a faster way of doing something, it being safer, or easier to read/understand. But if GenAI models end up writing most of your code, there's only marginal gain ('easier to understand'). Yet understanding comes from using the features, so if GenAI were to write code using new features, you might not even get that benefit anymore as _you_ don't understand them well enough.
However, there's a larger problem lurking here. GenAI models code based on what's represented in their training data. New language features are (mostly) absent from that. And if more and more slop is pushed to Github, that slop isn't using the new features either. Which ultimately results in fewer new language features seeing widespread adoption.
Maybe I'm too pessimistic, but I think it's a somewhat dark time for PL development. And to quote the article, I can't prove it matters..
10 comments
[ 0.20 ms ] story [ 36.1 ms ] threadAnybody with enough experience know that.
I can't say I've had any issues getting code using the new syntax through code review though. C# 14 has been out long enough that the team is familiar with much of it, and the IDE is helpful at reminding you to consider adopting new syntax. That aside though, the collection expression syntax is pretty familiar for anyone who's ever written e.g. JavaScript.
However, there's a larger problem lurking here. GenAI models code based on what's represented in their training data. New language features are (mostly) absent from that. And if more and more slop is pushed to Github, that slop isn't using the new features either. Which ultimately results in fewer new language features seeing widespread adoption.
Maybe I'm too pessimistic, but I think it's a somewhat dark time for PL development. And to quote the article, I can't prove it matters..
As for the rest of the examples, I would rather stick with the old way. Sometimes new syntactic sugar techniques make code more confusing.
Why should I create a medium account to read a blog? Am I so out of touch?
Shouldn't the author want the article be available to the biggest audience?