Comment by xigoi
3 hours ago
The kind of subtle difference I’m talking about is that in C, you have to put a semicolon after struct {…} and do {…} while (…), but not after other curly brace constructs. What would be the analogue in Janet?
3 hours ago
The kind of subtle difference I’m talking about is that in C, you have to put a semicolon after struct {…} and do {…} while (…), but not after other curly brace constructs. What would be the analogue in Janet?
The first one is pretty C-specific and for example doesn’t exist in Java. Similarly for the second one, the reason is that `while` is used for two different constructs, which for example isn’t the case in Rust. These are just accidental complexities stemming from unfortunate design decisions in C. Having a richer repertoire of syntactic constructs doesn’t by itself imply such complications.
My point was, replacing n syntactic constructs by n functions or macros doesn’t reduce the cognitive load of having to know how each construct works. To the contrary, one can argue that everything having the same syntactic form makes it more difficult to distinguish different classes of features.
> The first one is pretty C-specific and for example doesn’t exist in Java.
It was just an example, most languages have quirks like this. I don’t know about Java, but in Rust you have the turbofish operator, whose necessity stems from using the less-than sign as both an operator and a delimiter.
> My point was, replacing n syntactic constructs by n functions or macros doesn’t reduce the cognitive load of having to know each of them.
The difference is that if you don’t know a function/macro, you can just read its documentation. If you don’t know a syntactic construct, where do you look?
Another advantage is that if you want to create new functionality similar to existing language features, it won’t stick out like a sore thumb. For example, you could create an until loop:
In Rust, it would have to look completely different from a while loop: