One things I miss most about Go is Ternary, coming from PHP and Node where the ternary is already well established, I understand how it can be controversial since a lot of them use complex logic in them, that would fit more in IF, Else and Switch. It's like a famous quote form the comics: "With great power comes great responsibility".
The discussion about the Go and Ternary isn't new, there are a lot of issues on the topic, I'll leave her sources for you to delve into, I'll summarize the points that I think are most important on both sides, and the give my personal conclusion.
Benefits
1) Drastucally reduces the use lines code.
var n T
if expr {
n = trueVal
} else {
n = falseVal
}
var n = expr ? trueVal : falseVal
2) If the problem is the sumbols ?, :. =, an approach that not even in python can be quite interesting.
x := "yes" if true else "no"
3) Languages like C and C++, which are the example base of Go, have ternaries.
4) Variable that nedd a value default if no value is assigned to it, for example: service port.
Disadvantages
1) When there is more than one expression to be validated, the ternary operation is difficult to read and it becomes impossible to determine its consummation.
var b1, b2, b3 bool
...
var s = b1 || b2 ? b2 && b3 ? b1 && b3 : "s1" : "s2" : "s3": "s4"
2) Can replaced by operations like IF, ELSE and Switch.
3) Deviates from Go's goal of clarity and simplicity when it comes to syntax.
Conclusion
The ternary will not come to version 1.20 (at this time we are on 1.19), discussions are whether it will come in version 2 of Go, but the reality that will hardly come, the communicates it despite being divided the majority from what I see defends that should still keep the aspects of simplicity in its syntax, even if it requires more lines of code.
The biggest fear of communicating is the creation of more complex codes, I personally prefer autonomy for developers and their teams to choose the best way to write, one more option would be nice, since there are countless that make more sense to use just one line than to use at least 3.
No everything is lost, if it is really necessary there are packages in Go that create the ternaries in Go or leave it very close to what we have in other languages, example: ternary, logically we must have other scattered around Github.
Top comments (6)
A ternary is just syntactic sugar for an
if-elsestatement, it does not objectively provide any other benefit, hence it does make sense to have it very very low prioritized in the road map.By the way in the disadvantages list I'd add that it does not work well with doing a single thing in a condition (no else) even though some people uses ternaries that way (for any reason I can't comprehend).
The one use-case ternaries would have in go is assigning a conditional value to a read-only variable. Given the overall ethos of Go programming, however, that very use-case would be more of a code smell than anything else. So ultimately ternaries are just syntactic sugar with more drawbacks than benefits from a language design POV.
However, Go has its fair share of syntactic sugar already (e.g. optional return, and defer), so it's questionable whether that alone should be an argument against ternaries. Though I do tend to agree that it shouldn't make it into the syntax.
I was suprised too, but go does not provide it.
It's debatable, as it can be less readable in many cases, as some devs tend to chain ternaries :(
100% agree
Yes, they can be useful but maybe they're not in the "absolutely need that" box. I assume they'll add ternaries. The question is when.
ternary operations have become quite common in most languages nowadays that when you run into a language that does not support them in a certain way, you feel surprised. That's how I felt a few years ago with Rust, but Rust's compact if/else statement makes a lot more sense and its far more readable than a ternary operator.