CODEWITHSUNDEEP
structrust
Canleskis
Canleskis

Reputation: 113

When to use a field or a method in Rust structs?

being quite new to Rust and programming in general I have a basic question that is probably easily answered. Let's say I have a rectangle object, which has a width and a height, like so:

struct Rectangle1 {
    width: u32,
    height: u32,
}

I can create such an object with a new method:

fn new(width: u32, height: u32) -> Self {
    Rectangle1 { 
        width, 
        height,
    }
}

Now, let's say I want to use that object later and need its area, what is better practice? Have the area a field or a method?

So either I just implement this:

fn area(&self) -> u32 {
    self.width * self.height
}

Or, since area is inherent to that object, I give the object an "area" field:

struct Rectangle2 {
    width: u32,
    height: u32,
    area: u32,
}

And implement the new method like this instead:

fn new(width: u32, height: u32) -> Self {
    Rectangle2 { 
        width, 
        height,
        area: width * height,
    }
}

Now somewhere else in the code, when the area is needed:

let rect1 = Rectangle1::new(30, 50);

let rect2 = Rectangle2::new(30, 50);

println!(
    "The area of the rectangle 1 is {} square pixels.",
    rect1.area()
);
println!(
    "The area of the rectangle 2 is {} square pixels.",
    rect2.area
);

In this simple example I can't see when one would be better than the other. What should be preferred anyway? Is there one way that is less common because of something I am not aware of?

Upvotes: 3

Views: 677

Answers (1)

Ahmed Masud
Ahmed Masud

Reputation: 22402

TLDR; it depends on the use case.

IMHO

This starts to become opinionated very fast. Generally speaking your answer will be driven by use-cases.

On modern systems, experience says that it's better to keep dependent params as functions and only optimize the caching of results as a special case.

Example 1:

Length and height remain constant over life of the Rectangle; pre-calculating may be useful. (Consider having to do it for 10^6 rectangles e.g.)

Example 2:

Your height and length get modified ... then do you precalculate? Do you cache the result?

Example 3:

You are constantly updating the third param based on the user updating any of the other two :-)

@prog-fh made a comment about accessing memory is expensive. I would say that you have to consider that both ways. Fetching two values into the CPU and calculating it can be potentially more expensive than accessing exactly one pre-calculated value.

so IMHO, and in line with what every one says,

It depends :-)

Upvotes: 1

Related Questions