this post was submitted on 12 Jun 2023
492 points (100.0% liked)
Technology
37739 readers
500 users here now
A nice place to discuss rumors, happenings, innovations, and challenges in the technology sphere. We also welcome discussions on the intersections of technology and society. If it’s technological news or discussion of technology, it probably belongs here.
Remember the overriding ethos on Beehaw: Be(e) Nice. Each user you encounter here is a person, and should be treated with kindness (even if they’re wrong, or use a Linux distro you don’t like). Personal attacks will not be tolerated.
Subcommunities on Beehaw:
This community's icon was made by Aaron Schneider, under the CC-BY-NC-SA 4.0 license.
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I would argue that on the one hand you could say that the references to objects in garbage collected languages are also pointers.
On the other hand, you could argue that such references are not pointers, but then you might as well argue that references in rust are not pointers.
I just feel like "a language with pointers" is a weird way to describe a language and it isn't really something that causes the language to become fast. Pointers are low level constructs that are used all the time, and whether or not they are abstracted away in the high level language doesn't automatically make it slow or fast.
Hmm. Alright, what word would you use to differentiate C or Rust from, say, Python? I haven't actually made anything with Rust, but it sounds like you can store a reference in another structure like you can in C, but you can't AFAIK in Python.
You can store references in another structure, but you probably don't want to do this most of the time, since it's a major headache to get ownership and lifetimes right. You shouldn't think of references as pointers, but you should think of them as "borrows": you are temporarily borrowing something from the owner, but you can only do so if the owner exists. So you need to statically guarantee that for as long as you borrow it, the owner must be alive, which gets tricky when you store that borrow somewhere in some data structure. In practice, references or borrows will be short-lived, and most operations on data will be done by the owner of that data.
Underneath, references are represented by pointers, but you shouldn't think of them as pointers, but rather as something you have borrowed, so in that sense it's different from C.
Also, Python does use references everywhere, it's just implicit, and depends on the type. Try storing a list in a class: you've just stored a reference to another structure. Most things (e.g. lists, objects) are passed and stored by reference, some types like integers are copied. In Rust, you can choose whether you want to pass by reference, copy or move ownership. At this point we're still at a high level of abstraction, we don't think so much about whether this will be pointers at the low level.
But my main point is that whether you use pointers, references, or whether it's implicit or explicit doesn't make a language slow or fast, it just defines how programs are written. Rust is very fast because it's compiled to machine code and doesn't do garbage collection or have any other overhead from a "runtime". Python is relatively slow because it's interpreted. You could argue that more manual control over references/pointers can make a language faster, but it's not the main contributing factor.
I guess I could just say "Rust isn't garbage collected" but I feel like that would be meaningless to someone who doesn't think about compilation. I gravitated to manual pointer/reference control because that's the part you can actually see in the code, and it's pretty closely connected to the lack of garbage collection.
While I have your ear, "who" exactly are the owners in Rust? So far I've come to understand it from the aliasing XOR mutability perspective, so I don't really understand the more common terminology.
When you create an instance of a struct and assign it to a variable or field of another struct, that variable becomes the owner of that value. When you assign it to some other variable or pass it to a function that takes ownership, ownership will move. Otherwise, you will borrow. But there will always only be one owner for each value. That way you know exactly when to free up memory: whenever the owner is dropped.
So then when you say "most operations on data will be done by the owner of that data", do you mean ownership-taking functions? In my head a variable is like a bin, so it's odd to think of it doing any "operations" other than holding.
No, I phrased that poorly. What I meant is that if you have a struct that has some field (it owns that data), your operations on that data will be methods of that struct, and not some other struct that happens to have a reference to that struct. The latter is something people tend to do in OO languages like Java. In Rust, if a function accesses data, you usually "freshly" borrow from the owner of that data and pass it as argument to that function instead of relying on some "hidden" references somewhere in hidden in an object. Borrows will almost always be short-lived.
I don't know if any of this makes sense, I'm sorry for the bad explanation. It might make more sense if you play with Rust yourself.