4 ms·

These kinds of articles are always hostages to fortune. Because I guarantee somebody, any second now, will demonstrate that what the author wanted to do was tri

These kinds of articles are always hostages to fortune. Because I guarantee somebody, any second now, will demonstrate that what the author wanted to do was trivial, and if he'd spent a bit less time trying to start flamewars on Hacker News he'd have discovered it. Not me, by the way, I don't know rust. But I reckon it'll be about three comments down.

BozeWolf 10y ago
It is already in the comments at the site of the article. To quote "Will": "On the bright side it is possible to implement a pretty elegant solution to this particular problem"
mpalme 10y ago
But in any case, it will be helpful for other newcomers to Rust because they are likely to run into the same challenges.
Amezarak 10y ago
I think part of the problem is this: if you're say, a C++ programmer, you can write C++ in most other languages and have a working program, even if it's not idiomatic. You can't in, say, SQL, Rust, or Lisp. That doesn't make them a priori harder than other languages, it just requires a shift in thinking, whereas if you're jumping to say, Java after C++, you can get started right away and hopefully pick up idiomatic Java as you go along. Instead of thinking "here's how I would implement this in C++, now how do I translate that to Rust?" you have to cut that part out and think "how do I write this in Rust?" I mentioned SQL because that's where that really, really, clicked for me: if you ever write SQL and think imperatively, you are doing it wrong, you have to think in sets, and once you do that, SQL is the easiest and most natural thing in the world. The Rust Book is pretty good, but I think it would really benefit from an in-depth section early on explaining how writing in Rust requires a paradigm shift in your thinking, and exactly how to think Rust-ically. I'd write something here, but I'm still learning how. :)
- steveklabnik 10y ago
  We're re-writing the Rust book at the moment: http://rust-lang.github.io/book/ http://rust-lang.github.io/book/ It doesn't explicitly have a section like this, exactly, but there are several "thinking in Rust" style chapters towards the end.
- kazagistar 10y ago
  (SQL is still kinda a crap language for dealing with sets, because the synatx does not represent the order in which set operation are applied or anything else; instead, they went with the "cobol" strategy of trying to make things like english.) But I agree with your premise, certainly.
gcp 10y ago
If this were a C++ problem the author wouldn't have blogged because the answer would've been on StackOverflow.
- sambe 10y ago
  Not sure if your comment was intended dismissively/jokingly, but I think also somewhat insightful. I find myself much less often really thinking about how to do something these days; if it seems remotely standard or potentially troublesome I google it. As a result, I spend less time thinking "algorithmically" about low-level stuff and probably lose something. But I also spend more time thinking (perhaps algorithmically) about the business domain, and am ultimately more productive.
  - jandrese 10y ago
    Googling for Rust solutions is dangerous since frequently the advice you find only applied to some old beta version and no longer works. Then you find a second answer but it only applies to an even older beta version. None of the examples anywhere work of course. This is my biggest frustration with the language: the manual is good for what it is, but it only covers the basics. You can see this in the article, where the author tries to implement a simple DNS cache and fails miserably. He doesn't even get to the part where you have a separate thread managing the cache and expiring old entries in the background to avoid unbounded growth and stale data. This is the sort of thing I could whip up in literally half an hour in C, but I just know the borrow checker is going to be a pain in the ass if I try to do something similar in Rust.
    - sambe 10y ago
      And your comment sounds exactly like the author to me: "I can't move transparently from languages I know well to ones I don't". Would you be able to whip up same in Haskell or Erlang? How many people claiming this would include a few bugs that Rust prevents? Learning curve is a feature for sure, and the Rust community seems to be aware of this. But expectations of minimal effort are odd.
killercup 10y ago
That's why I'm really looking forward to the point in time where "non-lexical borrows in Rust stable" and "clippy installable via rustup" converge.

BozeWolf 10y ago

It is already in the comments at the site of the article. To quote "Will": "On the bright side it is possible to implement a pretty elegant solution to this particular problem"

mpalme 10y ago

But in any case, it will be helpful for other newcomers to Rust because they are likely to run into the same challenges.

Amezarak 10y ago

I think part of the problem is this: if you're say, a C++ programmer, you can write C++ in most other languages and have a working program, even if it's not idiomatic. You can't in, say, SQL, Rust, or Lisp. That doesn't make them a priori harder than other languages, it just requires a shift in thinking, whereas if you're jumping to say, Java after C++, you can get started right away and hopefully pick up idiomatic Java as you go along. Instead of thinking "here's how I would implement this in C++, now how do I translate that to Rust?" you have to cut that part out and think "how do I write this in Rust?" I mentioned SQL because that's where that really, really, clicked for me: if you ever write SQL and think imperatively, you are doing it wrong, you have to think in sets, and once you do that, SQL is the easiest and most natural thing in the world. The Rust Book is pretty good, but I think it would really benefit from an in-depth section early on explaining how writing in Rust requires a paradigm shift in your thinking, and exactly how to think Rust-ically. I'd write something here, but I'm still learning how. :)

steveklabnik 10y ago
We're re-writing the Rust book at the moment: http://rust-lang.github.io/book/ http://rust-lang.github.io/book/ It doesn't explicitly have a section like this, exactly, but there are several "thinking in Rust" style chapters towards the end.
kazagistar 10y ago
(SQL is still kinda a crap language for dealing with sets, because the synatx does not represent the order in which set operation are applied or anything else; instead, they went with the "cobol" strategy of trying to make things like english.) But I agree with your premise, certainly.

gcp 10y ago

If this were a C++ problem the author wouldn't have blogged because the answer would've been on StackOverflow.

sambe 10y ago
Not sure if your comment was intended dismissively/jokingly, but I think also somewhat insightful. I find myself much less often really thinking about how to do something these days; if it seems remotely standard or potentially troublesome I google it. As a result, I spend less time thinking "algorithmically" about low-level stuff and probably lose something. But I also spend more time thinking (perhaps algorithmically) about the business domain, and am ultimately more productive.
- jandrese 10y ago
  Googling for Rust solutions is dangerous since frequently the advice you find only applied to some old beta version and no longer works. Then you find a second answer but it only applies to an even older beta version. None of the examples anywhere work of course. This is my biggest frustration with the language: the manual is good for what it is, but it only covers the basics. You can see this in the article, where the author tries to implement a simple DNS cache and fails miserably. He doesn't even get to the part where you have a separate thread managing the cache and expiring old entries in the background to avoid unbounded growth and stale data. This is the sort of thing I could whip up in literally half an hour in C, but I just know the borrow checker is going to be a pain in the ass if I try to do something similar in Rust.
  - sambe 10y ago
    And your comment sounds exactly like the author to me: "I can't move transparently from languages I know well to ones I don't". Would you be able to whip up same in Haskell or Erlang? How many people claiming this would include a few bugs that Rust prevents? Learning curve is a feature for sure, and the Rust community seems to be aware of this. But expectations of minimal effort are odd.

killercup 10y ago

That's why I'm really looking forward to the point in time where "non-lexical borrows in Rust stable" and "clippy installable via rustup" converge.