3 ms·

"'[Rust is] Technology from the past come to save the future from itself' He goes on to elaborate a bit: 'Many older languages [are] better than new ones.

"'[Rust is] Technology from the past come to save the future from itself' He goes on to elaborate a bit: 'Many older languages [are] better than new ones. We keep forgetting already-learned lessons.'" Loving how this Graydon person thinks already. Same stuff pjmlp and I are always saying. Except he built something that took off massively with Mozilla's help. Extra credit there to both. Now, what might happen if this kind of thinking was applied all over the stack? I need 100 volunteers. Start with VMS-style versioning for critical files (esp system or application), clustering, distributed locking, and app API to leverage all that easily. Put that in the good Linux or BSD disros. Maybe they'll stop eating my files or the services popular on HN will crash less. If they get that done, we're adding Microsoft's approach to drivers that work if they reject that of MINIX 3 or Genode. We can also use TLA+, Coq, SPARK, and/or Rust since those were things in the past that worked too on the kinds of problems this effort will introduce. Systems and services should get really reliable after this. I won't even care about Linux kernel crashes as it will be a VM that goes down with my critical work checkpointed into another on same machine on a microkernel. A little driver reset happens then new one gets focus while old VM resets in background then joins cluster. For some extra money, this happens with physical separation in a desktop with two motherboards, one on each side, similar to old SGI Octanes. :)

qwertyuiop924 10y ago
>Loving how this Graydon person thinks already. Same stuff pjmlp and I are always saying. Rust was bootstrapped by a compiler written in OCaML. Like you and pjmlp, Graydon is a massive fan of fanatically strong typing, and most of that stuff was actually in many older languages that lost because C was an optimal solution at the time. However, at scale, C is no longer an optimal solution. Thus Rust can start eating C/C++'s lunch, and people will stop using C for what it does badly, and using it for what it does well (namely, writing small programs that do one thing and do it well, doing embedded work, kernels, and other bit hacking, and bootstrapping better languages).
- nickpsecurity 10y ago
  "Rust was bootstrapped by a compiler written in OCaML." Another wise choice. I gave them credit for that, too, when they told me. "because C was an optimal solution at the time." Whoa, we've talked about this before. It's closer to an optimal solution if you're squeezing max performance out of old hardware with a higher-than-assembly language. It wasn't optimal on ease of compilation, safety, large programs, or even performance optimization for general case. That all takes different design decisions than the personal preferences of its designers when building on Richards' work. It was also achieved by Wirth shortly after to point that C's remaining benefit could've been had as an option or tweak to Wirth's approach. Easier to tweak his for C's benefits than vice versa since his was actually designed and simple. Final proof of that was when C's inventors did that in their ideal language: an Oberon-2 improvement called Go. "Thus Rust can start eating C/C++'s lunch, and people will stop using C for what it does badly, and using it for what it does well" This is true. Again, though, due to social reasons more than tech. People been able to do that with GPL'd Ada. Rust takes off, though. The likely reason is a combo of developers preferring its syntax/semantics, how the community runs, and primarily Mozilla's effort [like Sun's Java & Microsoft's C# before it]. Like with the Richard Gabriel and the startups, details of execution matter more than how great the idea is.
  - qwertyuiop924 10y ago
    >It's closer to an optimal solution if you're squeezing max performance out of old hardware with a higher-than-assembly language. That was what was happening when C got popular.
    - pjmlp 10y ago
      Depending on the OS. On MS-DOS, there was hardly any difference between the code quality generated by the Borland C and Pascal compilers.
      - qwertyuiop924 10y ago
        But by that time, C was already very popular.
MichaelGG 10y ago
Reminds me of Don Syme, creator of F#[1] on getting generics into .NET: >It was seeing and experiencing polymorphic programming in OCaml and SML that made us persevere with the long and torturous process of convincing Microsoft that they should add such an "experimental and academic" feature as generics to their flagship runtime and languages. Of course we were in reality just making 1970s ideas work in practice, but at least now even Visual Basic has generics. http://caml.inria.fr/pub/ml-archives/caml-list/2006/11/d921ca59f12d434a3f005dfd07256d91.en.html http://caml.inria.fr/pub/ml-archives/caml-list/2006/11/d921c...

qwertyuiop924 10y ago

>Loving how this Graydon person thinks already. Same stuff pjmlp and I are always saying. Rust was bootstrapped by a compiler written in OCaML. Like you and pjmlp, Graydon is a massive fan of fanatically strong typing, and most of that stuff was actually in many older languages that lost because C was an optimal solution at the time. However, at scale, C is no longer an optimal solution. Thus Rust can start eating C/C++'s lunch, and people will stop using C for what it does badly, and using it for what it does well (namely, writing small programs that do one thing and do it well, doing embedded work, kernels, and other bit hacking, and bootstrapping better languages).

nickpsecurity 10y ago
"Rust was bootstrapped by a compiler written in OCaML." Another wise choice. I gave them credit for that, too, when they told me. "because C was an optimal solution at the time." Whoa, we've talked about this before. It's closer to an optimal solution if you're squeezing max performance out of old hardware with a higher-than-assembly language. It wasn't optimal on ease of compilation, safety, large programs, or even performance optimization for general case. That all takes different design decisions than the personal preferences of its designers when building on Richards' work. It was also achieved by Wirth shortly after to point that C's remaining benefit could've been had as an option or tweak to Wirth's approach. Easier to tweak his for C's benefits than vice versa since his was actually designed and simple. Final proof of that was when C's inventors did that in their ideal language: an Oberon-2 improvement called Go. "Thus Rust can start eating C/C++'s lunch, and people will stop using C for what it does badly, and using it for what it does well" This is true. Again, though, due to social reasons more than tech. People been able to do that with GPL'd Ada. Rust takes off, though. The likely reason is a combo of developers preferring its syntax/semantics, how the community runs, and primarily Mozilla's effort [like Sun's Java & Microsoft's C# before it]. Like with the Richard Gabriel and the startups, details of execution matter more than how great the idea is.
- qwertyuiop924 10y ago
  >It's closer to an optimal solution if you're squeezing max performance out of old hardware with a higher-than-assembly language. That was what was happening when C got popular.
  - pjmlp 10y ago
    Depending on the OS. On MS-DOS, there was hardly any difference between the code quality generated by the Borland C and Pascal compilers.
    - qwertyuiop924 10y ago
      But by that time, C was already very popular.

MichaelGG 10y ago

Reminds me of Don Syme, creator of F#[1] on getting generics into .NET: >It was seeing and experiencing polymorphic programming in OCaml and SML that made us persevere with the long and torturous process of convincing Microsoft that they should add such an "experimental and academic" feature as generics to their flagship runtime and languages. Of course we were in reality just making 1970s ideas work in practice, but at least now even Visual Basic has generics. http://caml.inria.fr/pub/ml-archives/caml-list/2006/11/d921ca59f12d434a3f005dfd07256d91.en.html http://caml.inria.fr/pub/ml-archives/caml-list/2006/11/d921c...