3 ms·

> While not perfect, given its copy-paste compatibility with C, you can opt-in into bounds checking by using the STL data types, or similar. I've heard of abus

> While not perfect, given its copy-paste compatibility with C, you can opt-in into bounds checking by using the STL data types, or similar. I've heard of abuses of templates to add something that resembles proper ranged types, but it is not a path I'd want to go down. Partly, because C++ templates are not an excuse for a proper type system, even if they are a tool that can be used to force a type system into the language. A proper type system lets you do things like add centimeters and meters together, divide by seconds, and pop out a proper unit and ensure no overflow happens. A proper type system lets you easily, without the visual noise or compile time hit that templates have, add bounds checking wherever it is appropriate. type Age is Integer range 0 .. 130 type Day_Of_Week is (Sunday, Monday, Tuesday, Wednesday, Thursday, Friday, Saturday); subtype Working_Day is Day_Of_Week range Monday .. Friday; Take a look at https://en.wikibooks.org/wiki/Ada_Programming/Type_System#Defining_new_types_and_subtypes https://en.wikibooks.org/wiki/Ada_Programming/Type_System#De... For all the flak it gets, Ada did this right long ago. And a huge part of all this is compile time only, and you can opt in/out of the runtime checks as performance dictates. I'd kill for a proper "milliseconds" type everywhere, one that can only have other milliseconds added to it. And then double the fun with a "seconds" type, and a language that keeps track of which is which for me. Why the heck am I writing timeInMs or timeInSeconds still? Subtyping integers should be a basic and primitive operation in any modern language.

pjmlp 6y ago
I know most Wirth based languages relatively well. Actually your examples are even possible in Pascal. type Age = 0 .. 130; Day_Of_Week = (Sunday, Monday, Tuesday, Wednesday, Thursday, Friday, Saturday); Working_Day = Monday .. Friday; However in 1992, C++ was a viable path forward on PC, whereas Ada was too expensive and Modula-2 compilers weren't getting the love they needed from OS vendors (https://freepages.modula2.org/compi.html https://freepages.modula2.org/compi.html), so C++ was the best alternative to move from Object Pascal into cross platform development. So while C++ isn't perfect, its copy-paste compatibility was key for its adoption, but also the reason why C++ will never be Ada in regards to security. Just like TypeScript or the JVM guest languages, have to deal with the semantics of the underlying platform. In any case, in what concerns opt-in, C++ provides the options, and using C already in 1992 was a matter of being religiously against C++. RAII, proper strings and arrays data structures with bounds checking and reference parameters instead of pointers were already available. From the 5 surviving Ada vendors, there is a free option, so anyone that cares about security with the issues that C brings into C++, can also have a look at it. In fact, GenodeOS, seL4 and NVidia Autonomous Vehicle project are all using Ada/SPARK to some extent, as they considered investing into Ada a better option than adopting Rust, given its history in production.
- q-big 6y ago
  > In fact, GenodeOS, seL4 and NVidia Autonomous Vehicle project are all using Ada/SPARK to some extent, As far as I am aware, seL4 does not use Ada/SPARK, but C (with some assembler snippets), and Isabelle/HOL (for the proofs). This is also what is written in https://wwwnipkow.in.tum.de/teaching/proof21/SS18/files/14-final.pdf https://wwwnipkow.in.tum.de/teaching/proof21/SS18/files/14-f...
  - pjmlp 6y ago
    My mistake, I thought that Componolit work was for seL4, but it is also for Genode. https://componolit.com/index_en.html#components https://componolit.com/index_en.html#components
jstimpfle 6y ago
> Why the heck am I writing timeInMs or timeInSeconds still? Because "keeping track of types" is much more complicated than one would think at first. How many languages are there that can do something like (3A * 4V == 12VA), (12VA / 3A == 4A)? What if you want to use a generic routine that is just about the integers? It will lose track of all the types. I claim that there is still no nice way to do this stuff except scripting languages (dynamic types). One of the languages that go a little more into types than mainstream languages but still are somewhat close to mainstream is Haskell. Have you ever tried to use it? You get quickly sucked down into a sea of language and type extensions that you have to use, and that are often mutually incompatible, most of them probably the result of a PhD thesis. And you never quite get at what you want (if you want sth slightly differend the card house tumbles and you have to pick a different set of extensions). That is unless you're very, very smart, and have so much investment into Haskell that you know the current limits, so you're able to choose a model that is still within the bounds of representability in the language. For all of us mere mortals and practitioners, we use simple integers and get our job done. For everyone still clinging on the idea of types, I strongly believe the type system vs runtime values dichotomy is a dead end. Sometimes you need dynamic computations with types - sometimes types of expressions are known statically, but sometimess not. That's why most mainstream type systems keep it simple and provide escape hooks (like void-pointers or RTTI).
- tome 6y ago
  > You get quickly sucked down into a sea of language and type extensions that you have to use, and that are often mutually incompatible "Often mutually incompatible" suggests that you should be able to name, say, ten such pairs of extensions. Can you? (I've used Haskell professionally for nearly eight years and I can't name one such pair. I could probably find a few marginal cases if I spent 30 mins, but they don't come up in practice.) > most of them probably the result of a PhD thesis Can you name any extensions that are the result of "just" a PhD thesis (as opposed to a peer reviewed academic paper published at a top conference)? > if you want sth slightly differend the card house tumbles and you have to pick a different set of extensions Can you name any examples where a card house tumbled and someone had to pick a different set of extensions?

pjmlp 6y ago

I know most Wirth based languages relatively well. Actually your examples are even possible in Pascal. type Age = 0 .. 130; Day_Of_Week = (Sunday, Monday, Tuesday, Wednesday, Thursday, Friday, Saturday); Working_Day = Monday .. Friday; However in 1992, C++ was a viable path forward on PC, whereas Ada was too expensive and Modula-2 compilers weren't getting the love they needed from OS vendors (https://freepages.modula2.org/compi.html https://freepages.modula2.org/compi.html), so C++ was the best alternative to move from Object Pascal into cross platform development. So while C++ isn't perfect, its copy-paste compatibility was key for its adoption, but also the reason why C++ will never be Ada in regards to security. Just like TypeScript or the JVM guest languages, have to deal with the semantics of the underlying platform. In any case, in what concerns opt-in, C++ provides the options, and using C already in 1992 was a matter of being religiously against C++. RAII, proper strings and arrays data structures with bounds checking and reference parameters instead of pointers were already available. From the 5 surviving Ada vendors, there is a free option, so anyone that cares about security with the issues that C brings into C++, can also have a look at it. In fact, GenodeOS, seL4 and NVidia Autonomous Vehicle project are all using Ada/SPARK to some extent, as they considered investing into Ada a better option than adopting Rust, given its history in production.

q-big 6y ago
> In fact, GenodeOS, seL4 and NVidia Autonomous Vehicle project are all using Ada/SPARK to some extent, As far as I am aware, seL4 does not use Ada/SPARK, but C (with some assembler snippets), and Isabelle/HOL (for the proofs). This is also what is written in https://wwwnipkow.in.tum.de/teaching/proof21/SS18/files/14-final.pdf https://wwwnipkow.in.tum.de/teaching/proof21/SS18/files/14-f...
- pjmlp 6y ago
  My mistake, I thought that Componolit work was for seL4, but it is also for Genode. https://componolit.com/index_en.html#components https://componolit.com/index_en.html#components

jstimpfle 6y ago

> Why the heck am I writing timeInMs or timeInSeconds still? Because "keeping track of types" is much more complicated than one would think at first. How many languages are there that can do something like (3A * 4V == 12VA), (12VA / 3A == 4A)? What if you want to use a generic routine that is just about the integers? It will lose track of all the types. I claim that there is still no nice way to do this stuff except scripting languages (dynamic types). One of the languages that go a little more into types than mainstream languages but still are somewhat close to mainstream is Haskell. Have you ever tried to use it? You get quickly sucked down into a sea of language and type extensions that you have to use, and that are often mutually incompatible, most of them probably the result of a PhD thesis. And you never quite get at what you want (if you want sth slightly differend the card house tumbles and you have to pick a different set of extensions). That is unless you're very, very smart, and have so much investment into Haskell that you know the current limits, so you're able to choose a model that is still within the bounds of representability in the language. For all of us mere mortals and practitioners, we use simple integers and get our job done. For everyone still clinging on the idea of types, I strongly believe the type system vs runtime values dichotomy is a dead end. Sometimes you need dynamic computations with types - sometimes types of expressions are known statically, but sometimess not. That's why most mainstream type systems keep it simple and provide escape hooks (like void-pointers or RTTI).

tome 6y ago
> You get quickly sucked down into a sea of language and type extensions that you have to use, and that are often mutually incompatible "Often mutually incompatible" suggests that you should be able to name, say, ten such pairs of extensions. Can you? (I've used Haskell professionally for nearly eight years and I can't name one such pair. I could probably find a few marginal cases if I spent 30 mins, but they don't come up in practice.) > most of them probably the result of a PhD thesis Can you name any extensions that are the result of "just" a PhD thesis (as opposed to a peer reviewed academic paper published at a top conference)? > if you want sth slightly differend the card house tumbles and you have to pick a different set of extensions Can you name any examples where a card house tumbled and someone had to pick a different set of extensions?