4 ms·

Engineers repulsion from Ada -- for being an icky DoD invention -- has probably held the art back several decades. We're only now getting measurable adoption of

Engineers repulsion from Ada -- for being an icky DoD invention -- has probably held the art back several decades. We're only now getting measurable adoption of languages like Rust, Scala, Clojure, or Haskell that attempt to assist programmers in writing safe code -- admittedly in ways that mostly supersede Ada's admirable but now outdated efforts. Now Spark is introducing further safety features, and you're jumping on the bandwagon again to make sure that no one bothers to learn from it. Please stop. If every CS101 student in the last 20 years had been taught to depend on, for example, bounded numeric types, verified pre and post conditions, and buffer and array overflow protection, the world would be a much better place. Let's try to not make the same mistakes (over and over) going forward.

hga 11y ago
Don't know that it's so much repulsion as the simple inability to play the game without playing $$$$$$ for a very long time (until 1995 for GNAT). Even now, I remember reading that the fully free GNAT runtime is said not to be all that hot. All that, and the costly tools above the language and runtime, is a major reason I'm ignoring this ecosystem in favor of (tentatively) the seL4 one. Hmmm, and how close are Ada and SPARK to being supplied by a single source?
- oconnore 11y ago
  NYU Ada/ED -- the free teaching compiler that went on to become GNAT -- was released in 1983, 4 years before the first release of GCC, the same year Stallman first published his Free Software Manifesto, and two years before the FSF was created. It's hardly for lack of compilers that people rejected Ada (I agree that the ecosystem there is lacking relative to other options in 2015). Also, FWIW, seL4 is not an ecosystem. It's a mish mash of Haskell prototype, Isabelle/HOL proofs, and C99 code put together by an extremely talented team. You might use some of their tricks, but there is no such thing as putting together an "seL4 ecosystem" software project.
  - hga 11y ago
    NYU Ada/ED was an interpreter, and per http://www.adahome.com/Resources/Compilers/Ada-Ed.html http://www.adahome.com/Resources/Compilers/Ada-Ed.html: Apart from the 100-odd tests of ACVC 1.11 that Ada/Ed failed, the major deficiency of the system is that, being an interpreter, it does not implement most representation clauses, and thus does not support systems programming close to the machine level. So it sounds suitable for getting your feet wet, but not "industrial strength", and not useful for a lot of the things we want to do. As for seL4's "ecosystem", well, it's certainly not polished like AdaCore's offerings (or so I assume, which for things like this is very iffy), but something I do believe I can clone, and seeing as I plan to start with a seL4 port, it qualifies for me.
- pjmlp 11y ago
  We used to pay for compilers for any language, not just Ada. You surely are aware that one of the reasons GCC became anything useful was the switch to paid SDKs on the UNIX world, which attracted developers not willing to pay for them?
  - hga 11y ago
    That's why I used 6 dollar signs, instead of my usual 3 (which I'll admit is not super clear). Or at least it was my impression that Ada compilers cost a lot more. And I most certainly remember that detail about GCC, having started playing that game no later than 1992 with Sun.

hga 11y ago

Don't know that it's so much repulsion as the simple inability to play the game without playing $$$$$$ for a very long time (until 1995 for GNAT). Even now, I remember reading that the fully free GNAT runtime is said not to be all that hot. All that, and the costly tools above the language and runtime, is a major reason I'm ignoring this ecosystem in favor of (tentatively) the seL4 one. Hmmm, and how close are Ada and SPARK to being supplied by a single source?

oconnore 11y ago
NYU Ada/ED -- the free teaching compiler that went on to become GNAT -- was released in 1983, 4 years before the first release of GCC, the same year Stallman first published his Free Software Manifesto, and two years before the FSF was created. It's hardly for lack of compilers that people rejected Ada (I agree that the ecosystem there is lacking relative to other options in 2015). Also, FWIW, seL4 is not an ecosystem. It's a mish mash of Haskell prototype, Isabelle/HOL proofs, and C99 code put together by an extremely talented team. You might use some of their tricks, but there is no such thing as putting together an "seL4 ecosystem" software project.
- hga 11y ago
  NYU Ada/ED was an interpreter, and per http://www.adahome.com/Resources/Compilers/Ada-Ed.html http://www.adahome.com/Resources/Compilers/Ada-Ed.html: Apart from the 100-odd tests of ACVC 1.11 that Ada/Ed failed, the major deficiency of the system is that, being an interpreter, it does not implement most representation clauses, and thus does not support systems programming close to the machine level. So it sounds suitable for getting your feet wet, but not "industrial strength", and not useful for a lot of the things we want to do. As for seL4's "ecosystem", well, it's certainly not polished like AdaCore's offerings (or so I assume, which for things like this is very iffy), but something I do believe I can clone, and seeing as I plan to start with a seL4 port, it qualifies for me.
pjmlp 11y ago
We used to pay for compilers for any language, not just Ada. You surely are aware that one of the reasons GCC became anything useful was the switch to paid SDKs on the UNIX world, which attracted developers not willing to pay for them?
- hga 11y ago
  That's why I used 6 dollar signs, instead of my usual 3 (which I'll admit is not super clear). Or at least it was my impression that Ada compilers cost a lot more. And I most certainly remember that detail about GCC, having started playing that game no later than 1992 with Sun.