Comments by "HalfSourLizard" (@halfsourlizard9319) on "Exploring SKALA: Chernobyl Reactor Control Computer" video.

  1. Are you writing from the 1980s or do you just not know how to use a search engine?
    3
  2. All the tech looks at least 10 years out-of-date, given that this thing was designed in the 1970s. Retronerds: Does that feel about right?
    3
  3. It's slightly horrifying that a nuclear reactor was running on core memory, reel-to-reel tape drives, and teletypes into the 1980s. The only more-distressing aspect is that it was coded in assembly -- and you damned well know zero formal methods / mathematically-rigorous verification was done of that code: It would surprise me if they even had unit tests.
    2
  4.  @alexhajnal107  Almost like management by political agenda is a really shit idea, eh?😉
    2
  5. Okay, that people are querying a nuclear reactor with a 1960s / Apollo-DSKY-style 'user interface' is equally disturbing.
    1
  6. At least it's ICs instead of tubes ig🤷‍♀
    1
  7. Like, how were they not on digital terminals by this time!?!
    1
  8.  @alexhajnal107  Um, you may recall the whole 'meltdown and cause an international incident' part ... which suggests broke af.
    1
  9.  @alexhajnal107  Yep! Just like most aeroplane crashes and car accidents ... Just have to eliminate the meathumans to increase safety.
    1
  10.  @alexhajnal107  Some glorious day, we'll be rid of humans entirely. Just us lizards + AGI. It'll be glorious.
    1
  11. I mean, emulate, yes!? This thing isn't doing anything beyond the Turing Limit.
    1
  12. But, they're not doing HPC ... I would have thought that reliability was a few orders of magnitude more important than perf.
    1
  13.  @ironhead2008  Formal proof of correctness seems significantly more important. A higher-level, referentially-transparent language would be significantly easier to prove anything about ... and -- because of better abstraction capabilities -- more tractable to make any mathematically-rigorous claims about.
    1
  14.  @ironhead2008  For instance: 'I wrote my assembly code reaaal carefully.' is an infinitely-less-convincing argument that deadlock is impossible than a proof in a formal logic.
    1