Comments by "Vitaly L" (@vitalyl1327) on "Continuous Delivery" channel.

  1. What a lot of programmers fail to understand is that Turing-completeness is harmful. Every time you do something in a Turing-complete language you must justify this choice, and most of the time you'll find that there is no such justification. There's a very good reason why, for example, full MISRA-C is not Turing-complete. Now, the good thing about functional programming is that it's easy to restrict it while keeping it still very powerful and flexible - Total functional programming, as in languages like Coq. Imperative languages turn out far more restrictive when you kick Turing-completeness out of them.
    2
  2.  @DyingKlokateer  by competent I mean capable of learning and solving problems. Knowledge is cheap, ability to gain it and use it is rare. And, sadly graduating from a university does not guarantee such an ability, especially these days, when it is very hard for a student to get kicked out.
    2
  3. The shortage is massive. But the quality of candidates is abysmal, so it's really hard to find a single competent one even if you go through hundreds of them. And all those hundreds will come here and babble that "there's no shortage, I went to a hundred of interviews and not a single offer". Well, guess why you're not getting any?
    2
  4.  @Toulkun  there's not enough worthy graduates to train, and tons of worthless ones diluting the pool, that's the biggest problem.
    2
  5. This is exactly the main driver of the current shortage.
    2
  6.  @davidjulitz7446  it is a distributed system, formal proofs of soundness are indeed possible (TLA+ or simialr). And yes, lack of QA loop is also developers fault. They shoudl have raised an alarm and decline to work in such conditions.
    2
  7. Only if your OS and bootloader are utter trash. Otherwise it is trivial. And if you do not have a hardware watchdog on a mission-critical system, you're criminally negligent and must go to jail.
    2
  8. Tools are not any better. If you think your bloated development environment based on the deplorable Electron of all things is any better than an average IDE we had since 80s, think again.
    2
  9. ​ @aabhishek4911 by enforcing immutability, allowing compiler to use the fact that once value is read it won't change. To get the same level of optimisation in an imperative language you must litter your code with 'const' and carefully annotate aliasing.
    2
  10.  @chakritlikitkhajorn8730  tooling - yes, absolutely, Javascript got the best JIT compilers out there, for example. I am talking about the language itself, and the language is among the worst ever created by humanity.
    2
  11.  @chakritlikitkhajorn8730  brainfuck is not bad at all, for its purpose. It was designed to teach Turing machines and to explore how potential future TM-based computers can be used (think of, say, DNA computing). Assembly is not "horrible" too - it's just a reflection of the actual hardware. Basic is remarkably simple and can be implemented on a really rightly constrained hardware. Not bad at all either. Though Forth is even better, but for the teaching purposes Basic was the right thing to happen. While for JavaScript there is no excuse at all. No single reason for it to exist. It should have never been created in the first place.
    2
  12.  @chakritlikitkhajorn8730  language is defined by its purpose. A language that fits a purpose, which have a nice, clean semantics that reflects the purpose, the problem domain as closely as possible is a well designed language. In this regard, Brainfuck is perfect, and not just that, it's beautiful. While JS serves no purpose. It was not derived in any way from its supposed problem domain. It's inconsistent, its semantics is even hard to define (it shows that the language was just jotted in a couple of days, and no spec was written before the implementation started). If a PL theorist tried to build a language for the niche occupied by JS, it'd be a very different language. Most likely, something similar to Scheme. If a PL theorist had to create an accessible language with a dense encoding for a 4kb RAM, 8080 CPU device, they're likely to reinvent Basic, given that Forth is deemed too alien to the target audience and teaching purposes.
    2
  13. My experience does not match yours at all. Portability across the very diverse Unix world was a solved problem (autotools). Portability between *nixes and VMS was not an issue either. Caring about crap platforms like PCs was an easy to avoid mistake.
    2
  14.  @mandisaw  or, you could stick to Unix and be happy. Did so since the 80s, and my software stack now is not too different from what I had for decades.
    2
  15.  @Meritumas  you must feel like you're in the wrong sector of this industry if it can even be outsourced. And if something can, in fact, be outsourced without severe quality and communication implications, it is criminal not to outsource.
    2
  16.  @travispulley5288  lol, that's a funny take. Which letter in the word "science" you failed to understand?
    2
  17.  @travispulley5288  so, you have no idea what science is. As expected.
    2
  18. now ask any PL theorist what they think of JS. Language is objectively very poorly designed.
    2
  19.  @CripplingDuality  try to apply your flawed logic to doctors or civil engineers. Expectations are not unreasonable, they're set by the very nature of our reality.
    2
  20.  @trignals  it is an unfirtunate and misguided response of the industry to an abysmal quality of the vast majority of the candidates. Imagine something like this happening in regulated professions, like civil engineering, medicine, accounting.
    2
  21. Do you realise that Lisp existed since forever? That we had Smalltalk since 80s? That we had Perl and Tcl?
    2
  22. ​ @TheBoing2001  Again, javascript is a really crappy language, and it's not equipped with any of the powerful tools for creating higher levels of abstraction. Clearly you know very little about programming languages if you don't see how limiting JS is in comparison to the real languages. If you're very determined, sure, you can build nice things with it. But it's a massive waste of time and effort, and a better tool would have been much more suitable.
    2
  23.  @TheBoing2001  your answer got deleted, apparently, for having URLs. "Javascript is fine." - no, it is not. It's a crappy language devoid of any tools for creating higher level abstractions without suffering too high performance costs. Namely, no proper macros, and no proper control over code generation. "With 3 other click you can C&P and run entire chess game" - mind you, a chess engine was one of the first programs Konrad Zuze wrote for his primitive computer in 1945. Hardly an impressive achievement in the 21st century. Clearly you're a fanboy and you don't know anything at all about PLT, so you cannot see how deeply flawed JS is. And, no, ease of deployment is not a virtue of JS the language. It did not facilitate this cross-platform deployment in any way. It's simply the result of web standards monopoly. If Tcl/Tk was such a standard, you could deploy a Tcl/Tk application everywhere equally effortlessly. So your argument makes zero sense. Now go and learn some computer science, you're clearly lacking the most basic education.
    2
  24. And the solution is always the worst possible - to produce a lot of very low quality developers, who, in turn, will create such enormous piles of technical debt that even more developers will be needed to maintain all that mess. Hordes of code monkeys is never a right solution, it's a root of the problem. Introduce the same level of professional quality regulation as in, say, civil engineering or medicine, and things will get much better in just a decade or so.
    2
  25. Software engineering is an engineering. Obviously. So why we're not regulating it like the other branches of engineering? Why software engineers can work without any certification, why they do not have personal liability for their incompetence? High time software engineering is regulated to the same level as civil engineering or medical professions.
    2
  26. Companies are picky because they have to - a wrong developer will cause a lot of damage, not many can afford it. What's the point in hiring a destructive incompetent junior, your losses won't be that awful if you just keep the position open until you find a right candidate.
    2
  27. We had much better software in the 80 and 90s. We have much much worse developers now in 202x. So yes, development was easier, we did not have to tolerate that much awful software written by awful people.
    2
  28.  @almari3954  ideally there should not be any concurrency whatsoever in a BLDC controller. Designs tbat utilise concurrency (such as ODrive) are just wrong. A proper real-time design will be an FSM with precise timings of all measurements and PWM settings, wile communications are handled entirely independently on a separate core.
    2
  29. Lol wut? Tests are the only way to force LLMs to generate anything remotely meaningful.
    2
  30. There's only one real 10x programming in this world - Fabrice Bellard. All the other self-styled 10Xs impostors, and of course they just poison everything around them and add no value to anything.
    1
  31. LLMs write a much better code with ReAct, in languages they were not trained on (ideally, languages that did not even exist in the training set).
    1
  32.  @BigWhoopZH  "Desktop operating system" is a dumb thing that should have never existed in the first place. We never needed "desktops". What we need is *terminals*. Immutable and expendable. Microsoft have a lot to answer for, and spreading this deranged idea of a "desktop computer" is among their top crimes.
    1
  33.  @BigWhoopZH  looks like you have no idea what you're babbling about. Try to define, in your own words, what are the use cases for this so called "desktop" abomination. You'll see that for anything you can come up with there is a much better solution. There is no single reason to ever use a "desktop OS" for anything.
    1
  34.  @BigWhoopZH  the entire desktop idiom is fundamentally broken. You don't need a desktop, connected or not. There is no single use case for anything "desktop". You need terminals for data entry (which is a naturally connected used case anyway), and you don't need anything "desktop" for the local use cases.
    1
  35.  @BigWhoopZH  you did not, you failed to name a single use case for this so called "desktop".
    1
  36.  @BigWhoopZH     e-gate is a terminal application, by the very nature of the data. You cannot store it locally or even cache it locally. Where they failed is in redundancy and in the system design from thr very beginning (what kind of imbeciles do such things over WiFi?!?). And how any of the things you listed is "desktop"?!? You seem to call a "desktop" any device with a local processing. It is a massive stretch. Desktop is a far more narrow thing - an environment with UX modelled after a literal desk top, with folders, documents, tools and all that trash. A game console is not a desktop computer. A CAD workstation is not a desktop computer.
    1
  37.  @peterdz9573  the asnwers make no sense - calling any device with a local storage a "desktop" is nothing but a poor trolling attemt and do not deserve a polite answer.
    1
  38.  @BigWhoopZH  you started with a narrow definition when you babbled about "no desktop OS" that would have an immutable root by default, and after this bait switched to a broad definition. Enterprise Linux distros clearly did not fall under your initial definition, embedded distros (Android included) neither, so it is clearly on you.
    1
  39.  @peterdz9573  go back to the beginning of this thread, to the outrageous claim that "no desktop OS have such a configuration out of the box", which pretty much defined what is supposed to be a "desktop" - a consumer PC or a laptop running stock consumer-tier Windows or one of the toy Linux distros, or a Mac. Ignoring all the corporate systems.where such a thing is very often a default, at a cost of very limited end user rights.
    1
  40.  @peterdz9573  I never said "remote". Terminal can be connected locally, to a single host (or a mainframe). And yes, 99.99% of the use cases of desktops today in the corporate environment should have been terminals, and the remaining 0.01% is for specialised devices with specialised OS configurations, not some generic "desktop" bought from Currys.
    1
  41.  @peterdz9573  you can have a gui on a terminal. A more sane corporate setup is where your laptop runs an immutable and a very trimmed down OS which can only run a VNC client and nothing else, and all the work is done on, again, immutable Windows VMs running remotely. A VM is destroyed immediately when the user session is done. I worked in such environments in a few major investment banks, for example. Is it still a desktop? I would not call it that. It's a terminal. And I first used immutable X11 terminals in mid-80s, all the technology existed since then. Not using it now is a criminal negligence.
    1
  42. you're just saying that "worse is better". You're right. Humanity have a tendency for falling for the worst possible solution when multiple options are available. Did it over and over again. Javascript is a perfect example of such, an extremely awful language design, made at a time when PLT was already extremely advanced and people knew exactly how to construct languages fit for any given purpose.
    1
  43.  @CripplingDuality  why is it a nightmare scenario? At least it is better than having to comb through hordes of hopeless juniors to find the few worthy ones.
    1
  44.  @world-9644  we need high quality juniors. The current horribly diluted junior pool have an opposite effect, making it impossible to detect and nurture any talent.
    1
  45.  @world-9644  yet, they are indeed a liability, even the brightest ones, but you have to put up with some liabilities long-term. It is an optimisation problem. What makes this problem hard is the very diluted junior pool.
    1
  46. @theafterman1848 you're not very coherent, but from what's possible to decipher from your rambling I can conclude that you worked for a failing company and if you were smart enough you ran from there very quickly.
    1
  47. @theafterman1848 you're pathetic. If your argument is that I can google rants of sore losers like you, conplaining about shortage of jobs - then you know where you must stick your argument. There is a massive shortage of the experienced developers. I struggle finding any, for months. There is an abundance of trash "developers" who should have never been in this industry, and nobody needs them any more, so you may get an impression that there is no shortage, if you're ome of them.
    1
  48. @theafterman1848 only CRUD crap is ever outsourced anywhere, India included. You're pathetic and you see the very bottom of the barrel, confusing it for the rest of the industry.
    1
  49. @theafterman1848 you will never earn a living as a developer because you're trash, not because someone is "crashing the labour market" (right, to an extend that starting salaries are pushed well into six figures for good).
    1
  50. @theafterman1848 the reality is that I cannot find competent developers. And I see an abundance of worthless "developers". Shortage is real. And stop babbling about outsourcing, only the bottom of the barrel projects are ever outsourced.
    1