Announcement

Collapse
No announcement yet.

Latest Trolling? The Linux Kernel In Perl

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #31
    Originally posted by Tiger_Coder View Post
    All that trolling aside, may be it is not so bad to have a interpreted language support in kernel level.
    Yeah, and maybe it is not so bad to have the critical components of an internal combustion engine made of playdough, you know, to add some "flexibility" and all.

    Comment


    • #32
      IIRC, there was an OS written in Haskell a while back. Development stopped, but it was fairly functional.

      In fact...here's a stackoverflow question with a fairly comprehensive answer: http://stackoverflow.com/questions/6...ten-in-haskell

      I think I was thinking of House.

      Comment


      • #33
        ^^ You don't have to do something just because it can be done. There is so much stupid stuff being done on this premise, utter waste of resources, and for what, just because you are in a desperate need to shine with something "original" and you could not thing of anything actually original and instead opt for something no one has done simply because it is pointless.

        Perl is anywhere between 2x and 200x slower than C - THE language of choice for kernel development, and the distribution is predominantly in the upper regions of this range. Also the very notion of using a dynamic interpreted language to write kernels sounds bad and dangerous to begin with. Surely, it might be useful to "open up" a C kernel for a narrow, safe and specific context to modify with a dynamic language, but writing entire kernels sounds like a terrible idea. And yes, I know performance is not the utmost concern when writing kernels, but it is still very important in the majority of cases, but running Perl code in kernel mode simply doesn't sound like a good idea.

        Comment


        • #34
          Originally posted by ddriver View Post
          ^^ You don't have to do something just because it can be done. There is so much stupid stuff being done on this premise, utter waste of resources, and for what, just because you are in a desperate need to shine with something "original" and you could not thing of anything actually original and instead opt for something no one has done simply because it is pointless.

          Perl is anywhere between 2x and 200x slower than C - THE language of choice for kernel development, and the distribution is predominantly in the upper regions of this range. Also the very notion of using a dynamic interpreted language to write kernels sounds bad and dangerous to begin with. Surely, it might be useful to "open up" a C kernel for a narrow, safe and specific context to modify with a dynamic language, but writing entire kernels sounds like a terrible idea. And yes, I know performance is not the utmost concern when writing kernels, but it is still very important in the majority of cases, but running Perl code in kernel mode simply doesn't sound like a good idea.
          Just FYI, Haskell can be compiled (see, ghc), and after compilation code can run at about the same speed as compiled C/C++ code, though may use more memory. That being said, it's probably not best suited for speed-critical or low-memory applications, but the language has other advantages if you're willing to overlook those flaws, or if they don't apply to the program you're writing (i.o.w., when written in another language it'd use the same resources/be as efficient). See The Benchmarks Game and this Stack Overflow question. Obviously, Perl is a different case altogether, but it does have it's advantages. Perl, IMHO, wouldn't be suited for a kernel (at least, unless some compiler is made for it, and then there may be other trade-offs).

          Now, to pull out my straw-man's hat. If everyone decided not to do things, because "it can be done" (as you say), a lot of good things would never be invented or so. Imagine if someone said "why are you wasting your time on creating that stupid levitating train; it's too expensive to operate, and we already have working trains that do just fine!" Well, guess you can forget about traveling 200+ miles in under an hour without flying. Or maybe someone said "internal combustion engine? pfft, we've already got steam engines, why would we want that? It's horribly inefficient anyway!" (this was done for electric vehicles a long time ago, by the way, though inefficiency wasn't the reason they were dismissed, but rather effective range of operation)

          Comment


          • #35
            Originally posted by ddriver View Post
            ^^ You don't have to do something just because it can be done. There is so much stupid stuff being done on this premise, utter waste of resources, and for what, just because you are in a desperate need to shine with something "original" and you could not thing of anything actually original and instead opt for something no one has done simply because it is pointless.
            And yet here you are, arguing on an internet forum. Just imagine all the productive and original things you could have been doing instead...

            Comment


            • #36
              Originally posted by Nobu View Post
              Now, to pull out my straw-man's hat. If everyone decided not to do things, because "it can be done" (as you say), a lot of good things would never be invented or so. Imagine if someone said "why are you wasting your time on creating that stupid levitating train; it's too expensive to operate, and we already have working trains that do just fine!" Well, guess you can forget about traveling 200+ miles in under an hour without flying. Or maybe someone said "internal combustion engine? pfft, we've already got steam engines, why would we want that? It's horribly inefficient anyway!" (this was done for electric vehicles a long time ago, by the way, though inefficiency wasn't the reason they were dismissed, but rather effective range of operation)
              It's really not that similar, as all those examples had the possible benefits explained before doing anything. A kernel in haskell has its possible benefits explained (haskell guarantees no side effects, so you get a trivially parallelizable kernel, which should help with scalability), internal combustion engine means you can use an alternate energy source (while previously it was coal or coal, and oil looked far more abundant at the time), the train, you named it yourself, it can help you travel fast and efficiently. Perl kernel? None explained. Maybe there are possible advantages, but the guy didn't bother explaining, if he even thought of any. That being said, it's his time. I'm never against anyone having hobbies. He shouldn't expect having it mainlined, though.

              Comment


              • #37
                Originally posted by mrugiero View Post
                It's really not that similar, as all those examples had the possible benefits explained before doing anything. A kernel in haskell has its possible benefits explained (haskell guarantees no side effects, so you get a trivially parallelizable kernel, which should help with scalability), internal combustion engine means you can use an alternate energy source (while previously it was coal or coal, and oil looked far more abundant at the time), the train, you named it yourself, it can help you travel fast and efficiently. Perl kernel? None explained. Maybe there are possible advantages, but the guy didn't bother explaining, if he even thought of any. That being said, it's his time. I'm never against anyone having hobbies. He shouldn't expect having it mainlined, though.
                Well, you have to drop a ball first before being able to say that it will (with certainty) bounce...unless you do some math or have prior examples.

                Comment


                • #38
                  Originally posted by Nobu View Post
                  Well, you have to drop a ball first before being able to say that it will (with certainty) bounce...unless you do some math or have prior examples.
                  But you don't have to to speculate about some advantages; the fact that the last word is in experimental evidence doesn't change this. I've seen not even a speculation of those, so what would motivate me (or any actually skilled person, instead) to help? As I said, he can do whatever he wants with his time, I'm not against the guy having a hobby, but adoption will require colaboration in some degree, and for that, his idea must awake some interest in skilled people.

                  Comment

                  Working...
                  X