ITT we plan the roadmap of computing for the next 30 years

The current software that we all use everyday is the most unnecessarily convoluted that it has ever been in history.

Windows is bloated buggy shit, Mac is bloated buggy shit, Linux is bloated buggy shit.

How do we change it? What is the solution?

New languages like Rust? Microkernels? Virtualization? Containers? Capability based OS? Model checking? Formal verification?

DMT Has Friends For Me Shirt $21.68

Yakub: World's Greatest Dad Shirt $21.68

DMT Has Friends For Me Shirt $21.68

  1. 2 months ago
    Anonymous

    how about you fix linux desktop first? Where is HDR support?

    • 2 months ago
      Anonymous

      >how about you fix linux desktop first? Where is HDR support?
      X12

    • 2 months ago
      Anonymous

      Linux is hard to fix because it's a legacy system full of 30 year old cruft. *Especially* X11. Windows is the same way but worse. Notice how every release basically they add a new GUI library but still keep all the old code running in the background for backward compatibility? On windows 11 there's a bug where Adobe Reader will switch to the Windows 7 style window decorations sometimes.
      And it's written in C which is an obsolete language. C was obsolete already when it was invented, for example Pascal was similar but superior. There are some pretty incontrovertible design decisions that should be made outright. UB as a concept shouldn't exist. Architectural differences should be handled with platform specific code, not different (unspecified) behavior for the same exact code on different platforms. Bounds checking should be enabled by default with a compiler flag to be disabled for very specific workloads if any. Strings should be length prefixed and not null terminated. What other changes to make beyond that depend on your philosophy. Do you want a more functional or procedure language? Should we prioritize simplicity or increase complexity adding features that result in faster code development or more safety and easier static analysis? Rust for example takes the route of having a more complex language and a more functional approach with static typing but still focusing on performance. Lua takes the path of having a simpler procedural language. Python focuses on having a complex dynamically typed language with fast prototyping that is colossally slow. Js has decades worth of cruft and is fairly complex but manages to be very fast and has more or less successfully tacked on many functional programming concepts.
      Haskell is functional and pretends to be "mathematical" but still is so complex that nobody has bothered to make a formal spec of the language, while even C has a formal spec (Frama-C). Overall the trend seems to be toward FP and more complexity.

      • 2 months ago
        Anonymous

        >Windows is the same way but worse
        yet windows has HDR. However bad the Linux/X11 code is, is it really worse than proprietary Windows code?
        IMO the problem is working on this stuff is not easy or sexy/rewarding. Proprietary devs do what they're told, so you don't have to motivate them.

        • 2 months ago
          Anonymous

          They have sufficient manpower to brute force the problems, but still it's going to have bugs that would've been caught had the system been simpler, and it's going to be much more expensive for x amount of features.

          correct answer

          [...]
          I dispute "bloated" because every layer is a necessary layer of abstraction, if not for the computer then for the humans who develop the software.
          autistics who cry about "bloat" and spend their time writing custom memory allocators in their programming socks have very little effect on the state of software development for actually productive humans

          No, the problem is that there are not enough layers of abstraction. Abstractions reduce the overall size of the code. C has a very low level of abstraction. C++ has some more but is much more complex which makes it doubtful whether it ends up being any better.
          >autistics who cry about "bloat" and spend their time writing custom memory allocators in their programming socks have very little effect on the state of software development for actually productive humans
          Autism is a fake illness because it groups smart people who lack social experience with actual mentally ill people.
          Would Dijkstra or Hoare be "autistics"? Do you think they would be ok with the current state of computer science? Do you think they would be the type to have homosexual sex and post pics wearing women's clothing?

      • 2 months ago
        Anonymous

        >he thinks Rust solves this problem
        ngmi, and lol

        • 2 months ago
          Anonymous

          Never said that. I don't like Rust.
          I don't know anything about Rust but right off the bat it's nearly undecipherable without learnint the syntax. I don't like that. Unless there is a really really good reason not too it should be easy to read for someone familiar with other languages.
          It's also too complex. IMO ideally a language should have a spec that fits in a 50 page booklet.

  2. 2 months ago
    Anonymous

    >Mac is bloated buggy shit
    Mine isn't.

  3. 2 months ago
    Anonymous

    You take away the interface and put it all in the cloud.
    The future of computing is a catgirl waifu maid personal assistant that only you can see via either AR glasses or a neural implant and who does all your computing tasks for you.

  4. 2 months ago
    Anonymous

    >How do we change it? What is the solution?
    I came to the realization that the only solution is to write my own software and minimize the use of other peoples software as much as is feasible.
    In the past people wrote their own tools because there were no tools. Then came a time where the tools people wrote were actually decent so people just used what was available. Now the software that's available completely suck (electron, java, name it) so we're back to writing our own software.

    We've gone full circle.

    • 2 months ago
      Anonymous

      All the software you write yourself is trivial. What can you write yourself, a window manager and a text editor? The non trivial parts of the stack (kernel, compiler, graphics system, web browser) are going to be written by someone else because by necessity to post on this site there are only two browser engines that pass the cloudflare checks.

      >Mac is bloated buggy shit
      Mine isn't.

      >have shitty Internet connection
      >applications refuse to start because macos needs to phone home every time you run a program
      Oh yeah, definitely a dependable and robust system.

      • 2 months ago
        Anonymous

        >What can you write yourself, a window manager and a text editor?
        >The non trivial parts of the stack (kernel, compiler, graphics system, web browser) are going to be written by someone else
        This entirely depends on your needs. You can cross out kernel and compiler if you write your own Forth, but a non-optimizing C compiler and a tiny kernel written in C should be in the realm of possibilities too, but a lot more difficult than a Forth.
        For the graphics: with modern CPUs and SIMD, 2D graphics without hardware acceleration can be decent. I don't know if it's feasible to write your own drivers for modern GPUs (or even modern hardware for that matter), I would have to read into how radeon/amdgpu drivers work under the hood but it's low priority for me and on my long todo list and I don't have high hopes for that. Maybe someone else has already looked into how hard that would be and can share their insight? I'm first focusing on user space applications that I use every day. The other issue with writing drivers would be the large time investment.

        For the browser I would be entirely happy with the functionality of an early 2000s web browser with text rendering and inline images as long as it rendered text and aligned blocks in a smart way making modern websites readable in them. From my experience, minimalist browsers that exist unfortunately don't render pages in a readable way. All I need is being able to read the pages, something like Firefox's reader view would be more than adequate as long as it was good. It doesn't even need TLS, you could just MITM proxy it on a Linux machine and call it a day (until you ever get the time to implement TLS, if ever).

        For everything else I would just use a second machine with a full modern software stack like Linux to do whatever legal shit and work I have to do, like online banking, shopping and humiliation rituals like filing taxes.

        • 2 months ago
          Anonymous

          >2000s web browser
          I like links2.

          >How do we change it? What is the solution?
          I came to the realization that the only solution is to write my own software and minimize the use of other peoples software as much as is feasible.
          In the past people wrote their own tools because there were no tools. Then came a time where the tools people wrote were actually decent so people just used what was available. Now the software that's available completely suck (electron, java, name it) so we're back to writing our own software.

          We've gone full circle.

          I completly agree on this one
          You still could read others peoples code to see if it's fit using it or not, But still maybe you don't get to understand it which is why you should then write your own.

          • 2 months ago
            Anonymous

            >links2
            I'll give it another shot.

            >This entirely depends on your needs.
            No, we're not talking about my needs. You were the one who claimed "we're back to writing our own software".
            >You can cross out kernel and compiler if you write your own Forth, but a non-optimizing C compiler and a tiny kernel written in C should be in the realm of possibilities too, but a lot more difficult than a Forth.
            Yeah, but you wont get Gecko or Webkit to run on your homebrew kernel and compile on your homebrew compiler.
            It being optimizing or not is probably the least of our problems, with modern hardware I bet even if you compiled the whole stack with -O0 it would be perfectly usable for web browsing.
            >For the graphics: with modern CPUs and SIMD, 2D graphics without hardware acceleration can be decent. I don't know if it's feasible to write your own drivers for modern GPUs (or even modern hardware for that matter), I would have to read into how radeon/amdgpu drivers work under the hood but it's low priority for me and on my long todo list and I don't have high hopes for that. Maybe someone else has already looked into how hard that would be and can share their insight? I'm first focusing on user space applications that I use every day. The other issue with writing drivers would be the large time investment.
            It's not about putting an image on the screen, that's the easy part. The problem is having an X11 and OpenGL/Vulkan implementation if you want to be able to run a web browser.
            >For the browser I would be entirely happy with the functionality of an early 2000s web browser
            So you would be happy not being able to post here or watch youtube without using an external player?
            At that point why bother?
            My purpose is making computing systems less frustrating to use, not more.

            >No, we're not talking about my needs. You were the one who claimed "we're back to writing our own software".
            It does depend on your needs, only you know what you want to do on your daily machine. While not everyone, yes, see suckless.
            >Gecko or Webkit
            >X11 and OpenGL/Vulkan
            These are your wants and needs, not mine. I'd rather have direct hardware access by targeting the GPUs instruction set directly, but that probably comes with its own slew of issues so I'd probably settle with 2D software rendering with SIMD in the meantime because it's attainable.
            >The problem is having an X11 and OpenGL/Vulkan implementation if you want to be able to run a web browser.
            If you pull in the entire stack and want to support everything under the sun like a modern browser, then you'll ultimately end up with something like Linux. You either want a simple system or you pull in the complexity that comes with all that, you can't have both. This is why I'm suggesting that compartmentalization may actually be a reasonable middle ground if it works for you. I don't know if it works for me, I'll have yet to figure that out.
            >So you would be happy not being able to post here or watch youtube without using an external player?
            My daily computing is mostly limited to searching and researching stuff online, mostly limited to viewing text/images. I could get by just fine with an external player for the rare occasion that I do watch Youtube videos but even then it would be possible to automate that. As for posting on IQfy, it might be a good time for me to take a few years off of IQfy again or end up writing a selenium driver or vnc viewer to post through a Linux machine whenever that time comes as I'd need a Linux machine for TLS anyway.

          • 2 months ago
            Anonymous

            >It does depend on your needs, only you know what you want to do on your daily machine.
            You claimed that you write your own software. That has nothing to do with my needs, it has to do with yours. But you are not posting using a homegrown kernel and browser, so it's nothing more than a wishful fantasy.
            >These are your wants and needs, not mine.
            These are your wants and needs by the fact that you post on this site.
            >While not everyone, yes, see suckless.
            Oh yeah, let's LARP by rewriting the trivial parts of the stack (the window manager and the wrapper around webkit) while ignoring the non trivial parts that take up millions of code each (Linux, X11, Webkit).
            >If you pull in the entire stack and want to support everything under the sun like a modern browser, then you'll ultimately end up with something like Linux.
            Linux is written in a shit language (C). Even assuming Linux is optimal, you could make the code safer and shorter by rewriting it in a better language. Even rewriting it in C, do you know how drivers are versioned in Linux? Each slightly different device model from a manufacturer gets the whole driver copy pasted and modified to work with the new device rather than modifying the original code to support another device even if the changes are minor.
            And you know Linux tries to maintain backward compatibility in perpetuity. So even if a new syscall architecture is devised, you will still need to keep around the code to handle the syscalls going back to the first version of Linux.
            And even if you end up with something like Linux, Linux has bugs. You could improve Linux by finding and removing all the bugs. And that's where model checking and specs and documentation and proofs and formal verification come into play. All of which are mostly absent from Linux.
            George Hotz with the whole AMD saga has shown how sloppy kernel driver development is and the kind of slop that gets into the kernel despite all the memes about Linus being super strict.

          • 2 months ago
            Anonymous

            >But you are not posting using a homegrown kernel and browser, so it's nothing more than a wishful fantasy.
            >These are your wants and needs by the fact that you post on this site.
            Clearly you are having trouble understanding what I'm writing so instead of trying to improve computing maybe you should first start with improving your reading comprehension.
            >while ignoring the non trivial parts that take up millions of code each (Linux, X11, Webkit).
            >Oh yeah, let's LARP by rewriting the trivial parts of the stack
            You didn't get the part where I said I don't want any of that. I don't want those million lines of code. A single person can not implement all those features. My point is to _significantly_ reduce the feature set to the point where a single person _could_ implement them. Would that reduced feature set suit your needs? Only you can tell. Would it suit mine? Maybe.
            >Linux is written in a shit language (C).
            I'll throw your feces right back at you: where's your Linux written in your great language with model checking, proper specs and documentation and proofs and formal verification? Thought so. Yeah let's LARP.

            >My daily computing is mostly limited to searching and researching stuff online, mostly limited to viewing text/images. I could get by just fine with an external player for the rare occasion that I do watch Youtube videos but even then it would be possible to automate that. As for posting on IQfy, it might be a good time for me to take a few years off of IQfy again or end up writing a selenium driver or vnc viewer to post through a Linux machine whenever that time comes as I'd need a Linux machine for TLS anyway.
            And what's the purpose of that?
            If you are worried about security then wouldn't it be easier to keep two different laptops, one that only can view static web 1.0 content and another laptop to post on IQfy?
            And what do you do that requires so much privacy anyway? Obviously you're not going to be doing online banking on links2, so what? Watching porn pics? Keeping a journal? Downloading bomb making instructions?

            >And what's the purpose of that?
            I don't want to depend on the modern software ecosystem and given my requirements it's actually feasible not to, for the most part.
            >And what do you do that requires so much privacy anyway? Obviously you're not going to be doing online banking on links2, so what? Watching porn pics? Keeping a journal? Downloading bomb making instructions?
            What do you mean? I'd be doing the same things I'm doing now: researching stuff online, reading, writing notes, programming. Why would I need those trillions of lines of code that do god knows what for that?

          • 2 months ago
            Anonymous

            >My daily computing is mostly limited to searching and researching stuff online, mostly limited to viewing text/images. I could get by just fine with an external player for the rare occasion that I do watch Youtube videos but even then it would be possible to automate that. As for posting on IQfy, it might be a good time for me to take a few years off of IQfy again or end up writing a selenium driver or vnc viewer to post through a Linux machine whenever that time comes as I'd need a Linux machine for TLS anyway.
            And what's the purpose of that?
            If you are worried about security then wouldn't it be easier to keep two different laptops, one that only can view static web 1.0 content and another laptop to post on IQfy?
            And what do you do that requires so much privacy anyway? Obviously you're not going to be doing online banking on links2, so what? Watching porn pics? Keeping a journal? Downloading bomb making instructions?

        • 2 months ago
          Anonymous

          >This entirely depends on your needs.
          No, we're not talking about my needs. You were the one who claimed "we're back to writing our own software".
          >You can cross out kernel and compiler if you write your own Forth, but a non-optimizing C compiler and a tiny kernel written in C should be in the realm of possibilities too, but a lot more difficult than a Forth.
          Yeah, but you wont get Gecko or Webkit to run on your homebrew kernel and compile on your homebrew compiler.
          It being optimizing or not is probably the least of our problems, with modern hardware I bet even if you compiled the whole stack with -O0 it would be perfectly usable for web browsing.
          >For the graphics: with modern CPUs and SIMD, 2D graphics without hardware acceleration can be decent. I don't know if it's feasible to write your own drivers for modern GPUs (or even modern hardware for that matter), I would have to read into how radeon/amdgpu drivers work under the hood but it's low priority for me and on my long todo list and I don't have high hopes for that. Maybe someone else has already looked into how hard that would be and can share their insight? I'm first focusing on user space applications that I use every day. The other issue with writing drivers would be the large time investment.
          It's not about putting an image on the screen, that's the easy part. The problem is having an X11 and OpenGL/Vulkan implementation if you want to be able to run a web browser.
          >For the browser I would be entirely happy with the functionality of an early 2000s web browser
          So you would be happy not being able to post here or watch youtube without using an external player?
          At that point why bother?
          My purpose is making computing systems less frustrating to use, not more.

        • 2 months ago
          Anonymous

          >You can cross out kernel and compiler if you write your own Forth, but a non-optimizing C compiler and a tiny kernel written in C should be in the realm of possibilities too, but a lot more difficult than a Forth.
          C is the reason modern software sucks.

    • 2 months ago
      Anonymous

      exactly this. I have written way more than 100 little programs and more than 50 really are used every day. From automation for simple tasks to things that convert huge junks of text from multiple sources to bring me the current news of current year.
      I also always reinvent the wheel so that shit like xz can not happen to me.
      Yeah I wrote my own command and control server with keys too.

    • 2 months ago
      Anonymous

      >I came to the realization that the only solution is to write my own software and minimize the use of other peoples software as much as is feasible.
      Have you written your own toolchain? Your own basic utilities? How do you know that you've not been supply-chained by a mix of malice and thoughtlessness? (If you have, that's based but you won't have any spare time to do new exciting stuff. And I bet you've not completely written your own web browser...)

      • 2 months ago
        Anonymous

        >Have you written your own toolchain?
        I wrote my own basic Forth, yes, but I want to implement the ANS standard before I start actually using it but that can wait.
        >Your own basic utilities?
        Currently working on this.
        >you won't have any spare time to do new exciting stuff
        I have nothing else to work on and I just do this for fun when I feel like it, but you're right it is a time sink with little tangible benefit because everything out there already exists and I could just use whatever exists.
        >I bet you've not completely written your own web browser
        No I have not.

  5. 2 months ago
    Anonymous

    I don't want to do that because I don't like the direction that technology and the human race is going in. The introduction of the smartphone alone has ruined the internet as well as an entire generation of people who've had one in their hands since being born.

    • 2 months ago
      Anonymous

      correct answer

      https://i.imgur.com/HlwEHnY.gif

      The current software that we all use everyday is the most unnecessarily convoluted that it has ever been in history.

      Windows is bloated buggy shit, Mac is bloated buggy shit, Linux is bloated buggy shit.

      How do we change it? What is the solution?

      New languages like Rust? Microkernels? Virtualization? Containers? Capability based OS? Model checking? Formal verification?

      I dispute "bloated" because every layer is a necessary layer of abstraction, if not for the computer then for the humans who develop the software.
      autistics who cry about "bloat" and spend their time writing custom memory allocators in their programming socks have very little effect on the state of software development for actually productive humans

    • 2 months ago
      Anonymous

      Well, that's the old Ted Kaczynski vs George Hotz thing.
      Technology cannot be defeated so you might as well try to use it in your favor.
      Both are about as handsome and smart.
      Ted was an incel who still was seething at normal people, George is rich and has a hot girlfriend.

  6. 2 months ago
    Anonymous

    Someone needs to figure out how to secure Kernel against cheaters that use kernel based exploits to hack your computer, online gaming is not safe anymore.

    • 2 months ago
      Anonymous

      I agree. But the problem is that the same technology will then also be used for example to prevent you from doing online banking if you are not using an authorized OS, or even the ISP could refuse to provide you service.

      • 2 months ago
        Anonymous

        In that case they would be liable for a lawsuit in multiple countries, besides many banks work on two factor authorization with a detailed key code thing that's physical and login that's digital.

        No way for hackers to mis-use my funds in that case if they don't have my debit card with its own code as well.

        • 2 months ago
          Anonymous

          There are banks already doing it for rooted Android phones.
          >No way for hackers to mis-use my funds in that case if they don't have my debit card with its own code as well.
          The 2FA frequently is a phone app and you don't need a credit card number to make a bank transfer. I suppose that's why they don't want to let their app run on devices which they perceive as less secure, because if your phone in that case logs all video and touchscreen presses then the hackers have everything they need.

    • 2 months ago
      Anonymous

      We already know how to secure the kernel against cheaters but you won't like it.

      start with fixed simple hardware

      Simple well documented hardware is what we need.

      start with a new internet, separate from the current corporate cancer

      We already have Gopher and Gemini. Neither of which has any content and user submitted content is awkward on both. Literally all we need is the early web without all the modern cancerous features.

      • 2 months ago
        Anonymous

        >We already have Gopher and Gemini. Neither of which has any content and user submitted content is awkward on both. Literally all we need is the early web without all the modern cancerous features.
        moronic opinion. In that model communities are associated with websites and we all know how that turned out.

      • 2 months ago
        Anonymous

        >Literally all we need is the early web without all the modern cancerous features.
        There is nothing stopping you from, right now, hosting an HTML4 website with PHP 4 and linking to only other morons that do the same. I don't see how modern ludditism will solve anything.

        • 2 months ago
          Anonymous

          IQfy is a web 2.0 site and so was 8 chan and so is tiktok. If he means web 1.0 (static content) then he sure is a hypocrite being here.

        • 2 months ago
          Anonymous

          I am running my own website that you can browse on an early 2000s browser. Works fine.

          IQfy is a web 2.0 site and so was 8 chan and so is tiktok. If he means web 1.0 (static content) then he sure is a hypocrite being here.

          My problem with Gopher and Gemini is that it's mostly just static content from what I've seen. Gopher does have the capability to "POST" through the search request but it's clunky. I don't remember if Gemini had, but I think it's intended to be read only. I did however say early 2000s web browser which implies a read/write web.

  7. 2 months ago
    Anonymous

    optimize existing programs instead of making a new one with half the stuff and double the size and requirements.

  8. 2 months ago
    Anonymous

    start with fixed simple hardware

    • 2 months ago
      Anonymous

      i cant properly solder

  9. 2 months ago
    Anonymous

    start with a new internet, separate from the current corporate cancer

    • 2 months ago
      Anonymous

      Might as well develop an intranet only for Temple-OS users if you are going to be that schizophrenic about it.

    • 2 months ago
      Anonymous

      I would like to propose a motion to the floor. We need a break away network that deters the normies, but doesn't attract the dredges of pedos and federal agents that .onion does. Something that can be fun and entertaining, but also decentralized. Imagine the distributed computing of software like BitTorrent, merged with the peer-to-peer discovery protocols of crypto.

      • 2 months ago
        Anonymous

        Web3 is already a thing.

      • 2 months ago
        Anonymous

        Web3 is already a thing.

        it already exists but the vast majority of people are too moronic to get on it.

        • 2 months ago
          Anonymous

          Hence:
          >We need a break away network that deters the normies
          Not having the vast majority of people on it is a good thing.

          • 2 months ago
            Anonymous

            yet here we are. because we'd rather have a site where we can have a semi fluent conversation with dozens of people, than a site where we can have a 1 post per day conversation with 5 people. stop larping. acting like communities need to be gatekept is the dumbest thing ever. I'd kill to be able to have IQfy levels of activity on a platform of my choice.

          • 2 months ago
            Anonymous

            I don't care what you want, newbie.

      • 2 months ago
        Anonymous

        This is something that has been on my mind quite a bit. I think the best way to go about this would be to build weirdness/memification in the platform itself. Along with tons of easter eggs etc. Make it so silly that no one even bothers astroturfing on it and add so much hidden knowledge that you can see a newbie from miles away.

        • 2 months ago
          Anonymous

          If it would take off, the glowies will have an entire department devoted to learning all the hidden knowledge to fit in.

          • 2 months ago
            Anonymous

            Yeah, I think in order to be resistant to normies and spooks there must be a certain board culture that keeps getting reinforced organically. I don't think things like these can really be solved through technology.

          • 2 months ago
            Anonymous

            This is something that has been on my mind quite a bit. I think the best way to go about this would be to build weirdness/memification in the platform itself. Along with tons of easter eggs etc. Make it so silly that no one even bothers astroturfing on it and add so much hidden knowledge that you can see a newbie from miles away.

            >create a distributed and decentralized network that relies on modular assemblies loaded at runtime
            >there is a neural network model virtualized and operating in the distributed computing system
            >it has access to all network activity and all the data and is basically a quasi-power user
            >is has the ability to write and compile code and push real time updates to assemblies on all active connected users
            >it has one single function: frick with the newbies and glowies

            It would be pure genius if such a thing wouldn't take over a year of dedication and work.

      • 2 months ago
        Anonymous

        >We need a break away network
        Amendment to motion: (You) build it, Chuckles.

        • 2 months ago
          Anonymous

          He's a /misc/tard newbie trying too hard to fit in, don't mind him.

  10. 2 months ago
    Anonymous

    It's too much Satanism for my taste, lads.

  11. 2 months ago
    Anonymous

    Everything will still be running on Unix-like operating systems and C. Don't fix it if it isn't broken.

    • 2 months ago
      Anonymous

      >Everything will still be running on Unix-like operating systems and C.
      That's the reason software sucks. C was bad 50 years ago. C is the cause of software stagflation.

      >Don't fix it if it isn't broken.
      Except it is broken. The amount of lines of code is not normal. Buffer overflows are not normal.

Your email address will not be published. Required fields are marked *