Future of Software and Software Development

Unix was designed in 1969 and so was C along side it Windows, based on CP/M released in 1974 Both are half a century hold at this point. Neither were designed with networking, security, multi-threading or any other modern features in mind. Back then you would have 1 computer for dozens of people. Today a single person can be spread across 3 devices easily. All the modern functionality have been patched in but the core languages are still unsafe which is responsible for most of the security problems. What will the future of software look like? Will we continue using old solutions because it is economically advantageous? Building a new system is expensive, but all it takes in 1 serious conflict where cyber warfare becomes accepted and zero days are dropped left and right. All the businesses running on modern systems will be wide open to exploitation and be risking all of their customer data whenever their services are running on the public internet, then we will see real demand for newer operating systems and software with security being part of the initial design. Programming and learning to program computers today is a nightmare. From dependency hell to poor or non-existent documentation(and lack of off line documentation) the actual skill of programming a computer is artificially becoming more difficult. Newer languages are easier to use but the development stacks that they use are so fragile a single maintainer deciding to delete some core repo no one pays attention to, and will shutdown half the services running on the internet. Also their performance suffers as a result. Software is all about manipulating and presenting information in certain ways. Almost all the basic ways this can be done is complete. Text, pictures, movies, music we all have a plethora of options to choose from. If you want to write a novel piece of software you probably have to be working on a PhD in the latest research. AI/ML is offering the greatest increase of functionality in actual computing in the last 20 years. It is next to impossible for the average programmer to provide any kind of useful new functionality to the market. And most of those features are not available to the public since the models and how they are trained are proprietary secrets. OpenAI is the exact opposite of Open. It seem to me that even the computer world is dying as fast as it was born. Outside of building niche custom software for smaller businesses (web development) or working for giant corporations there isn't going to be much opportunity. Stuff like cryptocurrencies is a one off thing and is primarily used for scams. We see projects like Google's Fuschia but I doubt it will become a large market player and primarily focused at mobile. What does lainchan think the future will look like? I think it will take something very very bad and very expensive for the industry to start moving forward again. How or when that happens is not something I can predict. There are great projects out there that were never finished properly and could be successful in the modern environment (A Xanadu network with trusted information providers that don't lie or mislead) and it would be great to see them revived but most people have moved on. There is a huge obsession with pop-culture inside the industry but it seems that pop-culture is just becoming pay2win games, NFT/crypto scams, and buggy programs.

>Unix was designed in 1969 and so was C along side it Multics is even older. By that point, APL, Lisp, Smalltalk, Forth, and many other languages already existed. >Neither were designed with networking, security, multi-threading or any other modern features in mind. Multics was. That was the entire point. >All the modern functionality have been patched in but the core languages are still unsafe which is responsible for most of the security problems. Multics used PL/I, which didn't have these issues. This has been known before UNIX and the C language existed. >What will the future of software look like? More garbage written in languages for nitwits by nitwits seems likely, ignoring all inconvenient historical precedent. >All the businesses running on modern systems will be wide open to exploitation and be risking all of their customer data whenever their services are running on the public internet, then we will see real demand for newer operating systems and software with security being part of the initial design. No. Just think about what this would do to the poor security and anti-virus businesses. Computers without these issues have already existed, but they lost on the market because the market doesn't care about quality. Every business becomes financial manipulation, because why provide a working product when the broken product sells? >Software is all about manipulating and presenting information in certain ways. Yes. >Almost all the basic ways this can be done is complete. No. >If you want to write a novel piece of software you probably have to be working on a PhD in the latest research. I've written novel software and I needed no such qualifications. All it took was a different perspective. >AI/ML is offering the greatest increase of functionality in actual computing in the last 20 years. It's an attempt to avoid needing to make a decent system, by instead trying to make a system that will predict the user's wants. >It is next to impossible for the average programmer to provide any kind of useful new functionality to the market. I don't care about the market. I've dedicated my life to programming because I figured I could breathe life into beautiful things through it. >We see projects like Google's Fuschia but I doubt it will become a large market player and primarily focused at mobile. That's nothing special. Look at what Alan Kay and Ted Nelson have been trying to do for decades. >What does lainchan think the future will look like? Hope is an evil, so I expect the future to be even worse than the present.

>>31886 >Neither were designed with networking, security, multi-threading or any other modern features in mind. Indeed, but it would be false to say that macOS or Linux (which is only a unix-like anyway) is the same OS as the Unix running on PDP-11 in 1970. Unix is a design philosophy for operating systems, the modern versions of Unix you'd actually use today are modern and are designed for all these features. >Programming and learning to program computers today is a nightmare Programming was also a nightmare previously because of relatively how little the resources were available to create pretty sophisticated programs. For sure many people miss the days of pre-"Bloat" and before "nitwit programmers", but the fact is computer programs have a LOT more non-functional requirements today that were impractical to implement previously. Okay you might have had a whole databasing system running on 512kb of memory, peak efficiency right? But what are programmers of old needing to leave out to achieve that? >Error handling >Security protocols >Concurrency >Accessibility >Cross-platform execution So yes, you would have a database running on a tiny amount of RAM, but with absolutely no capacity to run outside of a very specific operating environment, with educated computer scientists at the helm and it's incredibly easy to break and/or crack. This is unacceptable today with our internet connected computers and just regular garden variety people using them. >It seem to me that even the computer world is dying as fast as it was born. Not really, it's just the pioneering days of mathematicians and electrical engineers working at Bell Labs with blank checks, and their kids going mad with cheap 6502s are over. Computing just isn't in an experimental phase anymore, it's less about trying to discover what computers can do, now software development is an exercise in architecture and that is something that is always evolving to seek a more secure, more reliable, more maintainable system which is what people need at the end of the day. >What does lainchan think the future will look like? It's abstract. We use the incredible horsepower of modern computers to completely leave the realm of bits and bytes, software will be functional to provide an extra layer of security and reliability baked right into the language while making concurrency trivial (which is required since performance gains now will likely rely on many more cores) and this will be combined with ML assistance like GitHub Co-Pilot to provide programmers with bulletproof boilerplate for any kind of function you would care to do.

>>31886 >Unix was designed in 1969 and so was C along side it >Neither were designed with networking, security, multi-threading or any other modern features in mind Looks like you forgot the UNIX 2 candidates: Plan 9 and Inferno. Networking was their whole thing and multithreading was tackled successively by Alef, libthread and Limbo. They're generally much less janky in every way, since they actually took some time developping it, instead of being rushed by commercial interests. I think that the Lisp Machines still look much more futuristic to the trained eye (that can look beyond the sparkles) than everything else, even today. And Lisp itself is certified alien technology, so it fits right in.

>>31887 This lainon is basically right. There is way too much financial incentive for the garbage that exists already to endure and for people to pile more garbage on top of it in the hopes of modernizing code bases and design principles that are fundamentally flawed. Anything that tries to do something different and better will simply be crushed and suppressed if it tries to compete.

>Software is all about manipulating and presenting information in certain ways. Almost all the basic ways this can be done is complete. Text, pictures, movies, music we all have a plethora of options to choose from. If you want to write a novel piece of software you probably have to be working on a PhD in the latest research. fuarrrking this, doing programming today just feels like learning to put together the dependency hell and learnjng the ropes of some new version of an old language to come to a new iteration of the same solution. There are exceptions, of course, where creativity still plays a role.

>>31887 >I've written novel software and I needed no such qualifications. All it took was a different perspective. Like what? I understand you can just combine different features together to get new things but it doesn't fundamentally change how that information is being processed. I really don't see anything in programming you cannot do today with tools already out there.

I think the future of programming is going to be stuff like Rust and Mercury but better. We're going to see more and more previously runtime-only high-level features implemented statically at compile-time in increasingly efficient ways, until eventually we have languages with expressive-power nearing that of Lisp running as fast as C. We'll finally be liberated from the hardware and actually be able to express ideas in human-friendly ways. >>31887 >>31891 The status-quo of the market can only hold for so long. If there's a truly objectively better way to do things and it's a marked improvement from what's being done now, it will catch on sooner or later.

>>31896 >>28919 I created that thread.

>I really don't see anything in programming you cannot do today with tools already out there. well, fundamentally every turing-complete language is The Same, so yes as long as the hardware and IO are considered constant then every language can do anything that any other language can do. The question is whether what you want to do is so difficult that you even *want* to do it.

>>31913 Computers exist to perform computations, not the other way around. Real Programmers have cargo culted around what is fundamentally a tool for automation and deemed that the correct way to use it is the least efficient most pain in the ass way possible. It's like using a hammer in place of a nailgun because a hammer is a more "real" way of driving a nail. We should hide the low-level details as much as humanly possible. Knowing how a computer works from a physical standpoint is exactly as relevant to a computer scientist as knowing how a calculator works is to a mathematician, but since computers suck we're forced to deal with worthless tedium. Once computing science properly matures, we can get back to what really matters: math and logic. Until then we're forced to deal with computationally inefficient high-level and programmatically inefficient low-level languages.

>>31886 >Windows, based on CP/M released in 1974 Wasn't Windows basically redesigned from the ground up with NT in the late 1980s to early 1990s?

>>31956 >Windows NT and VMS: The Rest of the Story >Is NT really new technology? >the roots of NT's core architecture and implementation extend back to the mid-1970s. https://www.itprotoday.com/compute-engines/windows-nt-and-vms-rest-story tl;dr Windows is a clone of VMS with patches ported from OS/2. MS DOS was clone of CP/M.

>>31886 I think these issues are caused because software engineers put too much focus on programming language features. Even though C++ doesn't have reflection and that I did not want to write a header file for every class. So I decided to just create a struct to create a class, and change that generates the header and the implementation file. The only downside is that I had to wrap the c++ code as strings. Maybe I should have used lisp instead.

the computer world isn't merely dying, its being killed. its all a bait and switch. computers were made interesting to get people to start using them, and funds were provided to enable interesting things to be made for them. now that computers have become ingrained, inseparable, from society, there's no reason for them to be interesting or usable anymore. they're here to stay now so why bother making them decent or programmable by civilians. more than that the freedom of computers was the bait but never the intention, computers are a means to control and monitor people. the internet began as ARPANET - a military venture. darpa's LifeLog was "canceled" the same day Facebook was founded.

>>32167 Computers are tools. It's your fault for thinking of them as anything more.

>>32168 guns are just tools too, but when an enemy military starts stockpiling firearms you should acknowledge the implications of that rather than putting your head in the sand

>>32172 I don't get how that relates to my post.

Expensive? Sure. Bad? I doubt. You're correct that most of the computing world is on the wrong track today, but I think there's plenty of reason to be hopeful for the future anyway. Unix and Windows, and most things designed originally for "growth market" platforms, give a very bad impression of the whole of humanity's computing knowledge. Security, multithreading, machine networking and transputing, they were virtually unheard of to most computer users, sure, but up until some time in the 2000s most computer users owned the push mowers of the computer world, if that. Nothing's new under the sun in what is essentially applied math and electrical engineering. Someone much smarter than both of us with a million-dollar multi-hardware mainframe or wall-sized minicomputer at least tried the vast majority of non-gimmick computing concepts you could ever want, and left behind enough information for us mere mortals to bootstrap their accomplishments back together once the willpower or money comes together. Perhaps we need to start a nonprofit and make regular donations to s research team, but they'll have plenty to do. Languages and software, too. Lisp is great, but it's just the tip of the iceberg. There's scores of almost as good languages and concepts that just never happened to break through; they could be great foundations for a true catalyst today if only someone put the time into integrating the lessons learned since the cobol era. IEEE publications seem to have plenty of jumping-off points of abandoned white papers for this or that new way of doing things that never got their follow-through, from what I can gather around the paywall and poorly documented historical archives. Sure there's mountains of waste, too, but the juice is worth the squeeze. Nothing lasts forever, no matter how entrenched it seems. Linux swept away the garbage proprietary commercial Unices, Javascript swept away flash and applets, servers swept away terminals and thin boxes. Things get worse when better isn't easy, but the tide always turns eventually. And since some professor in the 70s has always already tried what you want to improve, it won't even be hard this time. Point is, we're light-years away from maxed out on the milage we could get just from what we're unaware we already know, so there's no need to panic over limited access to what we don't know, and thus don't understand how to use non-trivially. It just doesn't seem that way because the everything tech bubble has confused funding sources on where profits actually come from. Both facebook and google have become huge confidence-game players as much as they are tech players to avoid admitting this. The playing field will level as they keep ossifying and no longer look like safe money.

hey summer users, you have just a good of a chance as any of us at reviving xanadu if you've seen / heard of ted via herzog or lanier or lain then you know what's up try, it's worth a shot replace pdf with the docuverse replace sql with zig zag then do xanadu over gnunet times-a-wastin

>>32279 name it xanaGNU but wasn't the flaw with the original that it used soykafty blockchain microtransactions for GETs? would there be a good/recommended gnunet compatible replacement for that?

>>32279 What should it look like from user's point of view? Is it comfy to implement just the same as in https://youtu.be/En_2T7KH6RA ?

>>31886 I am no expert yet but I'm not optimistic. The vast majority of devs are openly hostile to functional programming let alone actually new stuff. I met one who didn't know or want to learn what a pipe or higher order function was.

>>32168 Computers are technology because they require an interconnected network for their production, what Ellul and Mumford called 'technics' or 'technique'. A spade is a tool. Do you know anyone who got addicted to a spade?

>>32445 And people also get obsessed with cars. Doesn't change the fact that, at the end of the day, they're fundamentally a means transportation. Maybe a computer isn't technically a "tool" in a strict semantic sense. But you get the point. It's a piece of technology that exists to make things easier via automation. You can use a computer for its own sake, becoming knowledgeable in all of its low-level functionality. But that's not what computers are made for.

>>32352 I don't think we need, or even want, to care. The vast majority of programmers have never been the driving force behind industry change. They're the receivers, not the creators. The history of computing seems to skew to lean teams of highly motivated and skilled devs making all the difference. The top 0.1% of all programmers is 20k people; we could be just about as selective as we are able in putting together the save technology team and still fill a stadium. Cultivating manpower isn't worth the time to think about it. Organizing the latent manpower we already potentially have is the whole of the question of how to improve things if you ask me.

>>31886 I think we well see rise of safe programming languages, like Rust (and perhaps/hopefully Ada) and functional programming (Clojure?, F#?, idk).

>>32451 >Ada Yes, but only in the right market sectors. NVIDIA adopting Ada/SPARK is big news. I think we'd all be better off if model checkers and provers were more prevalent in development of core libraries. There aren't that many people who've got those chops, though. >Rust Rust is Ruby as farce. The same cabal of charlatans that polluted and over-hyped that scene moved on to the next season's fashion. Rust has been hyped by paid evangelists since 2008; it's been out of beta since 2015; it's been mostly feature-complete since 2017; there is still almost nothing to show for it except a grep clone. Their acolytes have tried to force it into the Linux Kernel for at least a year, and I don't think anything has actually been merged. I think you should try to compartmentalize and discount information that comes from Reddit/HackerNews marketing campaigns and what actual people are doing on actual projects. Looking at job postings is a great way to know if something is being adopted or is useful. If you can be paid to know it, then it's probably used for work. Example: Java, a dead language, has about 80,000 job postings on indeed. C++, a scary, scary language, has about 21,000 job postings. Rust, the future of programming, has about 1,600 job postings. Haskell, big-think-lang, has about 70 If you're interested in the hobby-sphere of languages, here's a great visualization: https://madnight.github.io/githut/ I find it very interesting to compare pulls, stars and issues. Those roughly translate into 'how many people write it', 'how many people hype it', 'how many people are using it.' Here's one example: Rust is well above Scala in stars, but Scala is much higher in pulls. Conclusion: the ratio of hype to real work is much higher for Rust than Scala.

>>32456 >Rust is Ruby Stopped reading there. The two languages are nothing alike.

https://hyperworlds.org/ Highlighting the strengths and weaknesses of some of the software projects that will replace some of the World Wide Web, email, and chat.

>>32459 They weren't talking about technical similarities, nor in fact are they talking about any technical merits of either language. They're just trying to associate Rust to "people we don't like".

>>31892 True but most problems being solved are the same. Programming at SaaS companies feels the same, but that's because all SaaS companies are just a frontend and a database with some logic in between, so the code is interchangeable between languages.

>>31886 >Fuschia The future is awk. You are literally living a meme.

>>32459 >>32467 Ruby evangelists were overwhelmingly left-signaling non-programmers selling snake oil. The exact same people boosted Rust. See Steve Klabnik (note: a communist who ragequit Mozilla because he wasn’t paid enough). Both were kind of cool hobby projects that were hijacked by hacker-scene-groupies into pseudo-social crusades. You better start believing in leftist social-proof grifts in tech… you’re in one

>>32481 Okay so you're here to talk about worthless drama, not programming. Got it. >inb4 I used to be a "ruby guy" back when it was more of a thing, and I never had to deal with any of that stuff.

>>31886 >Unix was designed in 1969 and so was C along side it. Windows, based on CP/M released in 1974. Both are half a century hold at this point. Neither were designed with networking, security, multi-threading or any other modern features in mind. Back then you would have 1 computer for dozens of people. Today a single person can be spread across 3 devices easily. All the modern functionality have been patched in but the core languages are still unsafe which is responsible for most of the security problems. What will the future of software look like? Will we continue using old solutions because it is economically advantageous? Although I have no idea when it will happen (since it's been possible for quite some time, so it's prediction is a social rather than technical question), I'm pretty much certain these things will eventually improve, and probably within our lifetime if we find the new system early and help it along. Part of the reason the legacy systems have had such massive economic staying power is that, until very recently compared to the whole history of computing, the real money has been in repackaging, or even reducing, existing tech to make the most useful 20% of it cheaper or more accessible, rather than iterating on it to be more powerful, let alone inventing something to replace it. Even many supposed improvements were mostly just very careful PR management of this general strategy (Remember 32-bit processors with a 16-bit buss so your DOS programs can pretend to be as capable as a workstation? Remember Basic hooked into graphical consumer applicatons so they can pretend to be as capable as a database and programming enviornment?). The other lainon's mention of multics and unix is a great microcosm of this. Technologically, Unix was just Multics for some devs who didn't have time to wait for it, and then commercially it was just Multics for customers who wouldn't pay for it (admittedly usually because they couldn't afford it, but that's how econ goes). Both turned out to be the vast majority of people, so the question of technology turned out to be a non-factor. If there wasn't the potential for this trend to turn around, things would be very bleak indeed, but by my reckoning we're very lucky that the downward slide stopped somewhere between ten and thirty years ago, depending on where you want to draw the line. For all we cognoscenti might complain about it turning computers definitively into consumption toys for most normal people, the iPhone was the social phenomenon that seems to have finally set in stone in the general consumer's minds that it's worth it to spend more money on the thing people say is better. Even as hardware costs for the same tech have continued to drop, today's trend is that the most successful next generation is more expensive and powerful than before, not less. Netbooks (read: chromebooks) seem to be the equivalent case for actual computers; even more pared-down platforms exist for niche uses, but Google found the market's local minimum with those just-barely-not-toys and the average has been going up since. I think these hardware purchasing trends suggest the tech-buying public is, for the first time, latently primed to actually fork over the cash necessary for better quality software instead of just more garbage, meaning there's finally an alternative source of value to industry to counterbalance the cost of not just reusing the "good enough" tradition analogous to manual tools and hand-fitting. Again, though, no idea when this penny will finally drop. Social adoption of new technologies has always taken time. Arguably we could've had widespread adoption of, say, interchangeable parts or a hyperdocument protocol long before industry finally jumped on board, but it can be difficult to think through all the benefits of an initially costly major change over the drawbacks before it's finally got some publically-known traction. That's just how people are, unfortunately.

>>32466 >software projects that will replace some of the World Wide Web i really, really hope so. xanadu and related projects seemed really promising when they appeared

>>32493 sad isn't it? i really love lainchan and i owe a lot to people here but all the members intelligence and nuances seem to go away when they figured out that they can twist the conversation to soykaf on gay/trans/left whatever. i don't come here expecting people to say things i agree with, but at least expect them to talk in relation to the topic. its like a competition where you when when you are the first to say "tran*y" "leftist".

>>34972 i can see OpenXanadu making a comeback among certain posters - i think schizos will really love it. And I don't think I'm clear on who "we" are meant to be. Consumer software will remain completely unchanged or go through some barely noticeable iterations. The public doesn't want to learn vastly different material - it won't be easy, they won't go. Everyone else who cares about things will go where their subcultures go and rely mostly on community projects, occasionally those with back backers. I can also imagine new software coming out of China in the future?

>scroll to random position in page >see fuarrrkjob bewmer being retarded as usual >>31888 >security protocols what the fuarrrk do you even think you mean? what do we have now? DAC? firewall snakeoil? nothing is harder to hack now than it was when your LARPing diaper ass called it "cracking". >accessibility almost nobody uses that, by definition >a more secure, more reliable, more maintainable system which is what people need at the end of the day. computers are not more "secure" or "reliable" now, they are just more bloated and still as buggy. they also werent good to begin with for any retard who will agree with me. >copilot will be the future because its the only way to build a webshite without spending 1000s of hours understanding all the bad legacy decisions of w3c scum

>>32481 >Ruby evangelists were overwhelmingly left-signaling non-programmers selling snake oil. You know Ruby is a Japanese project. It's basically a lisp with sane syntax. Nips love lisp that's why it appears in the lain anime. Anyway it's a bit harsh to condemn the whole tech community just because some obnoxious english speaking losers grabbed hold for a while (and later jumped to rust anyway like someone else said).

>>32456 >there is still almost nothing to show for it except a grep clone it's in the linux kernel it powers Discord, Dropbox, Let's Encrypt, parts of AWS and a lot more. Microsoft is betting on it hard. >Java, a dead language, has about 80,000 job postings on indeed. no one (besides college freshmen) believes that Java is dead >C++, a scary, scary language, has about 21,000 job postings. that's a very low number for such an old programming language that used to be incredibly ubiquitous. A huge chunk of it got replaced, not by Rust but by higher level languages like Java. >Rust, the future of programming, has about 1,600 that's a decent number >Haskell, big-think-lang, has about 70 no one ever said Haskell had jobs. It's the quintessential memelang Anyway, your post does have some truth to it. There's way more buzz for Rust than there should be. It's meant to be a language for modern projects that in the past would've been built with C++ or C. But just like Ruby (it's still huge in Japan for example) it's more than a fad, just extremely overhyped and its market share will never be as big as the amount of propaganda articles being shat out every day

>>35211 >A huge chunk of [c++ jobs] got replaced, ... by higher level languages like Java. Did they though? Were c and c++ ever that popular to begin with? "higher level" langs like cobol shell scripts tcl basic probably had more job openings than c++ in times before newer abstract programming languages like java python etc

>>35212 >Did they though? yeah, most desktop applications were made in C++ before Java became popular

>>35212 TCL was never more popular than C++, just suggesting that defies belief. Not even in the same magnitude

>>35214 i meant those types of languages in general, not in specific. perl would be a more popular example of the past. >>35213 sure DOS & desktop apps were written in c c++ pascal before java but i don't think most devs of the past were making desktop apps more were probably working in langs like cobol for big business, which were like javascript of the past

>>35211 >it's in the linux kernel With a little toy hello world driver. There is no production code, it turns out rust is actually hopeless for re-writing low level C code because it all ends up being a big unsafe block anyway. The reason it's "in the kernel" is politics and money not merit. >Microsoft is betting on it hard. By rewriting some C++ structs into rust. Go microsoft. >no one ever said Haskell had jobs No one (besides college freshmen) believes that haskell is a memelang. It's used plenty in banking where correctness and auditability are key requirements. >>35212 >Were c and c++ ever that popular to begin with? Dude there's millions of lines of C running in you're laptop's wifi firmware alone.

>>35217 > With a little toy hello world driver it literally just got approved a few months ago, and it achieved what C++ would never do after decades trying - convince Linus that something other than C was viable in the kernel. >No one (besides college freshmen) believes that haskell is a memelang. It's used plenty in banking where correctness and auditability are key requirements. if it has less than 80 job listings on Indeed, it's fair to call it a memelang. >By rewriting some C++ structs into rust. Go microsoft. downplay it all you want, it's a big deal. Chrome has also been experimenting with a switch to Rust, but it's a huge ordeal as with any project of its size. Every big tech company is betting on Rust, some more than others but it's way past the "just a fad" phase

>>31888 >software will be functional to provide an extra layer of security and reliability baked right into the language while making concurrency trivial (which is required since performance gains now will likely rely on many more cores) Sounds like the whole idea and drive behind projects like Elixir which dusts off and make the advantages of Erlang and Beam accessible to basic coder monkeys such as I. Same could be said for Rust even though the borrowing mechanism isn't particuralry new and hasn't been a design choice from the first iterations. So currently these are interesting even though I greatly despise anyone calling themselves or others a cRUSTacian. *shudder* As others have pointed out it's not about really having something in which to write a program in. This possibility has been around for over half a century now. But indeed to do that but in an increasingly more economical, accessible, efficient, maintainable, reliable, secure and idiot proof manner, including possibly misguided attempts like this entire nodejs nightmare that breaks at every release cycle of some obscure npm package. And anything replacing the coming hot thing will likely move into the same direction unless something truly paradigm shifting happens like skynet taking over or the spirit of Terry reemerging and making everyone suddenly interested in tinkering within offline environments with an improved c syntax which compiles just in time as well as ahead of time.

>>35215 >but i don't think most devs of the past were making desktop apps not really relevant to what I'm talking about though. I was referring to the C++ job market, not the entire job market, and desktop apps were a significant part of it for a while until the Java era

>>35221 >Every big tech company is betting on Rust They're not betting on rust they're have rust forced on them from the top down. Just like diversity and global warming and all the other current thing nonsense every corporation pretends to care about to get their ESG money. Don't confuse politics with merit. >it literally just got approved a few months ago Doesn't that prove my point? Someone wants rust to be "in the kernel" and highly visible even when they don't have any real world use for it. **Politics.** >it achieved what C++ would never do after decades trying I think you're soykafing on your own point again here. Everything in Linus's famous C++ rant equally applies to rust so clearly the decision is not technical. **Politics.**

>>35226 > They're not betting on rust they're have rust forced on them from the top down nah, it's from the bottom up. Rust developers are very satisfied with it, it's in every survey. >Just like diversity and global warming and all the other current thing nonsense every corporation pretends to care about to get their ESG money. Don't confuse politics with merit sounds like you really, really want Rust to be a political thing only, but it's not. I get it, Rust zealots and cargo cultists are annoying, but that doesn't eliminate the language's legitimacy. It's a modern systems programming language with pros and cons that addresses some memory safety problems. Nothing more nothing less. > Doesn't that prove my point? no it doesn't, because it's clearly a sign that Rust is being taken seriously by key players. C++ devs did decades of evangelism and Linus never budged. It's a bit silly to imply that Rust only got into the kernel due to politics, as if it was ever that easy to convince him of anything. >"in the kernel" and highly visible even when they don't have any real world use for it. https://www.phoronix.com/news/sudo-su-rewrite-rust https://www.phoronix.com/news/Linux-6.4-More-Rust > Everything in Linus's famous C++ rant equally applies to rust lol no it doesn't. His main complaints are OOP, the STL and other inefficient abstractions, exceptions, and bad programmers that are associated with it. None of this applies to Rust, except maybe the last part which is debatable. > clearly the decision is not technical. **Politics.** when was the last time you read the lkml? You're making a lot of unfounded accusations and dumb assumptions here. I'd expect this on /g/ not here.

>>35227 >nah, it's from the bottom up You can't have it both ways. If all tech corps are suddenly adopting it at the same time then it can't be a bottom up phenomena. The way bottom up tech trends emerge is through startups and open source projects making it big and getting adopted by the corpos. This is completely the other way around, the corpos are pushing rust down onto everyone else. >no it doesn't, because it's clearly a sign that Rust is being taken seriously by key players Dude it's not like someone built a great module or driver in rust and then linus was like yes lets include this in the kernel. They've got nothing, it's 100% forced. >as if it was ever that easy to convince him of anything Err, remember in 2018 when Linus disappeared and then randomly came back with a CoC? He's been compromised for a while. >His main complaints are OOP, the STL and other inefficient abstractions, exceptions, and bad programmers that are associated with it https://lwn.net/Articles/249460/ His complaints are that it is complex and unportable. Hello rust the language with no spec and only works with llvm. >It's a bit silly to imply that Rust only got into the kernel due to politics Even more silly to imply that rust got into the kernel because Linus had a personality transplant and genuinely likes it even though no-one is using it for anything.

>>35227 A huge selling point of Linux is security, imagine if Linus refused to support rust. Linux would be endlessly attacked for being unsafe insecure and not taking security seriously. There's no way Linus could refuse rust, the rust marketing & propaganda is too strong.

>>35228 > If all tech corps are suddenly adopting it at the same time then it can't be a bottom up phenomena yeah it can. It was kept from irrelevancy for years by Mozilla, enthusiasts and academics, and all the buzz got to big tech eventually. Some of that buzz came from dumb web devs with FOMO treating it as a JS alternative, but a lot of it also came from really smart people who know what they're talking about. > The way bottom up tech trends emerge is through startups and open source projects making it big you mean like all the crypto startups from 2018? OneSignal, Dropbox? open source projects like Servo? Warp? or the Rust standard library? or the Rust programming language itself? > This is completely the other way around, the corpos are pushing rust down onto everyone else. the community and surveys say otherwise. The Rust community was big way before it started getting adoption in ~2020. There were systems programming courses using Rust at Stanford before any corp was involved, even before the Foundation was a thing. > Dude it's not like someone built a great module or driver in rust and then linus was like yes lets include this in the kernel You wanted people to do work for free that was almost guaranteed to be rejected? You're suggesting that people should've put the cart before the horse. Your arguments are barely coherent at this point. Rust is being pushed by evangelists pushing books like Ruby was, but no actually it's the leadership at corpos. Rust has done nothing of value-oh it did? doesn't count, politics. Oh it's in the kernel? It's a conspiracy. >Even more silly to imply that rust got into the kernel because Linus had a personality transplant and genuinely likes it even though no-one is using it for anything. Discord (their entire backend), Dropbox (a bunch of their services and infra), Apple, Microsoft, Meta, Amazon, SpaceX, Cloudflare, Figma, npm, Disney+, Tesla, Coursera. People are using it. You are delusional. >>35229 more conspiracy theories. Great. This place is just /g/ with a lain skin now. I don't know why I bother. >There's no way Linus could refuse rust, the rust marketing & propaganda is too strong. C++ marketing and propaganda was just as strong when he rejected it.

>>35221 >it achieved what C++ would never do after decades trying - convince Linus that something other than C was viable in the kernel. Linus today isn't the same as before. I suspect the LF is now the shadow leader. Btw I'm not against Rust (except for its poz), it's massively better than C++ (except for the package-manager from the beginning meaning you have npm tier dependency trees). But I know it'll never replace CL in my heart, it's too ugly and "organically growing" for that (still accreting stuff, like C++ did for decades).

>>35231 >Rust is being pushed by evangelists pushing books like Ruby was Ruby had Rails. Passionate engineers create startups, use rails to write their saas platforms, get bought by bigger companies, now the bigger companies are using ruby. >You mean like all the crypto startups from 2018? Were any of them bought? >open source projects like Servo? Warp? or the Rust standard library? or the Rust programming language itself? Yes the biggest open source rust project is the rust language itself. Well done. More to the point what happened to all those "rewrite in rust" projects. They're all either abandoned or switched to go or zig. >You wanted people to do work for free that was almost guaranteed to be rejected? I thought you said you were a kernel dev. Not everything lives in mainline. Tons of open source and commercial kernel modules live out of tree. Any one of them could have adopted rust and used it for real world applications. Then they'd have a strong case to say look at this cool soykaf we did with rust, maybe we should merge it into mainline. What happened was the exact opposite. They forced rust straight into mainline and now they're scrambling to find a use for it. >Discord (their entire backend), Dropbox (a bunch of their services and infra), Apple, Microsoft, Meta, Amazon, SpaceX, Cloudflare, Figma, npm, Disney+, Tesla, Coursera. People are using it. You are delusional. It's bad faith to take quotes out of context like this. I said nothing in the kernel uses rust so there was no justification to add rust support, or to push it straight to mainline. >Your arguments are barely coherent at this point. Just because you're upset doesn't mean I'm wrong.

>>35231 Come on, it's not a "conspiracy theory", regardless of your opinion on rust, it's obvious it would be really bad publicity for Linux to refuse it outright like c++. I don't think c++ claimed to be more secure than c, and certainly they didn't on the level that rust did. c++ has the dubious claim of being more productive & efficient & easier than c. Dubious as despite a few really useful features like templates it's full of garbage that if you try to use will make everything worse, the development & result. At least the booby traps in c you could work around and still do something useful. In c++ the whole feature is the booby trap, trying to use it at all is just pain for everyone. C++ people could never claim that Linux is insecure because it rejects c++ and be taken seriously by anyone.