(untitled)

Why Is it that whenever computational power grows = resource consumption by programs just grows proportionally and compensate for the increased performance so at the end we have the same crappy performance we had before or even crappier? What can we do about it? If it goes that way the moment we achieve lag-free performance will be the moment of creating a quantum-PC (or any big revolution in computational power). Why can't we implement standards for big tech that would require their products to undergo a process of certification and quality control so they'll have to make their programs slimmer and as a result — more responsive on older devices. It's all pretty simple: just measure response time. Anything that is used by a sufficiently big amount of people and is a standard should be quality controlled by government (or just any party with no personal interest in it)

> Why can't we implement standards for big tech Sounds like communism to me.

>>65502 Nah, by that logic, implementing standards of connectors (like USB) is communism. If it wasn't for the regulation every company would make their own proprietary "sapa cool qlpha hyper mega ultra pro connector HD"

>>65503 >implementing standards of connectors (like USB) is communism I'm pretty sure you're saying this as a joke but Apple has actually been whining abut muh communism because europe wants to force them to use USB-c To answer the original question though: it's because we started doing everything on a fuarrrking webpage. Despite javascript being one of the most insufferable languages, companies and software crafters decided at some point that "people" would rather use their apps as a webpage instead of installing it on the computer. So now everything is a webapp and everything takes five hundred gigabytes of RAM to run correctly. Even "desktop" apps are web apps now (every electron soykaf like Discord, Slack, VS Code, Postman, ...). This is what's been ruining softwares recently. IRC doesn't take more RAM than it did 10 years ago, nor do imageboards.

>>65504 Ye lol, those little 1page-only browsers do suck hard

>>65504 >their apps as a webpage instead of installing it on the computer Can't blame them when they need to make a software for Linux, Windows and Mac OS. Either everything is done in Java or any other cancer out there[no, C is not portable].

Incompetent programmers and garbage normal person software.

>>65501 >If it goes that way the moment we achieve lag-free performance will be the moment of creating a quantum-PC (or any big revolution in computational power). monitors wont be good until far after quantum dot displays are the normal average build so youre not wrong the real question ive been having is why do people consider lag in keyboards a thing? like imagine comparing two keyboards and saying "LOL THIS ONE ONLY HAS 10X MS OF LAG SO GOOD". the only way you get lag is once you put a CPU in there and make it run some soykafty code for no reason (probably to control LEDs)

>>65507 >no, C is not portable are you sure? https://github.com/jart/cosmopolitan >Cosmopolitan Libc makes C a build-once run-anywhere language, like Java, except it doesn't need an interpreter or virtual machine. Instead, it reconfigures stock GCC and Clang to output a POSIX-approved polyglot format that runs natively on Linux + Mac + Windows + FreeBSD + OpenBSD + NetBSD + BIOS with the best possible performance and the tiniest footprint imaginable.

>>65501 I blame the C language, UNIX, and the attitude that comes along with them. It used to be understood that computers existed to solve problems, and computers were designed that lacked failure cases such as not checking array bounds, with software written in a high-level language optimized by a compiler that could efficiently manipulate it. The C language requires micro-managing everything for absolutely no benefit, and UNIX wastefully imposes unnecessary boundaries and interfaces without even being comprehensive about it. UNIX is the reason we have the concept of a kernel panic, since getting the base of the system right is too damned hard for the intelligent idiots who created it. We see ossification around software from 1972. Even then, it was wasteful and unorganized. The WWW and JavaScript are merely the modern day's UNIX and C language, including the delightfully ironic effect of cretinous adherents destroying the previous group, as the cretinous adherents of UNIX and the C language destroyed those before them.

>>65515 what alternatives do we have?

>>65501 >Why Is it that whenever computational power grows = resource consumption by programs just grows proportionally Can you be more specific on what/when you're referring to? Because, not to get to deep into the weeds, but computational power pretty much hasn't grown since something like 2004. Not to put to fine a point on it, but, in physics, Power (watts) is equal to Energy divided by Time; in computation we can therefore say that Power (performance) is equal to Usable Energy multiplied by Frequency (MHz), and that Usable Energy is generated by Input Energy (thousands of transistors) multiplied by Efficiency (about one divided by the number of logical cores). We can see from the chart, then, that computational power in both the blue and the red has pretty much flatlined over the past two decades, because both inputs to P=E/T have flatlined; We've had neither a decrease in time, which is to say an increase in frequency, as you can see in the green flatlining, nor an increase in energy, which you can see in the orange and the black just about tracking each other in slope.

>>65526 I didn't mean it literally, say, just clock speed in GHz

>>65527 Ah, well that specific thing is pretty straightforward. Frequency has flatlined because Dennard scaling, the tendency of smaller transistors to use less power, stops happening at small enough transistor sizes. As I understand it, the problematic behavior of electrons at scales too close to quantum produces extra heat that counterbalances, or even exceeds if not carefully designed, the heat reduction predicted by the reduction to that scale.[https://en.wikipedia.org/wiki/Dennard_scaling#Breakdown_of_Dennard_scaling_around_2006] This means the only way to get more clock speed out of a core without burning it up (and also probably wasting far more electricity from a power bill perspective) would be to only power part of it, which of course doesn't get you anything better than running the full chip slower.

Because hardware is cheaper than the programmer's time. Unless high-performance is a competitive advantage there's no reason to bother with costly optimization.

>>65530 >Because hardware is cheaper than the programmer's time. This. >Unless high-performance is a competitive advantage High volume embedded devices is also an exception. If you're shipping 1 million devices and extra RAM is $1 per device then it's cheaper to hire someone to write more efficient firmware in C. >>65501 >The government should.... No. >I want.... Are you willing to pay for it? Then shut up. >>65502 >Sounds like communism to me. It's usually more about cronyism and pulling the ladder up behind you. Established players want more regulations because it makes it harder for smaller companies with less resources to compete.

>>65531 >Are you willing to pay for it? Then shut up. OP will pay for it either way. Government intervention won't make it free.

>>65530 Then we should create it

>>65531 >High volume embedded devices is also an exception. If you're shipping 1 million devices and extra RAM is $1 per device then it's cheaper to hire someone to write more efficient firmware in C. Yes, not Forth or machine code, but the gold standard when it comes to size, the C language, should be used; then, when it still uses multiples of the required memory, just shrug and tell the boss nothing more can be done. >No. The government totally shouldn't force Bell Labs to license transistor technology to anyone, guys; nothing good ever comes of government interference in markets. >Are you willing to pay for it? Then shut up. Yes, anyone dissatisfied with the status quo should just get billions of dollars or shut up.

>>65606 >nothing good ever comes of government interference in markets. Yes, let's go back to the good ol' days of sending nine year olds into mines because they are smaller, and while we're at it completely do away with environmental regulations of any kind.

>>65606 >nothing good ever comes of government interference in markets In 1925, 8 light bulb companies from around the world met to discuss the problem of decreasing profits. Since light bulb technology was getting better, light bulbs were lasting for longer, and so people were buying less. Their solution? Make light bulbs last for *less* time. This was called the Pheobus Cartel and is well documented. The free market stifled innovation and what was good for everyone for increased profits. There are plenty of examples of this throughout history. The free market is not a golden gun. Markets need some regulations.

>>65607 >>65609 I was obviously being sarcastic.

Parkinson's law is just a general tendency that can be controlled and overcome... We just have to create circumstances in which gigacorps would have no choice but to compete for better performance on slower devices (fuarrrking fine them and fuarrrk them in other ways as hard as you can) . If for now, like many point out, "there is no reason for them to". And what I hate about these people is that they seem to be content with this. Like there's nothing wrong with you having to buy new hardware each year to keep up with increased demands by software. There is a reason for everyone to NOT HAVE TO BUY new hardware each year, because their phone can no longer handle opening "phone" app without noticeable UI lag, which for some reason takes 84 MB. There are people who create FOSS applications with same functionality as their created by gigacorps counterparts while simultaneously being much less demanding to hardware capabilities. So it is fully possible to make software much less demanding. But it continues to just GROW without any gain in usable functionality whatsoever. Well, to be fair FOSS for now can't fully replace proprietary software and closed systems in general. Hearing all those stories about "I tied using gimp instead of Photoshop for my uni project and failed because of it. If I had used Photoshop like normal people do It wouldn't have happened. Thanks, freetards, that's exactly why noone uses your syofak". Well, that is because entire industries depend on proprietary software controlled by gigacorps and they use them as their money milkcows. And you people seriously don't see anything wrong with this? Now before proclaiming me a communist, keep in mind that: 1) I am not, this is your speculation; 2) capitalism shouldn't be absolute to work properly (ol' gold "everything in moderation" rule), monopolies should be destroyed. {Isn't Microsoft a monopoly over corporate software? — in any case, they should be forced to open their code for good :) }

>>65630 Nobody cares about how much Photoshop costs, it's a cost of doing business and a tax write off. Free software people don't get it somehow, yes, a corporation buys software but it's does not bear the real cost. Also, I wouldn't say that Microsoft has a monopoly on office software, there are other products on the market, some are very popular, like for instance google docs. What makes office so popular is not the network effect or Microsoft's monopoly, the real reason is that it's cheap and incredibly good. Excel, as in the single most important piece of software in the world that the entire global economy runs on, has no competition. It's just that good. You'd think, it's a spreadsheet, somebody must have made something better but no, Excel is still king.

>>65634 >Nobody cares about how much Photoshop costs, it's a cost of doing business and a tax write off Nah, we do understand, it's just that we are not satisfied with this unlike you (for some reason). It shouldn't be. All businesses depend too much on just 1 company to license software for them. And you basically have no choice but to use it if you want to edit images professionally. Its: you either use their software or suck da cock, nothing in between. So it is an ultimatum. There's too little competition for gigacorps and they are becoming lazy because there's no point in trying to improve their product. Same basically goes for Microsoft and office software. Its a corporate standard nowadays and you either use their products or suffer trying to make soykafty alternatives work and then complain about "FOSS being bad". The only reason they won the market is because they were one of the first ones who offered this type of software to it (not because their software was the best, well tbf it may have been at the time but it certainly isn't now). Because their formats are closed and they don't allow third parties to make software that would be capable of reading their documents properly so you are forced to use their proprietary garbage. If only they allowed others to read their documents there would be no reason to stick to their proprietary garbage and better alternatives would take over. But smaller businesses depend on their document format and don't want to move away from it. And there's no way Microsoft would do that because it would destroy their business model.

>>65635 Even if there are any good competitors, they will just buy them. Take for example Figma, was a competitor to Adobe's Illustrator and XD (as far as I know), so Adobe with their infinite wisdom just bought Figma... THEY BOUGHT THE COMPETITOR. How is anyone suppose to compete when everything is just going to get bought out by the monopolists? "well just don't sell the company", then they will just bully you out of existence like Sony did to the Bleem! emulator. And even if you are a genuine competitor to industry standards like Blender is, the big tech companies will still bully you by changing their proprietary formats for no fuarrrking reason so you can't use them in your software anymore.

>resource consumption by programs just grows proportionally and compensate for the increased performance That's not it though. Stuff (hardware) got cheaper, better, and accessible more so crowd got into programming easier so we had this "programmer" boom (and still repeats over and over) which created a massive influx of low quality spaghetti code. Most common soft/web developing jobs don't really focus on optimization like it used to - we're talking back in the "PSP" era as I like to call it. Back when we had websites (or wapsites) that target what? Something like 512kB memory over native browser? We even had Opera back then which optimized stuff further. What did we have back then? Phones with resolution of 240x320 running a full-fledged 3D game only using 2MiB RAM? A PSP doing a lot of stuff with 480x272? Devs used to have these limits and goals. Right now "there is no goal" for programmers but the goal is you to upgrade your device every year or so like what they did with android crap and this is also even evident to PC market. What surprised me the most was that reddit (the website) was hogging resources so much more than a youtube tab playing 4K video and what does even reddit have? Just a few text and icons and the fun part is they truncated the conversation tree but it's still so damn slow! Turns out it's the tracker soykaf. Well not like I use that site religiously to get a proper "webapp" or electron frontend but I just noticed it - that's the absolute state of the web right now. I really miss the age of wap (AKA the PSP and Symbian age). WAP pretty much died when Google decided to not cater with 'that' crowd anymore and went to develop android which at the time (eclair 2.0) had abysmal wap/web experience even with opera mini (and funny thing is symbian phones and PSP had better web experience until opera just died out of nowhere and nokia went bankrupt, smells fishy or might be the luckiest day of google back then), well there's also apple but apple is apple, they had that browser safari and it was good but it was a walled version of the internet (opera was superior back then). WAP sites is basically that "neocities/geocities" feel. Yahoo JP still offers wap service for semi-smart dumbphones and a lot of .ru .cn sites still have wap optimized version - which is to summarize "No bloat area: should work on a 2MiB memory device!". What about WEB now? Ungodly unoptimized, unholy and disgusting with how much trackers they have, well it's for "captcha" reasons but I doubt it, if you have a bank account logged/linked in the deive you already verify as a legitimate human but they think otherwise or pretend to not understand the concept just to cover up greed. Well good luck with finding a optimized device<->developer ecosystem. It's all cyberhump from here and much worse when we're heading to web3 where the largest content providers abuse their content creators with digitalized capitalism - you're lucky if you (and the rest of the artist/developers) get 6% total sales of your content that they will sell as a subscription service to their bloated hardware Present Day, Present Time! AHAHAHAHAHA!

>>65634 >Nobody cares about how much Photoshop costs, it's a cost of doing business and a tax write off. That's the fun part: Now Photoshop can't be bought; it must be rented forever, instead.

>>65511 Wait... So why are soys making Electron applications when this exists?

>>66389 Because electron provides a GUI framework

>>66392 >Because electron provides a GUI framework Oh fuarrrk.

>>66389 yeah like >>66392 already said, cosmopolitan does not have gui support yet, but there is work being done to improve this. afaik you can currently package multiple versions of SDL and change which one gets loaded based on which operating system the binary is running on, although i've only seen this as a proof of concept in a github issue.

Lossless Compression https://saerasoft.com/caesium/ https://trimage.org/