Should the web be purely static?

Is there an actual use-case for webapps? when you use a webapp you literally re-download the same program over and over again. you could get the same functionality if you just downloaded it and installed it once. I'm not talking about chromium programs since they're not really in the category of a webapp. so, should anything on the web (not the internet) even be dynamic in the first place? does it make sense?

>>88314 > when you use a webapp you literally re-download the same program over and over again. No, it's the opposite. With webapp you download it once, then your browser just calls API endpoints with data, receives JSON and renders it locally using already downloaded webapp. If you turn this into server-rendered site instead, then you redownload html layout every time in fully rendered form instead of downloading just data and turning into html locally, and this is wasteful af.

>>88317 You misunderstood him. You need to redownload the webapp everytime you need it, as opposed to a conventional app that you download and install once. This isn't such a big issue most of the time. Network speeds nowadays are high enough that downloading a program is often negligible. That's except for things like games or bloated corporate ware (although those could be designed such that you only download the parts you need), but you'll notice that they aren't distributed as web apps. This also excludes users with bad/unreliable internet (third world, isolated countryside, dictatorships). A case where this doesn't matter is when updates are more frequent than usage. (either because of frequent updates or irregular usage)

>>88318 JS bundle is obviously cached once you download it.

>>88319 No it isn't. My browser doesn't remember anything I don't tell it to when I close it.

>>88323 You're special. Don't use webapps then. Might as well disable javascript altogether.

webapps aren't a bad idea, they're a bad implementation. ideally there's client side ui description language implemented in the browser + some simple scripting to put logic behind the interface. blah blah blah blame google

https://www.gnu.org/philosophy/javascript-trap.html https://www.devever.net/~hl/xhtml2

>>88323 >choose not to cache program >complains when program is re-downloaded this is a problem of your own creation. this is an especially bad take since OP's image refers to PWAs which you can manually save if you, for some reason, are anti-caching.

>>88317 >With webapp you download it once, then your browser just calls API endpoints with data >server-rendered site redownloads html layout every time in fully rendered form >this is wasteful af Dumb-ass argument. It's not wasteful at all. Overhead of a static HTML layout is like 1 KB of compressed HTML text with each page load. For example loading >>>/Δ/ is 600 KB of which the main HTML is 27 KB (100k characters of text) most of which is text content and a small fraction is HTML markup, it basically compresses down to the same size as a JSON. The rest is all pictures. We're on a website where the efficient WebP and AVIF image formats aren't allowed, where people post images as PNG at 10x the necessary file size. Loading one of those wastes as much bandwidth as the overhead of thousands of static HTML page loads. Nobody gives a fuarrrk about efficiency on the Internet. It's all bloated media, bloated JSON, wasteful JSON fetches, spyware running in the background and doing network requests, spaghetti frontends using the soykaftiest algorithms developed by retards who shouldn't be allowed to write software. If you mean compute efficiency, observe your browser's CPU use over time on JavaScript sites versus server-rendered ones with JavaScript turned off. Visit your favorites. Check how many blogs and news sites won't load without JavaScript for no reason. Measure energy consumed if you can. On benevolent websites CPU use of the JS app and the server-rendered page would equally be zero. On a real-life website (e.g. news sites displaying 0.02 MB of text running a 2 MB obfuscated binary in the background to exploit your personal data, video sites doing complicated data processing instead of an HTML video tag) there will be significant energy wasted, almost always evil tracking code too which you never consented to. Your loss, their gain. There are very few sites with benevolent JavaScript left, more and more insert malware because people have no awareness, they don't care about waste, human rights, owning their data.

>>88329 RSS only pure data and style is left to reader. In essence you only need to define what data you get and thats enough to let the receiver decide how to style it.

I'll go you one better and say that the internet would be a lot better off without the world wide web at all. The whole concept is worse than spaghetti-coding, and it inevitably turned into one big scamming, time-wasting soykafhole.

Web-Apps at times for my phone running graphene is okay if it's not something I need to use all the time and saves the security risk of installing a new application where I like to have minimal bloat actually installed.

>>88632 I do the same. For any 3rd party app that has a web alternative without any significant downsides, I will always use the web app. I fail to see the benefit of installing apps like spotify, whatsapp, or even your preferred email client natively when they can be confined in a browser with telemetry, premium nags, and other annoyances/permissions blocked.

yes. lets all disable javascript NOWWWWWWWWWW

PWAs are extremely portable and that is in every way imaginable. OP's drawbacks aren't real. The bundle can be easily cached so they don't download over and over again. The fact they don't need to be installed is a feature not a bug. PWAs have some native integrations to go onto home pages or taskbar, etc. Stuff like photoshop is probably better as a native app, but more basic tools: like weight trackers, or calculators, chat, etc. are better off as a pwa. Even games. I actually think they're underused.

>>88314 >Is there an actual use-case for webapps? there is none >when you use a webapp you literally re-download the same program over and over again that is correct >>88317 >thinks optimized edge case contradicts the main point you are autistic

> Is there an actual use-case for webapps? I used to run the PWA version of the banking apps that I needed on my Degoogled custom rom. Kind of sad that having play services is a must unless you are willing to put in the effort.

web apps are the death of efficient programming >lets program this application using redux on top of react on top of html, css, and js that must be able to run on all three major browsers that run on the os on top of c on top of intel or amd theres so much bloat in every web app stack that it is literally impossible for them to run fast. how many layers of abstraction do we implement before we lose sight of the implementation? how many dependencies do we inject before we can no longer trust the code that we are running? how many hacks do we have to use for performance before our ultrahighlevelcode becomes unreadable? >>88317 >use caching this is cope for bad algo. switch algo before optimizing, thats basic soykaf >>88815 >confined in a browser a browser is almost entirely not contained. blocking what telemetry and permissions the browsers allows you to only further narrows down ur fingerprint for server side tracking. what browsers are attempting and failing to do with the permissions should really be done on the os side, in a fashion similar to how apple does on ios (altho its not great at it either, lol).

I don't know... I've kind of come back around to them honestly. If the web tech making PWA's possible didn't exist and the focus was on rich applications like it was 20 years ago, then developers would only cater to a specific device or operating system as a "standard" that would essentially bar any competition in this space from ever developing. It's largely why Microsoft Windows was dominate during the PC era, because developers developers developers! It's actually still somewhat the case to this day with plenty of software development companies restricting their products to one particular platform while knee capping their potential market options by not adopting an agnostic approach; but many of these remaining rich software development companies in my experience only produce leechware to prey on back-office middle management morons using the wrong tools for the wrong job, such as Adobe Acrobat or just about any Excel plugin for large finance calculations and business intelligence. If you worked corporate IT, you'd be surprised the culture of learned ignorance in the office space. For instance, if Acrobat (the PDF editor) were to suddenly be pulled away from Adobe, I guarantee you they'd implode over night while the rest of their services wouldn't be able to compensate for the silent behemoth that Acrobat truly is in the corporate world, and yet despite that, the PDF format could cease to exist and the office world would learn to get by just fine with postscript and word doc files... I don't have much a use case for them outside of work, but services like arcGIS, google maps, filing taxes, and some of the Nextcloud plugins as PWA's isn't so bad considering the alternative. As PC users, we'd all still be suffering the worst of Microsoft full blast as a standard operating system base otherwise. Some of these PWA's you can "install" to your desktop, which I actually prefer doing instead of using my web browser as some makeshift task manager. Tabs should be for actual web pages while web apps can live in their own separate windows in my opinion.

>Some of these PWA's you can "install" to your desktop How do you do that? The way I've done it for Discord (their electron app is like kicking dead whales down the beach way more than the web version), is to create a firefox profile, customize it to hide tabs / toolbar and prevent intempestive third-party requests, then make a custom .desktop launcher. I wish there was a more streamlined way to do that. At least for creating a new profile with sane defaults.

>>90843 You're confusing a bunch of things together. How PWA (progressive web app) works: >you load a webpage >webpage tells your browser "hey, here's a js file that can manage this page's offline cache" >you chose to install/pin this page >browser accepts the js file, runs it, js can fetch and save other files it wants, like html and css >you go offline >browser notices >you try loading the page again >js kicks in and says "I've got this", serves you the webpage as if you were online So there is a clear divide between standard webshit, and PWA. You don't need to be online in order for PWA to work. On Chrome there's even loads of APIs implemented that may mean PWAs can be more useful than a note taking app or a calculator. WebBluetooth and WebRTC would mean two instances of PWAs on different devices can communicate. Apple wants to kill this real bad. To install a PWA you need brainpower that's found maybe in 1% of their userbase, it's very unobvious. Why? Because a lot of their appstore could be replaced by PWAs, meaning lost income. The greatest thing about PWAs is that you can self publish, just get a domain name and you can provide the world your webshit. That's it, no intermediate players to tell what's wrong or get a cut of your pie. But yes, there are certain limitations what a webpage can do, mostly integrating to the rest of the OS, but that's by choice, since the scope of web APIs is mind boggling sometimes.

>>90846 This might be heretical of me to say around these parts, but I'm using Brave. Really any Chromium browser would suffice because they should all have an option in the address bar to "install" the webapp to your desktop. All that it's doing is writing a .desktop file into ~/.local/share/applications with a special StartupWMClass. The rest of the .desktop contents launch a profile for that specific webapp. Also from my experience, this doesn't work with containerized browsers (i.e. flatpak installed) since the environment somehow breaks the StartupWMClass specification making it to where every opened webapp will just open as a Chromium window that gets stacked in the dock bar rather than from it's own pinned launcher. Very frustrating! Take a look at my screen shots for Nextcloud Talk. Yes, I'm aware there's a dedicated desktop client for this, but I'm just showing this as an example. If you keep an eye out for that button in the address bar on sites you frequent, then give it a spin and see what you think. I should mention that at present, Firefox doesn't have any similar functionality on Linux desktops that I know of. I really wouldn't suggest wasting your time creating custom profiles to accomplish this. I prefer using Firefox and Pale Moon for web browsing and then Brave or whatever chromium-based junk out there as a webapp wrapper client (I only prefer Brave in this instance because of its built in ad blocking.) You don't need the browser UI or any of its junk because it's irrelevant insofar that it merely renders the webapp and nothing else. It's just another contrived abstraction layer by this point, but at least it gets you out of the browser as a pseudo operating system and back to your real desktop. This might make good bloggin material. Maybe I'll write on it and the workflow! >>90847 Uhh, yeah some of them work this way, but the offline functionality is far and few from what I've seen of em. Usually just by their nature they're online dependent and are useless without. And plus, isn't this the same thing that I'm talking about here? I suppose I am a little confused then because I have yet to actually see a PWA function in the way you're talking about. There are those elaborate conferencing Teams "phones" and camera setups that utilize WebRTC so I suppose the device interface is a PWA there, but they're absolutely useless without internet connectivity too. Idk...

>>90848 Well I just kind of assumed that if you mention things like maps or filing taxes you have straight up webshit in mind. Webapps make little to no sense in those cases. Useful PWAs I've so far discovered are: jspaint.app, audiomass.co, pairdrop.net, brie.fi/ng/. Pairdrop seems to be the one that may work in LAN only. Should test it.

>>90848 There is a way to use firefox for PWAs too. https://github.com/filips123/PWAsForFirefox I just tried it with librewolf, it works nicely. It's a bit messy, but you can use it with a "template" profile too, to use custom settings, userChrome and extensions.