BlackWindBooks.com | Newsletter! | risingthumb.xyz | achtung.risingthumb.xyz | github.com/RisingThumb | site map

risingthumb.xyz 'Tis but a scratch!

The modern web is bloated

This will be a short blog post, explaining in detail why I think the modern web is bloated.

The problems

Website sizes. Alright settle down zoomers, with your gigabig chungus internet speeds. A lot of people still have very slow internet speeds that cannot handle over a megabyte of network activity. I can sympathise with the webdev that has to add ads or trackers on account of their employer's demands, however I will go on record saying that these 2 points are particular areas where the web speeds throttle. You have lots of useless Javascript getting loaded, and uncompressed ads and pictures that aren't even properly resized, among other issues. It's a MESS, and it's precisely why I use Brave Browser(it filters most of this shit but not all).

=> Luke Smith demonstrating the average size of Chicken Parmesan recipes. Each chicken Parmesan can store DOOM at least 3 times if not more!

Processor speeds. You know how you use useless Javascript to add ads that contribute nothing to your website? Yeah? Well they have to do their processing somewhere, so they do it on your CPU. Each browser handles this slightly differently so performance differs between them(As an example Vivaldi and Qutebrowser are slower than Brave in this respect). However, the problem here is that if an i486 can run doom, it should be able to run your website, and for the VAST MAJORITY of websites this would almost definitely not be the case. Remember, a lot of people also don't have the processor needed to handle these websites. Raspberry Pis and low-resource machines(which are getting more and more common due to the internet of things) get utterly destroyed by the processor requirements.

Anti-features. The original point of HTTP is a Hyper Text Transfer Protocol, and the original point of HTML is a Hyper Text Markup Language. People think that on these 2 points, it's fine to manipulate the document all over the place. This is an example of anti-features, as an example, when I search something on google and click the first link, and find it's trash. I will back out to the list of links, and click the 2nd link, but OH NO WAIT, Anti-features, engage! It props a little window under that link with "Also searched" and a bunch of shit I didn't want to click. On this point, it's also very hard to copy and paste content properly. Quite simply people implement dynamic shit, and I theorise it's front-end idiots because they think that all they can do to justify their job is to keep making it flashier and more dynamic. NO, that's how you piss me off. Another point on this, is websites that override the natural scrolling system, into some vomit-inducing motion-sick mess.

Accessibility. This one's simple, firstly, FAR TOO MANY WEBSITES ARE SCREEN-READER UNFRIENDLY. Holy fuck! How did we even get to the point that presenting fucking text, is too much for a screen reader? Quite simple how: because people use HTML(structure and content) as a way to force style. You should only do the minimal required structuring for the sake of style. Another point on accessibility is far too many people think low contrast, gray text with light gray background is good, and it's hard to read.

Implementability. This one's simple, the standard for HTML is insanely complex now, that only 3(maybe more?) viable web engines exist for rendering HTML pages(okay, yes there's more out there, but they are unusable for any site more complex than an imageboard or a normal forum). I'll list them now, Webkit, Gecko, and Chromium. First off, just fucking cut out Webkit, it's lagged behind quite a bit. Gecko and Chromium are the only 2 engines worth using on a modern internet, and that's why almost all web browsers use them(As an aside, I said yes there's more engines, Lynx and w3m both use their own engines but they only display text in a terminal, and usually it's utterly broken for most sites on the same basis as the screen-readers being broken. It's a pretty good way to tell if it's broken though).

The solutions

Stop writing shitty websites, that pull a thousand JS dependencies through a convoluted CDN to deliver anti-features to people and consume their bandwidth, mobile data and processor speeds(the last 3 points are also responsible for major mobile phone drain). Write your HTML, write your CSS, and write your PHP or CGI. This is how it was meant to be done, and this is the way that will stand the test of time, rather than outdated laggy glitchy Javashit. Of course, most people won't adopt this method as most people use shit like Medium or Wordpress or God-Forbid some other trash like it, to make websites/articles.

I'll also point out other protocol alternatives. Both aren't viable commercially, but for hobbyists they may be of interest. Gopher which is pretty much plaintext with restrictions of 80 column width without word wrapping, and a directory link structure. This is pretty outdated, and the flaws in it are assessed by Gemini which handles word wrapping and uses a markdown-like format. The protocol(read PROTOCOL, not the text format) is close to being frozen if it's not already, and that's actually a good thing. It means that people will be able to make clients and servers for gemini for years to come in the future, in a short time span AND as the text format, and it means it's relatively inextensible so people won't be able to extend the protocol with useless shite(as has happened with HTTP). Also look at IRC, that chat protocol is so simple that it has stood the test of time FAR BETTER than any proprietary or extended protocol.

There's pretty much no easy clean solution. One is to stop using HTTP and the other is to enforce good practice of HTTP(ironic how few people follow good practice because of how advertising has made the modern web terrible). Also here's a game to try. Take a webpage that is obviously bloated and convert it into a website with HTML, some limited CSS, and the images that are useful, and see how much faster it is on bandwidth compared to the original. What? It's 10x+ faster than the original? UNBELIEVABLE! SHOCKING! TERRIFIC! Now go and put those gains into practice on your live sites.

Published on 2021/04/08

Articles from blogs I follow around the net

BAT-EYES #13

NOTHING HAS CHANGED Saito knew how to turn the thing off and then turn it back on again. This gave him a leg up on the competition, who considered his admittedly limited technical skills nerd stuff, a form of eldritch witchcraft. This was stupid, but sinc…

via I'm not really Stanley Lieber. November 13, 2024

Neurodivergence and accountability in free software

In November of last year, I wrote Richard Stallman’s political discourse on sex, which argues that Richard Stallman, the founder of and present-day voting member of the board of directors of the Free Software Foundation (FSF), endorses and advocates for a ha…

via Drew DeVault's blog September 25, 2024

reblog - Moments of Introspection: Fun With AI: Cute Doggies

Moments of Introspection: Fun With AI: Cute Doggies: I am starting to get into creating graphics with AI. The following doggies, I created with  Idyllic . I was trying to get a god that is simi...:::Feed:::

via Filozofia September 24, 2024

Generated by openring