Web 2.0 is a monstrosity that never should have been evolved the way it did. It is now officially impossible to develop a new web browser, and Google's most-used Chrome keeps on kicking people in the nuts by removing privacy-focused extensions from the Chrome Web Store - because it's "not in line with their business model". Did you know the W3C specification set is 114 million words long? That's simply insane.

Lately, I've been implementing the Webmention and Pingback W3C protocols in my jamstack-augmented microservice, and I've come to realize that, although Webmention is the post-modernistic hailed and IndieWeb-backed successor to Pingbacks, they technically both suffer form the same problems. I honestly do not see any real difference between POST-ing an encoded form or an XML-formatted message. The specs are full of ambiguities that leave the door wide open to (1) interpretation and (2) a proliferation of wildly diverging implementations.

Don't get me wrong, I love JavaScript. I used to be that guy that constantly yells "check it out, const()={wait...]!" But the more I tried implementing something inherently simple, the more I realized how many different ways the same problem can be tackled in just one language. And then I used .replaceAll() in an older version of Node. Many programming languages evolve by borrowing ideas from each other, but really, why would JS need classes? As Rob Pike argues, keeping things simple in a language can be quite complex. I hope Go will keep it together.

First, there was CSS. Then CSS2. Then things got complicated, and people started using Bootstrap. Next, Font Awesome. Now, the consensus is to ditch that in favor for <svg/>. Then, flexbox. Material stuff. CSS3. Scratch that, use SASS instead! Hey, what's Tailwind CSS? The fact that these frameworks evolve more rapidly than I can think might be an indicator of something...¹

According to the HTTP Archive, the average size of a webpage in kilobytes has gone up to 2038.4. Try opening up any news site, blockers on (Firefox' Enhanced Tracking Protection, uBlock Origin), and count how many things get blocked on average. It's simply disgusting. Some more nauseating trends:

• Corporate blogs full of nonsensical bullshit. Come work for us!
• Like? Share? Plus? Click here! CLICK, DAMN IT!
• Cookies? No? Well then our site won't work. Sorry.
• Hey, fill in your e-mail address, we have a mailing list! No spam! Pinkie swear.
• You sure about the cookies? No?
• Cookies? Can't save, sorry, that's broken. It's a feature!
• Infinite scrolls make my head hurt.
• Your browser is not supported. Download Google Chrome here!
• Click me! https://www.amazon.com/dp/exampleProduct/ref=sxin_0_pb?__mk_de_DE=ÅMÅŽÕÑ&keywords=tea&pd_rd_i=exampleProduct&pd_rd_r=8d39e4cd-1e4f-43db-b6e7-72e969a84aa5&pd_rd_w=1pcKM&pd_rd_wg=hYrNl&pf_rd_p=50bbfd25-5ef7-41a2-68d6-74d854b30e30&pf_rd_r=0GMWD0YYKA7XFGX55ADP&qid=1517757263&rnid=2914120011 No? Oh sorry, I meant https://www.amazon.com/dp/exampleProduct! Thanks for your data, though²!

Back to the 90s Web, then.

As Max likes to call it: things come around full circle. I don't see this happening very soon for the mainstream web, but yes, there has been a movement. Leaner blogs, the "host it yourself" approach, the blogroll comeback page, and even guestbooks.

Then there's the ActivityPub protocol - an even more complicated set of agreements coded in a bunch of REST services. It gave birth to Mastodon (where I learned a lot about these efforts, thanks!), Pixelfed as a federated alternative to Instagram, and they're working on BookWyrm, a social reading and reviewing platform on the same protocol that hopefully proves to be more interesting than the now ad-ridden Goodreads.

I've joined The 512KB Club, averaging out at 248.4 KB for an article. That'll teach them! We could go back in time and look at my awesome 1998 webpage that includes vibrant .gif images and of course <marquee/> tags. Silly as it is, I still like that: it's simple. Yes, applets existed, but besides that, there wasn't much to it - yet. My minimalistic blog now could be considered avant-garde while I wish it was mainstream.

A few other outstanding initiatives are taking ground, such as the Small Web. Even if Google tried, RSS is still not dead, and it allows me to read what I want in the same simplified layout. And then there are tilde.town-alike initiatives: shared computers one can ssh into and leave messages - build a community. Sadly, the community will be mostly limited to Linux geeks. Of course Bulletin Board Systems still exist and that's fun too. But one thing does stand out even more.

Revolutionize Internet Browsing

Abandon the http protocol entirely. That is Project Gemini's plan to revolutionize internet browsing. A gemini:// website-uhm, "capsule" - sorry, is a very simple page that only allows plain text and basic linking. That's it. A sample source page looks like this:

# Welcome!

That's right. Oh, inline links do not work.
But hey, that's life.

=> https://brainbaking.com   my site

## Subsection

Neat!

No italics/bold, no scripting (well... there's the return of the cgi-bin dir), and a very simple way to parse and render the page. This makes creating your own Gemini client and server almost as easy as baking bread. There's already a lot of them in many different programming languages. I honestly would have preferred a unified Markdown syntax instead of yet another derivative, but hey. Required TLS certificates for static pages is another head scratcher.

If you're keen to check it out, there are Gemini proxy webpages on http that render a Gemini capsule. Lagrange is my user friendly Gemini client of choice. I've gone ahead and registered wouter.gr - who knows in the coming weeks I'll host a capsule myself.

I must admit I'm very intrigued by these initiatives, but this does not fix the issue at hand. It's still a very esoteric platform that started in 2019. Http will never simply vanish - unless Google says so. Searching capabilities are still a bit shaky too. And the biggest problem is: writing on gemini means writing for virtually nobody. Drew fixed this by letting Hugo spit out a gemini-compatible .gmi file, stripping out unsupported markup. If you're interested, read more at gemini://drewdevault.com/2020/09/27/Gemini-and-Hugo.gmi.

You'll have to figure out for yourself how to actually get there!