-- Leo's gemini proxy
-- Connecting to cadence.moe:1965...
-- Connected
-- Sending request
-- Meta line: 20 text/gemini; encoding=utf-8; lang=en
> 1543 words, 34 paragraphs, about 5 minutes to read (300 wpm).
> First published on 2020-11-10.
> This article has been loaded 3266 times.
I have a lot of thoughts about Gemini and I'm definitely not going to be able to note them all down in this post. If you've already visited Gemini on your own you'll know about a lot of what I'm writing already, but this post is also supposed to be read by people on the other side of portals coming from my HTTP blog.
Inspiration:
Imagine web browsers. The original vision of the web was that a browser would connect to some server, download a hypertext document (text with links to other documents and media) and display it for you to peruse.
What we have today is ridiculous. We have newspapers that you can read once and then watch as they track you for the next five years. We have paragraphs upon paragraphs of garbage before a recipe listing. We have interactive experiences with sound and 3D graphics that can run inside the browser that, while definitely _cool,_ are also completely pointless. All of this adds bloat to the web, since all web browsers need to support it, even though almost no sites care about it. Competent web browsers need to add _all_ the features to make _all_ websites work, and that list of things to add is ridiculously long, constantly growing, and leaves the world full of holes.
Gemini today feels like the web did in the 90's, before corporations found it and ruined it, before half of sites needed to download a megabyte of scripts to show you a text article, before the flat and lifeless design of the mid-2010s, before every personal homepage being the same Squarespace template.
Gemini is hypertext and nothing more. You get text, you get links, and you get a tiny amount of text formatting to play with.
It's a cozier space.
Finding new things to browse is dependant on websites linking to them. Instead of recommendations from The Algorithm of what to look at next, you have to actively seek out new things by following manually placed links from the page you're on, and bookmark the things you find if you like them. There are also search engines that I haven't really tried yet.
The Gemini system adds back the personal touch and spirit that has long been missing. You won't find companies there - they have no one to market to. You'll just find people sharing their own creativity and ingenuity and passion.
Unlike the web (see Drew DeVault's articles) Gemini is designed to be actually possible for people to implement. The document that you are reading right now is sent to you over the internet by a program called Bliz that I researched for and created in one afternoon. Yes, one. There's a little more to do, and I'll continue to work on it, but all the vital features are in. Try doing anything in a single afternoon with the web.
Browsers for Gemini are slightly more involved than servers, since a server just passes text around while browsers have to create some kind of user interface. Nevertheless, this still isn't a huge deal. Having never done GUI things before, I imagine that I could research a GUI toolkit and build something that works on my own in under a week.
Try making a web browser, ever, at all. Maybe you could do it with a hundred people and a million hours. This isn't an exaggeration. Half the web requires a working JavaScript implementation to display any content, and implementing JavaScript is a gargantuan task on its own.
A hot buzzword you'll see on today's indie blogging scene is "static site". I'll explain what this means.
Web pages are written in HTML. HTML is a markup language, it includes the text content of the page and also directions that describe the structure of that text. Those directions include <p> to start a paragraph, <em> to emphasise some text (italics) and <nav> to label a site's navigation menu.
The HTML of a modern website will include the title of the page, a list of stylesheets and scripts that it imports, the entire site navigation, finally the content that people came to read, and perhaps a footer.
One _could_ create an HTML file with all that for every page on their website, but this is a bad idea: if you made a change to the navigation on one page, you'd have to make the same change on all your pages, taking care not to let it get out of sync! What a mess!
Additionally, HTML is not a comfortable language to write articles in directly. While it gives you the freedom to do anything a web page can do (because it _is_ a web page, of course), it's not nice to actually draft articles in because of "cruft" - parts of the language that get in your way. For example, to start a new paragraph, you must type </p> to end the previous one, and then <p> to start a new one, which is a lot of fingers dancing around the keyboard. A sensible idea would be to write your article in a separate word processor program (LibreOffice or Word), but if you do that, it's not HTML so you can't put it in your webpage directly.
Which brings us to static sites. A static site generator is a program that takes those standalone articles, saved as separate files, created with your word processor (or whatever article authoring tool you use), converts them to HTML, and combines them into a single copy of the site's navigation to make one final complete HTML file, which is saved. This happens for all articles. Once all of the files - which make up the "static site" - are generated, they can be uploaded to a webserver for the world to view. So the authoring process takes these steps:
Write your article.
Run the static site generator to get a bunch of HTML pages.
Upload the HTML pages for the public to look at.
A bit of a mess, wouldn't you say? It's certainly not the worst option for web publishing, not by a LOOOONG way, but it does leave a lot to be desired. For example... imagine if instead of needing to convert to HTML, the browser could directly display the same format that you write in?
That's Gemini's gemtext.
Gemtext is the markup language that most pages on Gemini, including this one, are written in. As I already said, it's a minimal set of hypertext features. One huge advantage of it is that the page source is already in a perfectly readable and editable form. To create paragraphs, in my editor I just type... paragraphs. To create headers, I put # at the start of the line. It's really that simple and it means that no static site build step is required, since the content that I write in my editor _is the website that people see._
Well, after all that hype, it would be sensible if I gave you an introduction on how to browse the Gemini system and what you can find.
You'll need to start by choosing a browser. Unlike the web, there are a few options. If you want a graphical interface, I'd recommend Castor (currently used by me) or McRoss.
Or if you want to skip the hassle of installing things you can browse through a WWW portal.
I call a Gemini site a star. It's cute. I don't know if there's an official name.
Without an algorithm constantly pushing boats of content at you it can be surprisingly difficult to find things to look at. To get you started, here's some stars that I like:
To find more, follow links you find from stars you already know about, try a search engine, browse an aggregator, or open something from the list of known hosts on GUS.
So you're sold on Gemini and now you want to serve your own content on it. Nice!
I think I may write a future article dedicated to this. For now, see Bliz.
I want to join Gemini's equivalent of a webring. I think that would be a lot of fun. Please contact me if you run anything, know of anything, or want to start a ring with me. Email cadence@disroot.org.
I don't know how long I'll stick with Gemini for. It could end this week. It could end this month. Or I could switch to publishing all my posts on it. I don't know yet.
The publishing process is certainly more fun on Gemini, so this may incentivise me to write more blog/gemlog articles.
We'll just have to wait and see what happens, together.
---
> No replies yet.
---
Looking for more?
-- Response ended
-- Page fetched on Fri Apr 19 19:33:18 2024