For those, who do not know what the Gemini protocol is, think of it as a modern, light-weight HTTP alternative without CSS or JavaScript. In layman term, you could see it as Web 1.0 reinvented. It uses GemText instead of HTML. For folks who want to try it out, you can either install a Gemini extension for your HTTPs browser (which kinda defeats the purpose, as modern browsers are heavy), or download a dedicated Gemini browser like Lagrange. Here’s a few sites you can access in Gemini.
Personally, I love it, although I miss a few stuff, like for example, multimedia, streaming and stuff like that. The memory foorprint is very low, and pages are super-fast.
You must log in or register to comment.
I like browsing it! You always find interesting people and writings about a wide variety of topics. It’s got enough users to have variety but not so many that it feels in any way corporate. It’s very much in line with the idea that limitations breed creativity. I’d highly recommend everyone download a decent browser and look around it a bit.
My only worry is that the Google Gemini AI thing will quickly suck all the air out of that name.
Oooh I didn’t know about this! Does it have a search engine?
EDIT: oh I see there is! Shoulda clicked through before asking lol
I just use noscript and call it a day.
It’s a cool idea and I was using it every day for a while. I love the gemtext format and I even made some “gempub” ebooks for fun. I have a site on flounder.online with some crap on it. Two things brought it all down for me.
The first is the hard TLS requirement. I’ve read all the rationales about this and still don’t see the point. I get the principle behind it but it’s not worth requiring that much infrastructure. It sucks all the fun and accessibility out of it. Which is my other issue.
We all know a platform can’t be TOO accessible without becoming like twitter. But if accessibility is too low you’ll end up with nothing but upper class tech workers moaning about the bougie problems that they created for themselves. The only capsules that had anything decent on them had HTTP proxies. It didn’t feel like a platform worth contributing to for someone like me.
I’ve heard about the Spartan protocol which is similar but has no TLS requirement. I’ve been planning on getting into that but all I’ve done is read about it.
I like it. Everyone these days seems to want web pages that are 5MB of dynamically generated junk.
My little website is just static hugo-generated stuff.
Sure, but you don’t need a completely new protocol to speed up websites, learn HTML and CSS and you can easily create fast pages for anyone to look at, not just those with a highly specific client.
People tend to really suck at limiting themselves. If you’re wandering around in gemini space you’re not going to run into pages with lots of ad banners, trackers and other monetization BS. You pretty much can’t. On the web, you can run into simple fast pages but it’s getting less and less the norm. And the lack of easy ways to monetize means it’s unattractive to corporations, which helps avoid creeping enshittification.
Gemini is light, simple, and easy to parse. It’s just lightly marked up text. Compare the size of Lagrange with the size of Chrome or Firefox. And nobody is forcing you to use it. 🙂
Ads are the main reason for all the junk though.
i use firefox with ubo and most websites are really fast this way.
My understanding is that there’s no way to make anything that’s not a read-only website with it. So Gemini users could have a lemmy proxy to get a read-only view of lemmy, but it would not be possible for a Gemini user to interact with any website without exiting the Gemini client and using another program. This makes the protocol more of a novelty/toy/dead-end to me, but I can imagine some people want to keep it this way.
You can use server-side forms to update pages, just like we did before front-end HTML became Flash 2.0.
Cause it’s so good to reload the whole webpage just to send a form
That only became a problem with giant ball of crap WWW sites. A <10KiB page is fine.
You can use server-side forms to update pages
source? documentation? code example? I’ve searched and never found anything
Here’s a microblog service that’s built on Gemini that uses forms to get data from users. gemini://station.martinrue.com/
I’m not going to install a random program just to look at a webpage that’s not open source. If forms exist in Gemini, it should be documented somewhere.
It is open source! https://github.com/skyjake/lagrange
Not possible. The Gemini protocol lacks anything like a POST method. The only way to provide user input to a Gemini server is through arguments encoded in the URL.
I understand the sentiment, but… HTML and some light CSS is just as fast and much more accessible. It just strikes me as something that defines itself in opposition to “thing everyone uses” for no good reason.
I got super into it for a while but there just isn’t enough “killer sites” to keep it interesting.
It really is just a little too minimalist. It’s not really easily possible to do forums etc.
It’s fine, I use Lagrange to read it sometimes, and there’s a few gemlogs I follow. But it’s in a weird space of “almost HTML, so why not just do HTML?”
Gopher still works fine, and has more clients (I still use Lynx). I like the clean separation of menus (even if you use a lot of
iinfo lines) and documents. There’s a bunch of gopher holes still out here. I haven’t updated mine in a couple years, but when/if I move it over to a new server I will, as kind of a back-channel to the site & blog.Why?
What problem does this solve over simple HTML/CSS pages?
Outside of a very specific niche I can’t see how anyone would choose this over normal HTML and HTTP/HTTPS, you’d need to run a new Gemini specific server to host Gemini specific files, created by Gemini specific softwares or Gemini specific developers, files that can only be read with a Gemini specific client.
This won’t happen outside systems with highly specialized requirenments.
The advantage is that it’s an obligate web 1.0 (-ish) experience. You aren’t clicking a link on a Gemini site that is going to take you anywhere crazy. There’s no tracking pixels and embedded content to get in the way.
It’s possible to attempt this by just following web 1.0 standards on your w3 site, and only linking to sites that do the same, and so on, but eventually there’s going to be a like button or an embedded video or something that ruins the experience. The web is messy.
Smaller spaces with constraints can be a lot of fun. Working within those constraints can breed innovation.
Sure, but untill web browsers support the protocol natively, it will never take off
This completely misses the point.
Why would you not want a broader adoption of the system?
because the point is not broad adoption, the point is not what features it supports, the point is the features that it doesn’t. It can’t track you, it can’t advertise to you (effectively), it’s meant to replicate that pre-corporate-enshittification feeling the WWW once had. The creators never imagined it would get as big as it even currently is.
I still don’t see why it would be bad for browsers to add support for the protocol, that would only ever increase usage of the system, promoting the system and the ideas of the creators.
Example, one of the first times I used IRC I fiest began using a web based IRC client called CGI IRC, then I got on Chatzilla, a Firefox extention, then moved to mIRC, then added the Noname Script to mIRC, then moved to HydraIRC, after that I moved to Quassel, then I settled down on irssi for several years, untill I slowly moved away from IRC due to lack of time.
My point is that with a low barrier to entry using a web client and then a Firefox extention before jumping all the way in made it easier for me to join, and never hurt IRC in any way.
The same is true for Gemini, having a simple system with a low bar to make text content, and a simple way to access it will only be beneficial, it will lead to others finding it quicker, creating a more robust system, then as users mature they will find better clients and use those instead.
To claim that the creators don’t want people to use the system is just ridiculous, if they didn’t they would have kept quiet about it, it slso goes directly against the goal of creating a pre enshittification community if no one can use it.
There are limitations in the system, let those be used to preserve the old time feeling, without stopping specific types of clients
Well, firefox used to have support for gopher, but maintaining it was too much work and support was removed in firefox 4.0. Even now, with it gopher and gemini being the most popular they’ve ever been, neither of them have built-in support from any major web browser.
Also, it’s not that the creators don’t want people using it, that’s not what I meant. It’s just that they didn’t expect the level of adoption they currently have.
There’s no tracking pixels and embedded content to get in the way.
Looking at the Gemini docs, I feel like I can recreate a way to add tracking and embed content. I could be wrong. But it looks possible.
And if that is the case, once marketers see the potential, all the tracking, popups and gated content we all love so much can happen on Gemini.
Possible only if you add that functionality to Gemtext, but currently not something you can do with existing clients. It’s pretty much just modern Gopher.
It may not be particularly useful, but I welcome a challenge to the current status quo. The Internet is a powerful resource, and we’re still building on top of the first protocol that worked back in 1991 to navigate it. Gemini isn’t something I could see having any mainstream appeal, but it’s absolutely worth experimenting with alternatives to the World Wide Web. Having more than one functional open standard could help revolutionize the Internet in novel new ways.
I think you’re thinking backward. Internet is what it is because a single protocole unified it. Without it, you’d have island working with only one browser each, some would eventually die and with them large parts of Internet would disappear.
Internet works on unified protocoles. Everything that challenged this model is bound to fail. That’s why javascript is so successful eventhough it’s so shitty as a language.
Evolution can only be iterative.
I don’t understand why this would exist over noscript + using the websites you actually want to use. If you want to spend time on minimal websites, there’s no reason you can’t use html and http to do it.
it looks like this was a college student’s weekend hacking project and some people took it way too seriously and now have this social idea about how intentionally inferior tech is going to revolutionize… devolutionize? the internet. This is not snapchat. This is cave painting.
You should try burning man. Or an art gallery.
Get over this “I don’t see why it should exist” thing.
Another one of those “solution in search of a problem” things. It really doesn’t solve any of the drawbacks of HTML/CSS, it just does the same thing in a different (way less supported) way.
I was an early adopter of Gemini and even hosted some stuff on there years ago, but ultimately I don’t see a point when things like gopher exist and Gemini is a wasteland when it comes to interesting stuff to browse. Though admittedly the concept of an encrypted gopher protocol is pretty nice to me. I feel a lot more of the old internet feeling on there, i.e everyone else using it is a like-minded hobbyist with no corporate overlords. But even then things like activitypub that we are using right now also have that so idk.
CSS and js are nothing to do with http. In fact neither is html really.
It’s just a protocol for transferring text.
