Bye, Gemini
makeworld.spaceI do think Gemini suffers from a dearth of interesting content, which is interesting because the content you find on there is pretty much exactly what you would expect and the things that feel missing are the things we often bemoan about the web.
If you look at an aggregator like [0] you see 10-15 new posts a day, which actually isn't a bad level of activity. But it's all just people's personal blogs. Which is fine and, again, is probably how it is intended to be. But unless you have a personal connection to these people or happen to be intensely interested in all the same things as them, it's difficult to stay interested in their personal blogs.
Things you don't really find in Geminispace (unless it is mirrored or proxied from the web):
- news or serious journalism
- any kind of social media
- any kind of interactive content
- any kind of multimedia content
Some of that is due to inherent and very intentional limitations of the protocol. Some of it is just due to the community being much smaller (ie, the web has plenty of interesting text-only content that could be on Gemini, but is not).
I still like Gemini and will continue to browse it, and it's important to point out again that it was never intended to replace the web. But I remember to check in on it less frequently now and it seems I am not alone.
0: gemini://warmedal.se/~antenna/
On the early internet, it was all personal sites.
But we had attempts at journalism and tons of interactive content on those sites.
When there was no social media, people learned HTML to communicate, and we got all the variety of the early web.
Now people make sites because they like coding, and use them to communicate about coding. Or they use social media and comment on random drama.
The internet eats it's own refuse like AI does. Without real life stuff to talk about it's pretty terrible. Video games were the main interesting internet native subject, aside from tech itself in the pure sense, but gaming culture has become almost a 4chan offshoot, less interesting to everyone else, while the games themselves are full of DLC.
Modern computers are fast enough for bloated sites. I don't think the issue is tech(unless you're really unhappy with the privacy situation). It's that all the content is made by people who spend all day on the internet, and it's all about tech.
And tech is just going in circles. With less real world connection, everyone just wants to be better at writing code, to try new languages, etc. It's philosophy as much as real tech, or maybe like some modern cyberpunk version of meditation, they're all just seeking simplicity, and it doesn't make much sense to people who didn't join the scene because they loved elegant ideas.
It's like reading 10 biographies and writing about them and your experience reading them, vs the old Internet where you went and did stuff and wrote about your life.
Most early sites were personal but they weren't blogs. Many of them were even technical, but people made an effort to write for posterity rather than just for today. People wrote a site like they were writing a reference book, even if most of them got bored halfway through.
The fact that grandparent looks at Gemini and sees "10-15 posts per day" rather than "x sites with y pages" is the problem. Today's culture sees anything written on the internet as disposable; at best it's a magazine article, more likely it's a leaflet.
> On the early internet, it was all personal sites.
Assuming you mean WWW, this is patently false:
https://en.wikipedia.org/wiki/List_of_websites_founded_befor...
"It's that all the content is made by people who spend all day on the internet, and it's all about tech."
And "news" services that utilise the internet are more often than not biased toward so-called "tech" news.
It's like the internet is constantly promoting itself.
Then one looks at the advertising. The businesses that advertise the most are the ones who are entirely internet-based.
I'm not suggesting this is a new trend. Definitely not. But I had hoped it would change.
Just noting the meta of this commentary being a common reoccurency on this very tech-news commentary forum ;)
Who is everyone? There are proportionally very few coders in the world and even fewer making their own sites compared to non coders.
> There are proportionally very few coders in the world
I wonder how many people can code but for whom coding is neither a major part of their work, nor a significant hobby?
My brother is training to be an oncologist. But he did a computer programming elective as an undergraduate and wrote some Python programs, and I believe got a decent grade too (need to maintain GPA). That was years ago, but if he felt the need or motivation he could dust it off. I’m sure he’s got the brains to learn more of it if he wanted to, but between a young family and a very demanding training program, I understand why in his limited recreation time he’d rather read a sci-fi novel than muck around with computers. But when his kids are older and he’s more established in his career, he’ll have more time-if he fees the itch.
I joked with him recently that there is nothing stopping someone with his career from doing a bit of mine as a hobby, but the reverse is not true: “hobby oncologist” sounds rather disturbing.
This is why I switched from programming professionally to doing something else. Pretty much every field of knowledge (that existed before 1950, at least) is interesting once you get into it. Most of them you cannot do for fun. I can program for fun. I can do something else for work. Doing programming as a job ruins my favourite hobby.
I think the point was that the people producing content for their own personal website are a small, self-selecting group, with social media catching most of the rest of the online populace.
You've really succinctly captured why I don't enjoy a lot of modern tech content online these days. It's too self-referential, too steeped in itself. The people involved are motivated by a reaction to other online content and want to produce new online content. Then there's the online beefs.
> Without real life stuff to talk about it's pretty terrible. Video games were the main interesting internet native subject, aside from tech itself in the pure sense, but gaming culture has become almost a 4chan offshoot, less interesting to everyone else, while the games themselves are full of DLC.
I think this is the exception. Most gamers I know find some social connection in it, even if connecting with other gamers about the game they're playing. The programmers just tilt at ever simplifying windmills pursuing their platonic digital ideal.
> On the early internet, it was all personal sites.
Personal sites maybe but people wrote about stuff they liked. There were thousands of sites that just covered some niche topic or what today would be called a fandom. Sure there were sites that were just someone's journal about themselves but easily just as many semi-authoritative topic focused ones.
The Internet Archive has a number of scans of old "Internet Yellow Pages" books and Wayback archives of DMOZ and Yahoo!'s directory. Take a look some time, for any given subject there's tons of sites listed with tildes in the path.
> the content you find on there is pretty much exactly what you would expect
Mostly blog posts about Gemini and Gemini software, in my experience...
Maybe it's time for Mercury, an even simpler protocol: you open a TCP connection, the server sends back a chunk of bytes consisting of ASCII 0x20 through 0x7e (and newlines). Clients must print the response to the screen verbatim. You "link" to other sites by including a hostname or IP address in the document; the user can then type or copy-paste that into the client's address field.
Make the protocol so boring you have to write about something other than the software? It might work!
I wonder if the mostly proprietary OSes and browsers has anything to do with the early internet's appeal. Open source invites you to improve and customize the software. Does the ability to do so change the psychology and put the focus more on the app itself? Is it related to entropy, where software has lots of other possible ways that it could have been done differently, all of which sit in ones mind like alternate timelines distracting from the present?
With open software software(Aside from ultra minimal software defined by a specific ideal) you can always improve it with some effort, so perhaps it's always disappointing to some degree especially to anyone who knows how to code?
It's one of the only things people do where there are frequent updates. A textbook might go years between editions. Painters will usually return to similar themes and subjects when they feel they could do better rather than redo a painting verbatim. Software is never done, and software people are never done thinking about it.
How can we get the same results in FOSS that we had on the early internet, when people just said "it works, lets use it!" without using ultra minimal stuff that once again is only interesting to tech tinkerers, and doesn't really do much?
neocities.org is way more interesting if you want to see people just building sites for stuff they're interested in
This makes me think of a gag from the Simpsons, where the only dialogue coming from a ham radio was "I have a ham radio"
When the Web first came online in the early 90s the only web sites I remember were "How to write HTML."
It took me about a year of having a browser installed before I started using the Web.
Finger already exists.
And it’s been used for blogging too! How history repeats itself… I remember John Carmack’s finger blog.
I recommend port 17.
Hush, this is a new movement, a rejection of old bloated protocols like Gemini and HTTP!
Also RFC 865 limits the QOTD response to 512 octets which isn't nearly enough.
Those are the things I am least interested in geminispace lacking. As you get older you realise that the "news" is mostly not news, especially these days. Social media, interactive media, multimedia content.. Not really the point of geminispace at all. Gemini is meant to emulate the old internet, when it took a minute to download a each photo someone posted of their trip overseas.
I think a much bigger problem with Gemini is that, at least when I last looked through it, it was mostly people blogging about Gemini, or about the small web, or about software minimalism, or about community building, or some other meta topic. I'm sick of metatopics! Discussion online has got too meta. (The irony of this comment is not lost on me, I assure you.)
It's also a very narrow strand of people in the "community" of gemini.
No wonder really. Markdown + images fits many use-cases, worse-than-markdown without any multimedia just don't. It doesn't even work for blogs, if you want to brag about what you did, no pics, if you want to have any viewer interaction whatsoever, no dice.
Requiring some wanky app to even enter it doesn't help the issue
I looked into Gemini. I love some of the core tenets of the philosophy (Drew Devault), but it quickly dawned on me after reading some of the stuff on there: I won’t name the ideology but it almost feels like Gemini was developed to propel it. There is not much diversity of ideas.
I think clans/cults/whatever are fine. Subcultures are ok. But you can’t pretend to create an “open” platform only to have the whole thing designed to be the opposite.
I feel the same about Mastodon. Cool tech, undeniably. An echo chamber, regretfully.
I feel the same way about Mastodon. The founding generation of that ecosystem has very strident views about certain matters that are foreign (and in the sense of "never even think about it", let alone "I disagree") to probably over 90% of the world population. Expectations are that the bulk of entrenched Mastodon instances are going to ban federations with instances that permit discussion beyond what that founding cohort is comfortable with, so here, too, the technology almost exists to propel ideals, too.
My experience with Mastodon was that you can choose between:
* Echo chamber islands with purity spirals that enforce the banning of Problematic instances
* “Free speech” instances where all the refugees from the above go, but which are just a different flavor of cesspool
Moderation is hard.
Yeah likewise. One of the reasons (well I'm an old-time SSB fan) I'm interested in Bluesky is that defederation isn't really a thing and so the failure modes of splintering communities isn't as big of a deal. When individuals have more power and mods have less, the impact of politics on the community would probably be more diffuse. Or that's the hope anyway. The mods = gods era of the internet was always a colorful one.
> I won’t name the ideology
Not sure what you're referring to here, but to me the ideology behind Gemini is self-indulgent navel gazing. Sure it's neat, but that's it. It's literally just a subset of something, but no emergent property or benefit is borne from this shrinkage, so the end result is not very interesting.
That's fine honestly, I don't think the authors wanted to change the Internet with Gemini, I'm more surprised by the early adopters that are now suffering from post-hype depression, like the author.
> It's literally just a subset of something, but no emergent property or benefit is borne from this shrinkage, so the end result is not very interesting.
Often this de-bloating is a benefit in itself, in a technical sense. But I suppose that's not what you were talking about?
I had no idea who Mr. Devault was before this article.
So I googled, did some reading, read his blog some, and yes, he strikes me as one of those people who have entirely too many strongly held opinions, about a great many things that are probably not very important to peoples day to do lives - and will exclude people who do not agree with him in entirety.
False, there are lots of gemini places, look at them at gemini://medusae.space, and most of them are not Gemini related.
Ditto in the Spanish geminiverse: gemini://caracolito.mooo.com/deriva
On news, well, seriously, RSS' are not web pages and often work better than the pages themselves, and gemini://gemi.dev it's doing a great job. When a gemini service as Gemi shows up the actual content at 5% of the size of the actual web page, something it's really wrong with the web.
Possibly dumb question, but why does gemini have to be difficult to use? Why not just serve simple renders of Markdown+inline media?
Why not just use something like Obsidian for this purpose?
> Gemini is a reaction to bloated modern websites
Correction: it is an overreaction to bloated modern websites. Interesting in various ways. Artificial constraints in an era of gratuitous excess can be useful to trigger creativity but my sense is that the territory Gemini revisited is simply too well trodden. Not a particularly fertile ground to seed an alternative universe. Heraclitus said, "You cannot step into the same river twice, for other waters are continually flowing on."
Added thought: The vision of a lean web is more than valid, but it needs to add something new and exciting even while it subtracts a lot of useless bloat.
Yes, I would like to see something between the web and Gemini where the style is always provided by the browser, like reader mode, and there is no javascipt, but there is a fairly rich set of media like images, video, and sounds.
You really need to easily "interact" with content by replying to posts, chatting, and up-voting, searching, and navigating.
This article really resonates with me. It lots of ways I want to like gemini but when it really comes down to it the content just isn't very good.
Secondly, the restrictions seem great until you really could use a diagram or an actual equation or a million other things that aren't just plain markdown basically.
I am squarely in the wheelhouse of people who should be into gemini. I love the terminal. I have used vim for 25 years and emacs for a good 5 or so years before that, I have used textmode browsers and extensions like vimium etc to make my non-textmode browsers be more texty. My own website is almost entirely just plain text.
If it was obsidian-style markdown (ie including embedded images, latex and tables and especially diagramming using something like mermaid or excalidraw) it would be amazing[1] but as is, it's just not quite enough to actually be used for anything other than what's on there which is basically people just navelgazing or talking about gemini itself.
[1] And I expect that slightly less restrictive palette would make it attractive to a lot more folks to produce content for and therefore it would be a lot less like a barren technohipster wasteland as at present.
I love Gemini, and I'm bummed a bit that the maintainer of my favorite CLI gemini client (amfora) doesn't have interest in it anymore.
For me, Gemini's a sweet spot.
Yes, Gemini is slow. I like that. I don't need dopamine hits of constant content, even if I _have_ become accustomed to them (he said, on a website designed to have fresh links for fresh dopamine hits all the time). I like a slow "let me check and see what's come in across the past day or two or three", with the selection being pretty manageable -- a dozen or so posts.
I like the lack of tracking, the lack of JS, the lack of CSS. It means that we're focusing on the words on the screen, not the gizmos and flair around the edges. It's plaintext at its best, IMO. You write your stuff, I'll set my client to render it however I like to read it, everyone's happy.
Yes, everything so far is just personal blogs. I love that. It's more genuine, because it's just people talking about what matters to _them_ -- not what "their audience" wants, or what drives engagement -- and that's where the real good writing lives. I don't want someone to give me some AI-generated spew so that they can get pageviews for ads, or to give me soulless marketing copy so they can promote their brand so they can get pageviews for ads or merch-affiliate links. I just want to pop into a little corner and hear someone tell me something about a subject they care about, and have my own world enriched by it.
My only complaint about Gemini, so far, is that I do wish discussion was more feasible. I've seen some duct-taped-together comment systems, but it'd be nice if that were more of an easy-to-use and easy-to-set-up thing.
Gemini threw out the baby with the bathwater out while chasing simplicity.
The impossibility of inline media makes a huge amount of otherwise interesting content (tutorials, travel blogs, museum sites, personal hobby sites) boring and tedious.
The impossibility of "easy-to-use and easy-to-setup" commenting system stems from the same "so simple as to be nearly unusable" ideology.
I don't know much about Gemini beyond a cursory look, but I don't think inline media is impossible. IIUC it is up to the Gemini client to decide if it displays media inline.
My understanding is that Gemini clients don't display inline media in order to defeat privacy-compromising techniques such as tracking pixels. A Gemini client could very reasonably decide to show images inline if they are larger than 16x16 for instance.
It means you can't create content that you will know will display the same in different clients.
I love the idea of Gemini. But I also find myself forgetting it exists. Maybe the addictive way we do web design this days is the culprit. But, honestly, maybe is just the content quality.
Maybe, once you take out the rose colored glasses, the past internet wasn't so great?
I remember being a kid and having a blast on the 2000's internet, but it was really hard to bump into a really good site. Tbh most of them were kinda crappy.
Like with past friends and past games: I missed the good times, but if I tried to recreate today the same things like I did in the past, I realized that everything changed: Some friends stayed friends, but some others simply didn't. And some interests didn't interest me anymore.
I also realized that what I don't really miss the old internet. I miss being a kid, having all the time in the world in my hands, with total freedom to play with the most absurd activities with zero regards about productivity and being amazed with the most simple things.
Quoting one of my favorite HN comments that captures this sentiment really well [1]:
> I think it's easy to forget we're not the same people we were 20-25 years ago. I see all sorts of lamentations about various things in the world changing and becoming less magical. But as far as I'm aware, I could still go and argue about bands in chatrooms. I could talk to other writers and dream about my future best-selling novels. I could go read random opinions about any subject and get into an exhilarating flamewar about it.
> I don't want to do any of those things. I'm in my 40s and I have 3 kids. The internet 15-year-old me experienced was magical because _I_ was a blank slate. Every new friendship was thrilling, every new skill opened up infinite horizons, every nook and cranny felt like somewhere I could belong. But life moves on. I'm more than half-way through my career, perhaps not the one I was expecting. I didn't marry the girl I met on IRC. I don't have strong opinions about Radiohead anymore. I find people, however delightful and kooky they are, quite tiring having got to know 10,000 of them at this point.
> I know all this is true because my kids love the internet and find their place in it with all the joy I used to. And I'm pretty sure older generations frowning upon it all is part of the rush anyway.
That quotation is funny, because the forums I know where people argue about bands (or classical or jazz composers or performers) are largely kept active by users who are already in their late 30s or 40s and up. Passionately arguing about band minutiae requires typing longform text, but younger generations are mainly using their phones where they don't write much longform text, or even don't visit third-party forum websites at all.
Kids these days, amirite?!
I still feel weird about video. Gen Z just confuses me.
Gen Alpha will embrace telepresence. Then Gen Beta will reject the esthetics of disintermediation altogether and only relate face-to-face. Then Gen Gamma will create kinetic visual vocabularies, a synthesis of dance, sign language, and improv.
The pendulum will continue to swing.
Each cohort has their own thing. Different medias aren't better or worse. They're just different.
For the record, my post above made no value judgment about which medias are better or worse or ragged on “kids these days”. I just wanted to emphasize that the forum behavior which the quoted person claims to have aged out of, is actually being kept alive by his very generation.
Forgive my non sequitur. Rereading my words, I don't think I was replying to you or your points. But now I can't tell who (or what) I was responding to. So I'll just blame my pre-senility.
Younger generations are more engaged in Twitch streams, talking to each other while a central person, doing as "presentator", streams video.
Younger generations are consuming product from a millionaire, and communicating back to them with hearts and cash, as a chat runs along the bottom of the screen.
> I remember being a kid and having a blast on the 2000's internet, but it was really hard to bump into a really good site. Tbh most of them were kinda crappy.
> Like with past friends and past games: I missed the good times, but if I tried to recreate today the same things like I did in the past, I realized that everything changed: Some friends stayed friends, but some others simply didn't. And some interests didn't interest me anymore.
I think it runs both a bit deeper and broader than just "what I liked then, I don't like now".
In the 2000s and prior, the promise of everything involved seemed infinite. We looked at some small or bad website and saw the potential for it to become bigger and better. If it can do x, maybe it can do y, z, and more.
Today, we know what the full potential of a website is, and it's not nearly what we thought it was. So when we see a small or bad website now, we know that all it can do is maybe be a little better, but trying to improve it could just as likely make it unwieldy and worse.
In our disillusionment, we try to recapture that awe-inspiring feeling of infinite potential, but instead of making something new that's full of potential we only pantomime what we were doing 20-30 years ago, by forcing arbitrary virtual restraints on ourselves inspired by what were once the limits of our reality. Gemini, Hypercard, virtual game consoles, etc.
At best the result feels as artificial and empty as a more emotionally detached examination of the motivation might've suggested, and upon recognizing that we move on from it. At worst, we recognize how much of that potential wasn't ever realized — and the often arbitrary and cynical reasons why — and feel even sadder, because our memory of that infinite potential is tainted by the reality.
It's like the moonshots of the 60s and 70s leading to people thinking we'd be colonizing the solar system or breaking the speed of light by now.
Just like any limitation, all of these virtual-limitation experiments have some legitimate uses and can inspire creativity. But it's become clear that the people who dove into it for the doomed nostalgic hope of recapturing that feeling of potential have recognized that and moved on, either to newer pastures of actual limitation-breaking potential or other nostalgic boondoggles.
Yup, nostalgia projects often don't work because people don't really want the exact same thing as before, they want something which realises the potential of what there was before.
Good example is the game Minecraft - when it was new, everyone liked it for the sandbox nature and the endless possibilities. Now that most of the creative things in the game are either dead or stagnant, and the developers are actively working against the game's potential, now it feels like a dead end. It's similar to the web in this regard.
Nostalgia projects should just start from the old principles and build something which is an improvement over the things we had back then and which we have now.
> Today, we know what the full potential of a website is, and it's not nearly what we thought it was.
I mean, people attribute Arab Spring to Twitter (maybe it's even when traditional news outlets started paying attention to social media? I don't remember, exactly), so I'd say the potential of a website is pretty significant, although it really has nothing to do with it being a website. More about being a place for people to communicate.
I still have basically the same rose-colored glasses for the web. Sure, websites are not as exciting as they once seemed, but not much about the web's potential for people to connect has changed, aside from increased awareness by parties who would stand in the way, but in response ways of avoiding their attention have also arisen.
>I love the idea of Gemini
Yeah, same. I think it's just that so much of the Internet comes to you now (via social feeds, emails, or RSS feed), the idea of surfing to find stuff is mostly dead to us. And you have to fire up a different browser to do it? Just too easy to forget.
There will never be a killer app for Gemini, and pretty much anyone who's good enough to be worth following will never fully commit to the protocol.
I like that it exists, but that's about it.
I like the idea of Gemini, but having just a subset of the web's functionality is not enough.
While it might not be a perfect analogy, many prefer using the terminal/shell/TUIs is because it is more powerful and efficient than the alternatives.
And even though I appreciate Gemini, I wouldn't say that it is more powerful and efficient. Granted, that's not its goal either.
> And even though I appreciate Gemini, I wouldn't say that it is more powerful and efficient. Granted, that's not its goal either.
Correct! Gemini is a mix between an art project and a political statement. It doesn't, and is not meant to, provide significant value over the web - a quick look at the specs makes this pretty clear.
If you view it through this lens instead, it's pretty neat, though.
i love the spaces that gemini created, but agree that (from a protocol and technical standpoint) gemini and gemtext were more "frustratingly inferior" rather than "elegantly simplified".
This is my favorite take—constraints can breed innovation, but usually only when the constrained system doesn’t already have a complete substitute that isn’t constrained.
I’d love to see a system that went hard on elegance. If we’re tilting at windmills, the alternative to web I’d like to see, is the fully Scheme web: S-exprs for content, DSSSL for styling, and Scheme for logic.
It was the web we’d have had in a different timeline.
Or it could use PostScript for ... everything.
At that point, you have a digital fax machine that prints out a message when you you dial it.
Which beckons the question: does anyone know of or remember an automated system like this, where you rang a number and it would ring your fax machine back with a document? I like to imagine I could have made my millions on 'dial-a-newspaper' back in the 90s!
> [D]oes anyone know of or remember an automated system like this, where you rang a number and it would ring your fax machine back with a document? I like to imagine I could have made my millions on 'dial-a-newspaper' back in the 90s!
There were many, though they were not seemingly used for newspapers. Generally, you subscribed to a newsletter, or were subscribed against your will to a mostly-ads-but-pretending-to-be-a-newsletter. These would arrive on your fax machine, usually on a schedule. Or you dialed another fax machine that would respond with some infopage, specs sheet, or government document (and sometimes let you choose from a published menu by entering other codes after connecting).
I was only ever on the receiving end, so don't remember the names of any of these systems. I believe the dial-a-menu faxes mainly existed on high-end machines owned by large corporations.
> At that point, you have a digital fax machine that prints out a message when you you dial it.
Isn't that mostly what we want from internet?
How much sense does it make to use Gemini as compared to just keeping your HTTP implementation and your web pages markup as minimal as possible?
The idea behind Gemini is that it's useful to have a clean partition where you know the social "crowd" matches a specific type. For someone to use it they've gone and installed dedicated software which intentionally limits itself to what this crowd likes to engage in. This signals they are interested in the same kind of space. The utility of the dedicated software is not in providing some other level of convenience, it's separate strictly so the user "crowd" stays separate. Otherwise I could just register gemini:// as an aliased protocol of HTTP and it'd work in normal browsers or gemini clients all the same.
If Chrome supported Gemini tomorrow a significant portion of the user base would stop using it. Of course that's unlikely, Gemini's trust-on-first-use security model is something unlikely to be added net-new to a consumer focused browser.
> This signals they are interested in the same kind of space.
It appears to have achieved this goal to perfection, since most posts in gemini are about ... gemini.
It matters to some people that their preferences are enforced by the tooling.
How much sense does it make to switch to some immutable data structure or even a language with support for such rather than just not mutating the one you have? (Works for me.)
There is a philosophy behind Gemini which is to fundamentally simplify the web, not jut individual web pages that opt-in.
No site is an island. Even if I keep my own website's implementation minimal I still end up interacting with the rest of the internet via web pages with less minimal markup.
However, if you also had a like-minded community of similarly minimal websites, a mechanism to minify websites, search engines that prioritized minimal (or minifiable) websites, and a social convention of avoiding linking to any website unless it was minimal (or minifiable) then you'd end up with something like the Gemini community.
In fact, I suspect something like that will supplant Gemini in practice. Browsers already have reader modes and some search engines allow you to block certain domains. I could easily imagine search engines or browsers providing greater end-user control over this sort of thing in the future.
> search engines that prioritized minimal (or minifiable) websites
https://search.marginalia.nu tries to be this, and is now being worked on full time by its creator.
Also https://wiby.me
> search engines that prioritized minimal (or minifiable) websites
I actually like that idea.
A few have popped up on HN. The problem is, the quality of search results is often terrible, because the assumption that complexity of layout and density of information, much less relevance, are mutual opposites isn't valid.
> the quality of search results is often terrible
Having used a few, I agree there's room for improvement.
> the assumption that complexity of layout and density of information, much less relevance, are mutual opposites isn't valid
I agree, but has anyone claimed that there is a correlation between a page's complexity and the quality of its information?
My experience has been that the folks who willingly disable javascript, use niche minimalist search engines and browsers, or otherwise prioritize a minimal ("smol") web experience are acutely aware that they are making a tradeoff. They sacrifice access to a significant fraction of the overall web including many high quality pages in exchange for a user experience they prefer.
Like many who have tried it, I found the experience too limiting, but I appreciate what they're doing. If nothing else, their efforts are making the web better for people who rely on accessibility technology and are forced to make such tradeoffs.
Gemini also removes cookies, actual forms (there's a way to submit text, but not as a form), and any possibility of extending the protocol.
It throws the baby with the water, a bit.
Despite all that, it is still useful, but not enough to ignore all that it could have had but is missing.
Indeed. What would it be like if websites were designed to "look good" using a text based browser, like Lynx?
In the early days I used to use Lynx, and it was actually really good. It was always quite fast, easy to navigate around with the keyboard, able to zip through things like navigation menus quickly.
Didn't work to well with image maps! (A thing back in the day), but it was a nice experience.
The endless dropdown menus people stuff websites with spoil the fun as there is no nice way to render that.
If a menu is implemented in a semantically-correct markup a browser potentially can grab it and display in a sidebar.
A modern TUI should be able to handle that nicely but not sure if Lynx is still being actively developed.
Readers don't need a heavy web browser at all.
Even for a minimal html, everyone uses a different "minimal" subset.
Gemini felt to me like the web's quiet cabin on a lake. I'd view through a terminal emulator on a Mac Classic.
The lack of infinite content makes it peaceful and manageable. The lack of search (or maybe it exists and I just didn't care to find it) makes finding little haunts and communities more rewarding.
It's just some personal blogs about synthesizers or cereal recipes- but it's genuine and not people trying to build a brand. No one is trying to sell you anything. It's felt totally uncorrupted.
when i consider alternatives to http, i find that the main thing http is missing is a directory index at protocol level. it should be possible to browse the structure of a site without having to load and interpret a full html document. an index would help discovery of the site, especially, but not only for users of accessibility tools.
those who don't like html can easily host only text files if they want to. the main limitation here is the browser lacking a two pane directory/view mode
I think Gemini just went too far with how much they got rid of.
I think the ideal middle is using a stripped down HTTP subset, with HTML5 (maybe with some things removed, like script tags), and the latest CSS.
No JS. Focus on the document. Make any changes needed to improve things. No crossdomain cookies. Maybe no cookies at all. Only GET and POST maybe. Only allow some very specific headers maybe and strip anything else. Figure out a more novel system for secure transmission, maybe by default.
I'm thinking of starting something with Servo, I know it's not HTML5 yes but it could be a good way to limit things. Or maybe Libweb from the Ladybird project.
I think Gemini could have been fun for vintage and low powered systems, but then they slapped a mandatory TLS requirement on top of it. That's the entire reason I have no interest in Gemini.
It doesn't provide anything that HTTP or Gopher can't. I don't see the point.
I actually consider TLS an anti-feature in the case of Gemini.
What I mean is: what use do you have of TLS that is not about commercial applications and gatekeeping?
The obvious use of TLS is for e-commerce, so that you can pay securely. And for your passwords when you log into an account. But Gemini is not designed for e-commerce, it is designed for reading text documents. And it doesn't look like the philosophy is to restrict access to documents, so no need for accounts and passwords. What do you have to hide on Gemini?
Ok, imagine you have something to hide on Gemini. Here TLS offers only limited protection: it won't help you if the server is compromised, and it doesn't hide your connection to the server, only the data. For that, there are other privacy-oriented protocols that are much better suited for that.
So, TLS is a great protocol for the commercial web, and a rather weak one for privacy. Why do you think Google pushes so hard for https?
And for the anti-tampering part that is worthwhile, I think it would be better done at the document level, for example with PGP-style signature blocks.
TLS runs fine on 100-ish MHz systems. That is less powerful than most modern embedded devices and covers vintage deep into the 90s.
TLS was unnecessary, I agree, but if it was plain text it would not have been more interesting to me, a computer nerd, than the current iteration.
I might have written a client, and just forgot about it because a plain text internet is really not very interesting for any demographic.
The author of that post makes a good point: using Gemini offers no advantage over how I already use the Web. I think that goes for more than a few people, and not just the author.
Admittedly, I've never dived deeply into Geminispace, but it's never held my interest. Like the author of that post, I can get all of what I need online with my web browser and RSS feed reader.
But, hey, if you enjoy using Gemini (as a reader or a creator or both), then stick with it!
IMHO the problem with Gemini is that "a subset of the web's functionality" is a very broad definition and it's hard to find consensus on whether that subset corresponds to 1% or 99% or something in the middle
I think the ideal project starts closer to the 99% but allows for graceful degradation down to 1% with little beyond a TUI, some colors and hyperlinks. That requires creating the right infrastructure for servers to offer a "stack" of content built something like (text + (hyperlinks + (images + (css + (video + (animations + (...)))). Clients can then be developed to handle whatever portion of that stack they want to.
This at least has the advantage of lowering switching costs for users who want to explore this "new web" while also allowing nerds like me to experience the web "the right way" with no distractions
However, the elephant in the room is that is is hard to imagine content creators would be in any way incentivized to think about their content in terms of that stack. Reddit just wants to force everyone to use their app, text-and-hyperlinks be damned. There's a reason RSS feeds have died.
> IMHO the problem with Gemini is that "a subset of the web's functionality" is a very broad definition and it's hard to find consensus on whether that subset corresponds to 1% or 99% or something in the middle
The problem is that it consist of below 1%. Markdown (let's say some fancy variants with tables) + hyperlinks + multimedia gives you a lot. Cut multimedia and now not even "look what I made" blog makes sense on it. Cut on basic formatting features. and you can't even make it nice to read
(Also from another recent HN discussion on Gemini I was left with the impression that the project's direction was defined in a pretty heavy-handed way, and I think its chilling effect has likely further diminished the chances Gemini has to succeed)
I do like Gemini but still question whether the development of a protocol was necessary. Couldn't you replace it with webrings and a crawler to index them?
Or create a subset of HTTP/HTML, maybe more like HTTP 0.9.
Also, it's trivial to create HTML subsets or custom vocabularies using SGML including Markdown syntax conversion - after all, that's what SGML was designed for.
Some of the Gemini goals such as ease of implementation, no user tracking or extensions, and a short specification (16 pages) are much easier with a new protocol.
Not a perfect solution but with webrings you could dissuade user tracking while promoting certain style guidelines. Sort of like a Home Ownership Association sans all the bullshit.
no. because one of the key aspects of it is that every client can interpret the content in whatever way the user prefers. Also the restrictions.
Gemini is _incredible_ for accessibility, because there's no disconnect between what's displayed and what a blind person can consume, and it's not complicated by layout things or developers who failed to account for the fact that maybe you need huge fonts.
Also, the moment you use normal HTML someone is going to stick an image in it. Now, i 100% support this idea. I've never embraced Gemini because the idea that the web is better with NO images is just absurd to me. I don't care how good of a writer, no textual description of something you've seen will come close to conveying what a photo can, and if you're trying to describe abstract art you're doomed to failure.
> Gemini is _incredible_ for accessibility
Incorrect. It is actually bad for accessibility.
Because gemtext was so deliberately hobbled in functionality, it is actively hostile to users with accessibility needs.
No tables means people have to do things with ASCII art which makes them incomprehensible to screen readers.
Standalone images (i.e. not online with the content) means you cannot provide any textual clues or context to what the image is about - you just get the image on its own. And the pervasive use of ASCII art "images" obviously is terrible.
No way to specify descriptive labels for links.
No way to visually-hide content so that it is only visible to screen readers.
No way to use aria-style tags to semantically mark up the different sections of a page (this is the navigation are, this is the gemlog listing etc)
A bit web-centric but I think important considering the ASCII table issue, no way to use JavaScript to provide what is known as "managed focus" handlers to allow navigation of cells/content areas using cursors+space+enter etc.
I can't remember, but I don't know if it even has in-page anchors?
They should have made gemtext slightly more sophisticated - inline images, proper tables, and basic visual styling would solve a lot of the a11y issues.
It is a shame. I got into Gemini a bit a while ago (even wrote a basic library for deno @ https://github.com/matt1/deno-gemini (essentially a fork of my own https://github.com/matt1/deno-gopher) and started work on a GUI client). I started to experiment with a totally-non-spec "style" spec that my client would support - essentially a line-by-line thing similar to CSS to apply the most basic of styles (colour, font weight, size etc) to gemtext lines, and which would be loaded from a well-known file (similar to robots.txt etc - compatible clients would load styles.gem or whatever).
But ultimately I was scared off by the short-comings of gemtext, and the staunch refusal to consider any changes or improvements. Which is fine, it is their spec and they don't have to change anything, but its why I left.
> Gemini is _incredible_ for accessibility
It depends on how you use it. If you want to make ASCII tables, then a screen reader will not have an idea of how to navigate the table. However a simple HTML <table> is well understood by screen-readers. I would contend that properly structured HTML documents are far better for accessibility.
Tables are not part of Gemini markup. What it offers is accessible. What it does not offer is really beside the point.
I disagree. What it doesn't offer, people will try to do anyway, each in their own ad-hoc way. See how people used to abuse <div> before HTML5 (and still do)
What it doesn't offer makes a significant chunk of content that people actually write and read inaccessible.
For example, the lack of simple text styling makes Gemini significantly less accessible than it could've been. Because now a screenreader will say something like "Significantly asterisk less asterisk accessible" instead of "Significantly emphasis less accessible".
I don't think images can have alt-text in Gemini so that blind users could get info from there.
Etc.
Edit: someone else has a much better answer: https://news.ycombinator.com/item?id=37052729
Gemini is incredible for accessibility because there is nothing of value on it so blind people aren't missing out in any way. Based on my attempts to use gemini I'm not even really being sarcastic.
The idea that accessibility means everyone's experience is nerfed into the ground just seems fundamentally wrong. We need to do much better than that and enable a richer experience that still is accessible. There's really no reason that can't be the case.
I appreciate visual artifacts that convey meaningful information, such as actual photographs, graphs, diagrams, and similar elements.
Which is probably a relatively small subset of images on the web.
> Gemini is _incredible_ for accessibility
An all new protocol stack focused on accessibility could be pretty cool. Flip the problem. Instead of layering accessibility on top, make it the foundation.
We might learn that an accessibility-focus stack proves better for all use cases. Kinda like accessible consumer goods are also better for every one.
Something fun to think about.
> no textual description of something you've seen will come close to conveying what a photo can
For strictly descriptive writing, maybe. But novels remain popular even in 2023 and while they mostly do not have illustrations of any kind, good ones paint some very vivid pictures.
> I've never embraced Gemini because the idea that the web is better with NO images is just absurd to me. I don't care how good of a writer, no textual description of something you've seen will come close to conveying what a photo can, and if you're trying to describe abstract art you're doomed to failure.
Gemini does not prevent anyone to put images. What it doesn't provide is the semantic separation between a link to an image and embedding that image into the page. The choice is up to the reader. Actually the default setting of the lagrange gemini browser is to download and show the images by default.
Gemini has images, just get an image viewer for a terminal Gemini client, or a graphical Gemini client which displays images inline, and there are several to choose. Lagrange it's pretty good.
If only I had an application that shipped with my system that did that already. One can dream I guess.
Amfora is the best console gemini client I've used. Recommended.
For Emacs users elpher.el is a good option. There are also some additional packages to help with authoring .gmi files either locally or over Titan.
https://thelambdalab.xyz/elpher/
https://git.carcosa.net/jmcbray/gemini.el
https://alexschroeder.ch/cgit/gemini-write/tree/README.md
You can find these in MELPA too.
Bombadillo it's better, it has basic HTTP/S support thru Lynx and it can... search inside the pages.
I have used both, just compared them to remind myself of them. Amfora reflows text so it's actually readable, bombadillo doesn't.
Bombadillo has settings to limit the width.
I loved playing around with Gemini. Something loosely along those lines can definitely work. Who knows what will become popular in the future. Extreme web page bloat leaves the door open.
Here are a couple of other alternative ideas for the web:
https://github.com/runvnc/tersenet (an idea, not implemented)
I tried to get interested in Gemini, I really did. The problem is that a lot of it is self-hosted, and there isn't really a great index or search engine that I've seen. That self-hosting means I'm locked out of participating for the time being, and the lack of search means it's hard for me to find anything of real value.
I want to see it grow. But right now, it's inferior in my eyes to even the current Gopher hole hosters online today.
There's gemlog.blue if you want to have a simple blog without hosting it yourself.
kennedy.gemi.dev and geminispace.info are both good search engines, though maybe not "really great".
I'm pretty sympathetic to this. I want to like the idea of Gemini, but it currently doesn't have, and is difficult to impossible to build, anything to really draw users away from the conventional web.
I think the fundamental idea of a new protocol that's much more limited than the current HTTP + HTML + JS web platform still has merit, but Gemini is just too limited to get going. It's tough though to design something that's flexible and feature-full enough to be really useful and not too hard to implement, but not so much that it will inevitably be extended to be basically what the web is now.
Thinking about it a bit, I think the idea for required TLS and client certs as the only way to ID users is good, as is lack of support for arbitrary HTTP headers and JS. I think what really kills a lot of potential functionality is the total lack of support for sending any data back to the server by anything other than basically query params. I think it ought to be possible for somebody to build a Super-Gemini that just support some equivalent of POST and some simple equivalent of HTML forms in addition to what it had, and should still be pretty easy to implement.
See the "titan" protocol.
https://transjovian.org/titan/page/The%20Titan%20Specificati...
It's "intended" to post a whole page/resource, but obviously it could be used to send an arbitrary blob just like an HTTP POST request.
Huh never mind before it actually works fine at home, must have been a pesky work network filter against it for some reason. I've never seen our work filters do that before though.
That site appears to be completely down, can't even find much of anything under archive.org.
That's actually pretty nice!
Most importantly for me is that nothing on HTTP protocol stops you from designing basic websites. A protocol is about the transport, not the content. This is the main fallacy about Gemini to me.
So, nothing requires you to first jump onto an obscure protocol in order to present websites that support text mode web browsers and minimalist reading experiences. HTML is great for this. It has a decade long history and text browsers are actively supported.
Then there is the protocol simplicity argument. HTTP is not a very complex protocol with great overhead as-is.
Rendering pages is not a HTTP hurdle or even a HTML hurdle but a CSS and JavaScript hurdle. That’s where the complexity arises and you walling yourself off basic renderers begin. But doing so is a choice, not anything coming from HTTP or HTML.
In fact, the situation has gotten _better_ over time with HTML 5 with improved semantic elements and deprecating styling elements. This makes it easier than before to understand and present a website in various forms.
Or it should do so. Now there is the topic of obnoxious websites burdened with JavaScript and heavy, complex CSS but since this argument is about people willing to use Gemini, they are not the kind of demography that would design obnoxious websites, but clean and well-behaved ones according to the spec. So using HTTP and HTML is thus not a problem for Gemini minded people. Or it shouldn’t be.
I think Gemini is mostly a psychological phenomenon. People were willing to do something for themselves that captures their ideology and intentionally walls them off because they wanted to more clearly feel part of a subculture. This is harder to achieve if you just subscribe to a minimalist, well-behaved blog service like Bear or Blot.im.
The one thing I’d love to have today is native Markdown support in web browsers that could have a custom style sheet applied to them. You’d enforce simplicity in a similar way as Gopher or Gemini but in the normal web = way more convenient and lowering barriers.
Gemini might go away, but Gopher is forever.
Actually both communities overlap. Also, Gopher sucks in any device smaller a DEC terminal. Gemini gives you methods to display your media on nearly any size of screen.
>both communities overlap.
The overlap is much smaller than gemini proponents would have you think.
E.g. I wouldn't touch gemini with a 15ft pole.
>Gemini gives you methods to display your media on nearly any size of screen.
But it still requires SSL and it still doesn't tell you the size of the resource you're fetching.
The Gemini protocol can't be fixed, as by design it cannot be modified.
Overall, Gemini just smells like an expression of the authors' narcissism.
Narcissism, no. Gopher sucks when you can't display it properly in the 99% of the phones out there. Gemini it's like gopher on features plus TLS and not bound to a DEC terminal. And I say this by daily using OpenBSD in a netbook with xterm/tmux as the main interface.
On TLS, you can get an RPI far cheaper than a second hand 486 machine back in the day in 2001. Starting from a Pentium MMX (better a Pentium II) you can TLS just fine. Heck, even an Amiga with a 68030 can do TLS with a bit of ptience (few seconds on handshake/deal).
And any competent hoster would provide both services for its site, as lots of Gemini users do, not leaving anyone behind.
For instance, gopher://midnight.pub and gemini://midnight.pub
I don't know, Gopher being removed from Firefox and Chrome over the last few years kinda disagrees with that sentiment.
You can still connect to Gopher server using "proper" Gopher clients, so the parent commenter is still right that Gopher is not going away. But then Gemini too is not going away. These niche protocols will always have their niche communities using their niche clients so they are probably going to survive even if they remain obscure.
FTP was even removed 2 years ago.
Really, look that up.
The fact that we'd have to look it up is why it was removed. I haven't used FTP in at least a decade—scp does the trick for my server management needs, HTTP is sufficient for download-only use cases, and Dropbox-style applications have supplanted FTP for shared files. FTP is more general than any of these three replacements, but it isn't ideal for any one of the use cases.
I only use X, YouTube, Insta, and TikTok. Browsers should just refuse to connect to any other site besides those
Web browsers should stick to web browsing and other tools should handle other protocols at this point in time.
I take it you have a use case for FTP still that's not yet been supplanted by other, more specialized tools? Please share, rather than snarking.
To name a few, I interact with a number of sites on shared hosting that have been using FTP for years. I also work with a data server that primarily interfaces with its data store through FTP. It also used to be easier to just host an FTP server and manage its access independent of a website, but still link to it in a browser where people who are not technically inclined could browse and interact with files without needing a separate dedicated client. Even for me, it would be simpler to just see an FTP file listing in a browser sometimes
The point is that browsers used to be able to browse the Internet, but major players such as Google have been working hard to limit all of these things to just HTTPS. Excusing things away by saying "well I don't use it, so it's useless" has only helped to narrow this path
Web browsers aren't internet browsers and shouldn't be. There's way too many things on the internet for them to do a good job anyway.
Exactly right. This is why we must continue to fight for our FAANG-only locked-down future!
Netscape used to come with a browser, a mail client, a newsreader, an HTML editor.
We've moved beyond that now, why worry about a browser being able to read Gopherspace or Gemini pages? Putting those into dedicated browsers of their own has been the perceived action for some time.
The reason is to connect them together in one experiential flow. That was the purpose of hypermedia, the interconnection of platforms and data sources.
People aren't doing that, I realize it, but maybe we should fix that instead of abandon the dream.
I’m not against it but I’ve browsed the web and the internet I guess on a NeXT and not seen that. I’ve seen far too many shitty user experiences. The best thing for anything on the net that’s not the web is to write a proxy and host it very prominently and encourage browsers to use it or make a simple extension to use it.
2023 me doesn’t want to install a client for WAIS or Gemini but I would do it if it made sense.
TBL's browser on the NeXT isn't a great reference point here.
Think more of hypercard and go back and visit the computer Chronicles episodes on YouTube for a refresh if you want
The difference between the two is TBL's WWW was good for browsing citation based research data that's spread across a network of global institutions while hypercard was more about information organization and presentation.
The distinction is subtle in theory but in practice it manifests as inherently different patterns.
For instance, the fidelity of a WWW link is coarse. It could be richer, but it's not. I'm familiar with the W3C semantics group btw and I've read their literature.
The modern web is neither of these systems but it's also something I believe most people would agree is lacking.
Perpetuating the suboptimal anarchy of the present by reinforcing it isn't going to help. More agency needs to be exercised.
I used the early web via Line Mode browser for almost seven years.
I don’t think there’s much to be had from a suboptimal experience.
If you want or need to use FTP you use FileZilla or Transmit.
All of those things center around the same set of technologies. Browsers today even come bundled with developer tools which still include HTML editors. But it's better for there to be two separate applications that both use separate copies of the same rendering engines and UI chrome? We might as well just make everything in Electron
Also, it's not like browsers are any less complicated now. They support things like 3D rendering and even MIDI device interfacing which are arguably more complicated for less benefit than supporting alternate protocols and technologies
Bah, gopher is just a lousy hytelnet wannabe.
I am going to give my cynical self a pat in the back. Back in March 2022 I was calling out Gemini as a fad that had no reason to exist. https://news.ycombinator.com/item?id=30802376
Someone losing interest in Gemini isn't really evidence that's true.
But the people using Gemini are still using flawed reasoning and there is still nothing in the protocol that can not be done with HTTP.
> and there is still nothing in the protocol that can not be done with HTTP.
This argument itself is using flawed reasoning. Gemini was purposely designed as a subset of HTTP rather than a superset, starting with a tough approximation of HTTP 0.9.
Of course everything that is supported in Gemini can also be implemented in HTTP.
The same arguments from that linked comment still apply and make sense to me but don't seem to hold water for you, so I won't rehash them.
µprotocols are fun. try text://protocol for that truly minimalistic avant-garde experience. [01]
There's a certain sense of irony in the fact that the same people who don't see the point of gemini "when HTTP exists" and "when there's nothing stopping you from writing a simple no-frills html page" also seem to be the same people who love 50 shades of markdown and bend over backwards to support it, yet would never dream of writing the equivalent in html unless it was as part of a framework.
I expected Gemtext not allowing emphasis as the true dealbreaker.
I am doing something using Gemtext, but not the Gemini protocol.
Basically - it's a decent way to "start over" and not be HTML-compatible, which is important if you are moving away from the owned-server model of hosting(which I am). It recognizes that it's not Markdown. The inline limitation is something I am probably going to breach at some point.
Gemini itself is something I admire, but like Gopher, it offers few frontiers for exploration.
what was this, just MySpace for 30-something software developers?
It's still around, but I think you got the gist.
WWW is a superset of all these small fun protocols, none of them offer anything the web isn't capable of, but they are great examples of sometimes what's fun is the limitations, not the features.
There is always a song: https://youtu.be/XRnjPSkVdt8
Gemini also doesn't really offer anything significantly better over gopher, imho, and only adds things that make it, ironically, less useable.
Dearth of content and politics aside, it is really nice on mobile.
Gemini was doomed from the start. It barely extended the functionality of Gopher, it was basically just Gopher + Markdown, and it wasn’t even the full .md. I’m pretty sure it was impossible to really make a good index of pages because of how limited it is. I think even Gopher has some search engines.
It was probably even intentionally limited to be just protocol for making a blog without any user input. No possibility of making something useful on it. It needed just the forms support, but AFAIK the developers didn’t want that.
Basic HTML/CSS (without JS) would be minimal enough and even more accessible. See img’s alt attribute. You can view it on lynx or Netsurf, there’s no need to use huge clients for that. And even better, it was the way already.
I don’t get why would someone make a new protocol except for filtering people out. Why won’t they just focus on making minimal web services? As long as there will be something that’s not just software related, like it is often with Gopher/Gemini.
> I don’t get why would someone make a new protocol except for filtering people out.
That actually appears to be one of the motivations. Which irks me out.
> I don't get why someone would make a new protocol except for filtering people out.
That could have been there intentions, I have idea, but if you think the web shouldn't have support for cookies or javascript, for example, then you really a new protocol. Even if you build a minimal web service, it still lives in a sandbox of third party cookiesand scripts, browser plugins, etc
Quite literally: the real Gemini was the friends we made all along