Replacing the World Wide Web

Why? - Web browsers have become far too complicated. The difficulty of implementing a sufficiently feature complete browser for daily use, is nearly insurmountable. They are also far too resource demanding. Problems with the existing alternative - Gemini has superseded Gopher, making the latter not even worth talking about. Gemini limits its use case to document sharing, the original intention for the WWW. However, its document format is more primitive than preexisting ones, like PDF. The most glaring omission, is no support for embedded images. This makes Gemini pages less info-dense than even print media. It's also not even a theoretically practical replacement for the WWW because it doesn't accommodate the most fundamental use cases of Web 2.0, what really made the WWW interesting. You cannot host a search engine, message board, or online encyclopedia, with Gemini. What's needed - A user interface markup language. This could be an extension of DSSSL. One document type with S-expressions, that could explicitly describe both what is displayed and how. If interaction with the UI changes the page, just re-render it. This is far simpler to implement than an HTML and CSS parser. Mathematical calculations, and the use-cases of HTMX, could also be supported by the engine, removing any practical need for JS. Identification. How can a site differentiate between users? Users would have a private key, and different public keys for each site(SLIP-10). A handshake process would be done for sensitive requests.

Finally... someone who says 'World Wide Web' and not 'The Internet.' Let's be realistic, at this point it's impossible to replace the Web—it's too mainstream for companies and individuals to want to change the status quo. It's simply not feasible. Think about it for a second: Even Tim Berners-Lee himself knows that the Web is completely ruined beyond repair. He is trying to replace the Web with his new project, ‘Solid,’ but... how many people are aware of this initiative? Exactly. It doesn't help that, unlike simply installing, e.g., Firefox and using it, to use Solid you have to go through a whole setup process. People are lazy and prefer things to be easy. There are many details that make replacing the Web nothing more than a fanciful and idealistic idea. Also consider this, suppose you manage to create an alternative that is superior AND more popular than the Web. Do you really want it to suffer the same fate as the Web in, say, 10 years? You know: Junk content with no value whatsoever, censorship imposed by despicable governments and monopolies, sites with no soul or personality at all, arrogant idiots who want to put "A.I." in everything, communities that engage in questionable activities... name anything you can think of. Sometimes I think it's best to just ignore the circus and focus on yourself. Appreciate the good communities that still exist and don't fall prey to the temptation of negative leisure. Sorry to be the voice of discord, but someone has to point out what could go wrong—it wouldn't be the first time something good ended up turning into a disaster.

>>90410 >A user interface markup language. This could be an extension of DSSSL. One document type with S-expressions, that could explicitly describe both what is displayed and how. I I like it but how hard to implement? You need a rendering engine. I don't think there's even one for html. I was looking into this, a minimal html renderer, but there isn't one. It's a bunch of crap. So?

>>90413 >He is trying to replace the Web with his new project, ‘Solid,’ but... how many people are aware of this initiative? I heard about it years ago, and never thereafter. On a conceptual level, it seems confused. >A user stores personal data in "pods" (personal online data stores) hosted wherever the user desires. >Before you use any of the Solid apps, you need to get a Solid personal online data store (Pod) first. Why do pods have to be online instead of stored locally? Why do I need to use a third party service to store my data? Why can't I self-host a pod? And what's to stop an application's server from making a copy of my data anyway? And look at the financial side >In 2015, MIT received a gift from Mastercard to support the development of Solid. >n 2018, Berners-Lee took a sabbatical from MIT to launch a commercial venture based on Solid a commercial venture based on Solid, named Inrupt. The company's mission is "to provide commercial energy and an ecosystem to help protect the integrity and quality of the new web built on Solid. It looks like a tenured professor's noncommittal business venture. It's no mystery why it hasn't gained any traction. >suppose you manage to create an alternative that is superior AND more popular than the Web The superior part is the only thing I'm interested in. >>90414 >You need a rendering engine. The parser translates the document into instructions for a 2d renderer, like Skia, which Chrome and LadyBird both use.

>>90415 >Skia nice >replacing the WWW I agree and I have some good ideas about this but there isn't a place online to talk about it

>>90415 >>The superior part is the only thing I'm interested in. Me too. Unfortunately, we live in a world where many people (with or without experience in software issues) believe that “popular” is synonymous with “efficient.” I assume you remember Brendan Eich (let's put biases against him aside for now—we need to be objective at the moment). Did you know that he originally wanted to implement Scheme in Netscape? I'm dead serious: https://exploringjs.com/es5/ch04.html >>In 1995, Netscape hired Brendan Eich with the promise of letting him implement Scheme (a Lisp dialect) in the browser. Before he could get started, Netscape collaborated with hardware and software company Sun (since bought by Oracle) to include its more static programming language, Java, in Navigator. https://2ality.com/2011/03/javascript-how-it-all-began.html >>The lure for him was that he would be able to base the language on Scheme (a Lisp dialect). Scheme’s influence led to JavaScript having closures. Another influence was the prototype-based programming language Self which is responsible for JavaScript’s prototypal inheritance And, for those who have used JavaScript, it is clear that Eich was heavily inspired by LISP. Unfortunately, it was the people with the resources (money) who decided that it was “better” to implement a new sister language to the popular (which is not synonymous with better) Java. Needless to say, the consequences are still evident today. I appreciate your insights on this matter.

>>90418 >I have some good ideas about this but there isn't a place online to talk about it This thread is a place to do it. >>90419 >Did you know that he originally wanted to implement Scheme in Netscape? Yes. There's no corporate mandates with this. Opting to use a new standard also creates a selection bias for people who aren't primarily concerned with popularity.

Why not just run a kind of mesh network that has exit nodes to the WWW when you need it? Any protocol change if it cant load up all the old websites is just throwing out all the good stuff with anything bad.

Gemini is the answer. >you cannot host >a search engine already exists (they're basic and you can't ask them questions like google but that's a backend issue, not a limitation of gemtext): gemini://kennedy.gemi.dev/ gemini://tlgs.one/ >message board already exists, gemini://bbs.geminispace.org >online encyclopedia wikipedia frontend: gemini://gemi.dev/cgi-bin/wp.cgi/ Gemini is the solution precisely because of its arbitrary limitations. You cannot grow gemini into such a monster like the WWW. Enshittifying such a limited document format and protocol (only 1 request header) is a lot of effort, and even if you managed to do it, the fact that gemini's hotspots are community-maintained means you'd just get delisted from every search engine and aggregator for being niggerlicious. Anything more complex than gemini that doesn't have a hardcore 'no we will not add any more featuresn' attitude is bound to morph into a mess like WWW when it reaches enough people. Also, no embedded images isn't even that bad, since a lot of images are superfluous (for example, all of the images in this thread) and not necessary for understanding the related text.

>>90443 I see the protocol allows sending text data. So it's up to each service to define how to separate information in this one block, instead of having a form which visually indicates this. And the client has to be flexible enough to allow multi-line editing. >a lot of images are superfluous And the ones which aren't? If you're on a website about statistics, you can't see graphs. Or a blog about anything involving visuals. An image booru is mostly comprised of images. The only work around is the client-specific feature of loading links to image files within a page. Also, like that anon said earlier about selectivity of users, do I want to select only for people with little to no appreciation for art? Not really. >bound to morph into a mess like WWW when it reaches enough people It became a mess because it's designed by a committee. Reaching enough people that it become of commercial interest, is also nearly impossible.

Gemini will become a protocol for blind autists no one else uses.

I feel like anything we do is going to get Jew'd as long as ZOG still stands

>>90445 >I see the protocol allows sending text data. The gemini protocol allows sending any kind of data. Konpeito was a seasonly mixtape project that operated over gemini which distributed mp3 files. Gemtext is simply the HTML analogue, so most clients will only support rendering gemtext or plaintext documents. You could absolutely send HTML or your DSSSL extension, you just have to respect the 'one request per document' policy or whatever it's actually called. >So it's up to each service to define how to separate information in this one block, instead of having a form which visually indicates this No, the gemtext standard defines the newline as a separator. So for every separate 'block' of text you make a newline. Yes, this means you have to write a whole paragraph in one line, yes it's a bit retarded but softwrap exists and it's the simplest way to separate text blocks. >>a lot of images are superfluous >And the ones which aren't? If you're on a website about statistics, you can't see graphs. Or a blog about anything involving visuals. An image booru is mostly comprised of images. Depends on the client, I've only used lagrange, where it simply loads the image inline when you click it. I do agree that hosting a booru or gallery on gemini would be fuarrrked since you can't organize images horizontally, only vertically. >The only work around is the client-specific feature of loading links to image files within a page. Some clients support this, like Lagrange. The default is that you have to click every image (one request per document policy), but you can make it load images on scroll. >do I want to select only for people with little to no appreciation for art? Not really. People in geminispace definitely appreciate art, and they do host photography and drawing pages on their sites. If you think the text focus means your site can't have an unique identity or whatever, you can use ASCII art to for visual identity, and after people start reading your writings they'll recognize you by your writing style too. I think most people who complain about gemini's limitations haven't actually given it a serious try. I've been lurking geminispace since the beginning and I haven't ever really felt like it's too limited for anything non-interactive.

My idea barely touches the topic, still why not throw it here. Pic related. DE-level browser-like wrapper for everything. More custom clients - the core. Boorus are all the same, IBs are all the same, etc. Also, with a totally URI-based OS, you can probably have some fun with links or object embedding, even in gemtext. Not sure about security.

>>90464 >You could absolutely send HTML or your DSSSL extension Sure, but I don't see any advantage to using Gemini over HTTP. You're throwing out all of that optimized, battle-tested infrastructure for no reason. The hard part of implementing a web browser, and what makes them consume so much resources, is not HTTP. It lacks features present even in FidoNet. No data compression is ridiculous. >No, the gemtext standard defines the newline as a separator. Look at the Lainchan form. The text inputs are name, email, subject, main content. With Gemtext, okay you can separate these by new line, but the user has to somehow know what to put on each one, or how to label them, right? It's like using a mainframe. >>90471 Reminds me a bit of Android.

>>90472 > but I don't see any advantage to using Gemini over HTTP Real talk, I think the main issue are search engines. Them and their results are the reason people hate the web now. If we could get more people to have their personal search engine, I think any protocol could succeed! The Darknet Lantern is a good example: http://lantern6vrwheap4iahuktz3jtwvovrzhxgwgzvrvp46brdzzekxibqd.onion HTTP is like kicking dead whales down the beach because it's really hard to make your own browser implementations. If it can't run in Plan 9, it's not good for preservation. Gemini and Gopher are great, but it would be better if they were more accessible. Better tooling and services would be great! My point is, maybe we should focus on the environment rather than the protocol itself (although I don't feel confident about Gemini because of the name expropiation by Google and the lack of spam prevention measures)

>>90410 Isn't this already happening? Almost nobody uses instagram's website for example. They use the app, just like with facebook, snapchat and even youtube. People use the browser less and less. Even banks and governments have their own apps. Of course most communication happens using HTTP, but it's not limited to it like a browser.

>>90410 I think a Python library would be better for your use case. Then you don't need a whole engine or browser, but simply a python interpreter running in a container. Will it take a bit longer to run? Sure, it's a bit slower than JS. This is compensated, however, by the brevity of your code - very few packets need to be transmitted when you can implement each document feature in only a few lines.

>>90413 >Sometimes I think it's best to just ignore the circus and focus on yourself. Appreciate the good communities that still exist and don't fall prey to the temptation of negative leisure. A lot of wisdom for SomeRandomJerk

Some further details I've decided on. A constraint-based layout will be used. Constraints will be solved by an implementation of Cassowary. https://en.wikipedia.org/wiki/Cassowary_(software) The name will be PantsuNet. This gets rid of any concern over commercialization. >>90476 Everything has Python bindings, so I'll probably go with that when I implement a client

>>90472 >>No, the gemtext standard defines the newline as a separator. >Look at the Lainchan form. The text inputs are name, email, subject, main content. With Gemtext, okay you can separate these by new line, but the user has to somehow know what to put on each one, or how to label them, right? It's like using a mainframe. gemini://bbs.geminispace.org has a feature where if you prepend the first line with # it sets that as the title (otherwise it leaves it untitled). E-mail and name fields are rarely used on chans and are vestigial at this point. E-mail is used for sage, which you could implement by checking if the first line of a reply is 'sage'. Of course, both of these need to be explained to the user, which you can do by having a 'help' page linked on your homepage (like the BBS does). Name field is only useful if you need to prove that your posts really are made by you, which gemini makes extremely easy to do securely because identities are managed through cryptographic certificates (kinda similar to your proposal), so there's no way to steal someone's name. And the certificates themselves require no identifying info to be provided. The name creating flow would look like this: >have 'create name' link on homepage >click 'create name' link >client asks you to create/provide certificate (1 click in Lagrange, and through the sidebar you can turn it off and back on anytime you want to be anon or named) >server detects a new identity, sends you a page with a link to set your name >click link, enter name >now whenever you have that identity turned on, the client will automatically send your certificate along with your post so you don't ever have to enter it again, and the server will attach your name to your posts automatically. This might sound like a chore but it's 3 clicks and typing your name, and then you can just turn your identity on and off willy-nilly, without ever needing to remember a username & password. Seriously, give gemini an actual try. The infrastructure, standards, and community is already there, so you just need to focus on making a site that you'd want to use (we've focused on discussing imageboard issues for now so I assume you want something like that). You'll get used to the limitations (and working around them) pretty quickly. P.S. >>90445 >I see the protocol allows sending text data. So it's up to each service to define how to separate information in this one block, instead of having a form which visually indicates this. And the client has to be flexible enough to allow multi-line editing. I misunderstood this and only now do I realise you're talking about an user sending stuff to the server, I thought you were talking about how gemtext recieved from the server decides where paragraphs end in the context of rendering text... >///< I'm actually out of my depth here, so your best bet is to ask the guy maintaining the gemini BBS, skyjake (jaakko.keranen@iki.fi), how he did it.

>>90474 I think browsers and websites could work a little like webapps where there is basic things needed on the user side what normally would be handled by a CDN calling on libraries etc but locally hosted and the website just calls on these with whatever permissions you allow it to do. So by default it would act like a browser with uMatrix and LocalCDN installed but more baked into the websites themselves where that is how they are intended to be handled so there is no compatibility issue where pages render wrong. This would potentially reduce bandwidth usage by a lot.

>>90499 >E-mail and name fields are rarely used on chans and are vestigial at this point. Whether or not they're useful, isn't the point. It's an arbitrary example. I'm surprised that the "form" even needs justification as a UI element. >You'll get used to the limitations (and working around them) pretty quickly. I'm sorry, but it's not interesting to me. If the point is to work within stringent limitations, I'd rather try to make a Wii homebrew service or something. At least that enables developers to make whatever UI they want(within the hardware's ability to render it). >>90473 >name expropiation by Google This is also kinda a deathblow to be honest.

We don't need to replace it, we just need better communities that aren't just circle jerking the "old web". The old web was new at the time and that's why it worked, still using the same imageboard format for interacting decades later is stupid. We need new ways of interacting and filtering users.

>>90507 >I'm surprised that the "form" even needs justification as a UI element To elaborate more on this, even Terminal applications can have a form inside of them. https://www.terminal.shop/ SSH apps lets you transmit any TUI over the network as a service. So Gopher and Gemini are a step backwards even from that. BBS software had more interesting interfaces than Gopher allowed for, to be accessible to any terminal. But this has nothing to do with the reality of the modern world. https://en.wikipedia.org/wiki/Door_(bulletin_board_system)

>>90499 >I'm actually out of my depth here, so your best bet is to ask the guy maintaining the gemini BBS, skyjake (jaakko.keranen@iki.fi), how he did it. I use it every now and then and the way multiline editing works is you end a post with a \ and that converts the post to a draft (or you click the link for creating a reply as a draft). The drafts page has "add text" input link that let's you add a new block of text, and for each block of text you've added, it has an actions link editing/reordering it and a delete link. It also supports the Titan protocol if you don't want to deal with the hassle of editing through gemini or if you want to upload an attachment. >>90509 >SSH BBSes These tend to have their own issues where writing posts is like kicking dead whales down the beach because you're often doing it inside a crappily implemented input form (shout out to the ones that at least let you edit your post in vim) and copy/pasting gets annoying. But the major downside to SSH BBSes is that you have to maintain a connection while you're browsing. If your connection is slow or spotty this can be very annoying, especially during post editing. With web browsers all the editing is down locally but when SSHed in, you're editing on the server so you have to deal with network overhead for every fuarrrking letter you type.

>>90514 >editing is down locally meant "editing is done locally" >Titan protocol I realized after posting that if you search for Titan protocol, it's hard to find the right one because of other projects with the same name. So here is the correct spec for curious lains: https://gmi.blakes.dev/titan-spec.gmi

Hot take: a protocol made to host any arbitrary interactive system will get niggified just like HTTP and HTML did. The Web has grown into this behemoth because it is treated like the omniplatform - you can host anything and everything in it. I see people complaining that alternatives like Gemini and Gopher aren't any good for hosting stuff like message boards, boorus, search engines, etc - but do they have to be? The Web should have stopped on 1.0. This is a scalding-hot take most of you aren't ready to hear, but bear with me for a few paragraphs. The moment people decided stuff should be interactive apps instead of documents and files was when the gates of Hell were opened. That is pretty much the entire reason Javascript exists. Even CGI originally was meant to create interfaces between internal databases of institutions and HTTP requests coming into their Web servers - even then "dynamic" page were still supposed to be documents. Now dynamic pages are used to create abominations like video streaming so cattle can consoom their goyslop, which even if you think this is something that should exist, you should at least agree it should be carried using a different protocol/interface. But why do they insist on servicing such things on the Browser? Because HTTP can serve any filetype, and HTML can be used to render any layout (thank you, W3C bloat!) and absolutely everybody has a standard Web browser, no matter what system/device they are using. Therefore it is a nobrainer that anything and everything should be developed using that as its interface for maximum compatibility with the masses. Even most phone apps these days are just wrappers for a Web browser. A follow-up to the above: Web servers really don't have to serve HTML if they don't want to. Despite the name, HTTP can transfer anything, so you could host gemtext and gophertext over HTTP. Obviously regular Web browsers won't understand it, but you could modify a current Gemini client to communicate with HTTP (making an usable subset of the protocol yourself can be done in a day) and display the gemtext. Likewise, sites that have nothing to do with transferring documents, like Youtube and Netflix, could still use HTTP to transfer data, but not rely on HTML to display it - for example if they had their own programs, or "desktop apps" if you want to be trendy, to retrieve and display videos. Again, why don't they do this? Because they would have to write a new interface, and it is much simpler and cheaper to abuse HTML instead. That and installing random companies' programs on your computer isn't nearly as normalized as it is on your phone. (The same argument can be made for message boards, boorus, collaborative wikis, and other types of website you autists may consider the "good parts" of Web 2.0. Message boards in particular can contain nothing except message data that you render however you want.) TL;DR: HTML is considered the universal interface, and it has become such an grotesque bloated mess that only Web browser titans can bear its burden. This is what gave the Web cancer.

>>90542 You haven't read the thread.

>>90544 You haven't read the post.

>>90410 You're not going to make anyone use a kneecapped version of what already exists, even if you forced them to use your standard they'd just expand your idea to the same garbage pile and ignore you, you're better off just treating it as the generic OS that it's used as in practice rather than create any specialized API that'll only be bloat in hindsight.

>>90410 >Gemini has superseded Gopher, making the latter not even worth talking about. What is the status on Gemini's encryption? Is TLS still mandatory? I host my stuff over Tor and I2P, so as long as they try to force me to encrypt my encrypted connections, I will see it as a silly goober protocol made by and for skiddies. Gopher+TLS already exists, and it is optional for those hosting in the clear.

Another option is to create a subset of what is good about the web and stick to it, preferable something that lightweight browsers can support just fine. That way you don’t need to create a browser or a new protocol, maybe just a parser that would sign your page as “clean”.

>>90757 TLS is still mandatory although there's a separate protocol whose name I can't recall that is essentially gemini minus TLS.

I am genuinely not trying to be an asshole, but... is anyone already creating a functional prototype ready to be tested from this whole bunch of ideas, or are you just role-playing that you're in a cyberpunk anime? Seriously, I've read the entire thread since I last commented on why trying to replace the World Wide Web is futile, but all I see is a bunch of well-meaning people daydreaming and philosophizing about why this or that is like kicking dead whales down the beach or doesn't suck. Only one or two lainons have dared to propose real alternatives, but from there on I only see scattered ideas but nothing real. So what's the point of this whole thread then? Repeat the obvious for the hundredth time? I'll say it again, I'm not trying to be a jerk just for the sake of it; if anyone has a working prototype of an experimental protocol, I would like to use a piece of equipment I have sitting there gathering dust as a node just to contribute to the cause. I want to test software, not play an ARG or whatever this is. I'm going to share this resource again so you can see that I genuinely want you to succeed: https://intronetworks.cs.luc.edu/ Also, I think more than one person is familiar with this famous tutorial, but it doesn't hurt to share it again: https://beej.us/guide/bgnet/ Now it's up to you whether you put it to good use or prefer to continue filling this thread with anarchist fantasies. Oh, and I implore you: Don't give me the excuse that it must remain secret or something like that. If you're afraid that the CIA or whatever will find you, then why are you even challenging the system in the first place?

>>90789 Everyone has their own goals. Just like not everyone want to "challenge the system" or whatever you decide to call that. You just put your expectations on other people, that's not good. Some people keep an eye on different projects. For example, I find the discussions around Gemini quite interesting. I do not feel like Gemini is a good solution to my needs, but it does some things right. And seeing other people's opinion helps me refine my understanding and concepts. Some come to just let their frustration out. And that's fine too and could even prove beneficial. You just have to ask the right questions, both them and yourself. Discussion is the mother of truth. Even if it all looks like boiling soykaf, it might help to come up with ideas.

>>90792 First, I genuinely thank you for expressing yourself respectfully without resorting to cheap condescension or irreverence, as is so common on the web nowadays (specially on these kind of sites). I am guilty of being what you would call a pragmatist. https://www.newworldencyclopedia.org/entry/Pragmatism https://plato.stanford.edu/entries/pragmatism/ (as with everything in Philosophy, there are various interpretations of this ideology, but I hope that this will help to form a general idea) For example, regarding this actually interesting phrase: >>Discussion is the mother of truth. Even if it all looks like boiling soykaf, it might help to come up with ideas. What I think is, what good are ideas if they remain just fantasies? Maybe it's the environment I grew up in, but just dreaming about something isn't enough to achieve it. To survive, you have to take action; facing things with sweet words is just complacency. However, it's wonderful that your point of view is more focused on empathy, unlike me, who is more... bluntly direct (not to say obnoxious, hence my alias). I hope you don't think I'm trying to be the villain of the moment. I genuinely would like to see a great project which came about as a result of this thread being made. If that weren't the case, I wouldn't even waste my time writing this. My father used to say: Soldiers die on their own. Best regards, and I wish you a wonderful life.

>>90789 OP here. I've become invested in another project(pic rel) and interest in this one seems limited. >>90560 In between that and work, I don't have time for this endeavour. As for that book, I'm only interested in the presentation layer.

>>90796 *application layer [Spoiler]Post too short[/spoiler]

>>90796 >> OP here. I've become invested in another project(pic rel) and interest in this one seems limited. Thank you for your reply. I hope I don't sound generic, but it's important to create projects that make you happy and make you feel satisfied with your work, not to think about what others need (even if that comes up as selfish and brazen). A fact that seems like a cliché from a Pixar movie, but is genuinely true, is that the key to success is actually enjoying what you do. A very common mistake, even among people with decades of experience, is to take on the burden of gigantic projects such as an operating system or the design of a new programming language (assuming it is not a hobby or recreational project) without having the time or interest to see it through. I hope your new project moves your soul in a positive way and that it is successful. Oh, and sorry if I came across as too harsh in my posts. Even if it's in an unorthodox way, my goal is to encourage people not to give up. I wish you all the best and success in whatever you set out to do.

>>90794 Thank you for the explanation. I respect your action-oriented approach to life, thought I cannot follow the same ideals. Currently time is the most valuable resource for me. So pursuing any endeavor, especially such a big one, should not be taken lightly. And in order to prove it worthy for myself, I need some prerequirements. The most important one is the clear goal. The things I want from this new hypothetical tool do not align into one cohesive vision. For example, my government places a lot of restrictions on stuff on the internet. Trying to fit into the legislation proves quite difficult, but I still prefer it this way. I try to not pick fights I cannot win and this opponent is too heavyweight for me. The second one is the required knowledge. I have to say I'm not good enough with network programming at the moment. I collect learning material, so that if and when I decide to try it, I have everything ready. All in all, this research for me is a more of a teatime activity between my projects, than a fully pledged development. I see it as trying to differentiate between the things I can do and the ones I cannot do. I prefer to not waste time on the latter. You may call it pointless dreaming and you may be right at that. But that's how I operate with my current understanding of life. And I have to say that while it pleases me to be called empathic, it's a conscious strive and not a genuine worldview. Regards.

>>90845 Hello. I'm hope you're doing awesome. I hope you are still here to read this. >> thought I cannot follow the same ideals. You don't have to. You must follow your own path, not the one from a random dude on the Web. I admire the fact that you remained firm in your position. >> Currently time is the most valuable resource for me. I know that many of those who frequent this site are young, and although I myself am not that old (I'm in my 30s), I am already at the stage where I realize that I have wasted too much time. I don't want the same thing to happen to you. I know I have nothing to gain, but I think it's a way to give meaning to my worthless life. >> And in order to prove it worthy for myself Genuinely, I believe you already are. A person who knows how to recognize their strengths and weaknesses has already taken a big step forward. >> my government places a lot of restrictions on stuff on the internet As always, politicians doing what they do best: annoying everyone else. Those bastards are nothing but a plague. I'm sorry you have to deal with that crap. >> The most important one is the clear goal. Life isn't linear, remember that. I started out shaping clay and never expected to end up writing code. >> The second one is the required knowledge. You will never stop learning. The subject of networks and the Internet is enormous, and that's without even considering computing in general. I even had to learn about electronics at some point. But don't worry, initiative is genuinely a good start. You're on the right track, just don't give up. >> you may be right at that. Probably not. I'm not anyone special, but I want you to be. >> it's a conscious strive and not a genuine worldview. That already makes you better than me. I mean it. I sincerely hope that you achieve your goals.

As a LISPer, I'm in favor of pervasive s-exprs (Less for me to implement!). The possibility of some agnostic form of transmission like EXI (to allow both s-expr and XML handlers, with the additional benefit of lower bandwidth) might be worth considering. I haven't looked too much into DSSSL and am unsure of how well-designed it is. A few seem to suggest that it shouldn't just be raw data and should have producer-side styling (rather than consumer-side styling as is often the case in Gemini browsers) in order to maintain the soul of small scale sites. Styling adds to the technical complexity of implementing such a system. I don't think styling would get in the way. Clients can simply ignore stylesheets. There's a consensus in the thread that it should be cleaner and easier to implement. For this to become a reality, it should be substantially so. Gemini succeeds on ease of implementation but, while it beats out Gopher, cleanliness remains suspect. Gemini is being stretched beyond its initial focus. The original aim explicitly rejected multimedia hypertext. Clients seem to blur this distinction, for better or for worse. These stretches mean that implementations of complex ideas are necessarily clunkier as they struggle against their tools. This is seen plainly with OP's dialogue where he's hashing out the image board implementation on Gemini. Should we simply content ourselves to a very minimal subset of the WWW's functionality or instead cut out the superfluous and add in some lost potential from the early hypertext movement (i.e., bidirectional links)?

>>90410 dunno if it relevant to this thread but a decentralization of social media is one of those needs. One good way would be the merge of webpages and social media feeds via p2p and webrings

Did this a bit ago. It's extremely bare bones. It just displays a window with some squares specified by s-expressions, and automatically resizes them when you resize the window. https://github.com/Dry-Leaf/pantsu-net

>>90410 I think this is a noble idea. However, the history of the WWW shows that anything we put together will naturally expand, if we are focused on replacing it. CSS/JS/etc all came about from non-standards bodies, just doing it because it was cool or requested. This was a natural side effect of the WWW being an unstructured graph of anything, which was the intended goal of it. Therefore, any protocol or presentation layer that is anything stricter than "anything goes, always" can never replace it from a taxonomical level. >ITT: Gemini good actually, no it's not, etc, etc. Gemini is a skiddie protocol that relies on a model of encryption that makes it impossible to grow, and is less useful than old TUIs. If it ever achieved large-scale adoption we'd be seeing TOFU certificate MITM's every day. So instead of replacing the WWW, let's make a system that has stronger protections from enshittification, but is still useful. This is distinct from Gemini's restrictions against anything that could be enshittified, which leads to it largely being useless. That with its standard, which actively conflicts with other standards it uses in technical details such as encodings. >>90488 >constraint-based layout This is the way. Protocol-wise, it would be reasonable to build a reduced form of HTTP/1.1 that does away with some of the user-extensibility and server-client state management, but maintains the simple HTTP structure. Things I'd propose removing: * The multiple different forms of allowed targets in a Request-Line. We should only allow an absolute path as a target, ensuring the host is specified through the "Host" field. This allows things like reverse-proxies to work without TLS-host verification or needing to fully parse a URI. * Remove the 7-bit ASCII encoding restriction within the Status-Line and Fields. I think the reasoning is obvious, we should enforce a modern encoding such as UTF-8. * Remove a majority of request methods. GET, HEAD, and POST/SEND are all that's actually needed. This avoids niche behaviors like OPTION and TRACE, "application" level behaviors like PUT/PATCH/DELETE, but still allows for interactivity and caching behaviors. * Remove a majority of response codes. Anybody who has implemented an HTTP server should know there's far too many, and most of them can be consolidated. * Possibly removing "keep-alive" connections as they were added as a response to using HTTP as an application protocol. This is fairly arbitrary however, as reusing connections is just an obvious efficiency step. * Restricting fields, instead of allowing any arbitrary one. This is where we truly separate from HTTP's "graph of everything", since we'd be enforcing that arbitrary standard extensions are not valid and should be rejected. This way, alongside the presentation-layer, we'd have a protocol that is possible to use for interactive content but causes friction when using it for APIs or more "application-oriented" systems. It would still be technically possible, of course, but the lack of support for it would be dissuasive at the very least.

I like the idea of Gemini and having a strictly text-oriented document protocol that precludes feature creep by design. It's a rather rebellious move in this day and age to focus on text content when the average person's attention span has been reduced to a new nadir of infantile retardation and nobody has the patience for anything that doesn't give them instant sensory overload with pictures and bright colors and loud noises.

I like the idea of Gemini's emphasis on text because it forces substance to be the forefront of a page rather than shallow soykaf like fonts, animations, colors. It's easier to make a website look nice than have anything worth reading on it, so everything has trended towards that. For a web to push back on this superficiality is more valuable to me than minimalism or stopping bloat. Luckily none of those are mutually exclusive

>>90410 https://arcan-fe.com/2026/01/26/arcan-explained-a-browser-for-different-webs/

>ctrl+f "reticulum" >0 results

>>92696 You're getting your OSI layers mixed up.

>>91558 > It's a rather rebellious move in this day and age to focus on text content when the average person's attention span has been reduced to a new nadir of infantile retardation and nobody has the patience for anything that doesn't give them instant sensory overload with pictures and bright colors and loud noises This is just not true. A lot of people are tired of overloaded screens with pop ups and have to contend with frozen up browsers. People have the attention span to watch documentaries on their phones and make long walls of text