10 ms·

Poor Man's Web

122 points by Looky1173 3y ago

danishbread 3y ago
The article title didn't express what this was about, but it seems seems to be a revival of the old, pre-WWW web, via a new protocol Gemini. I'm not sure why it's "poor man's" web, other than being minimalistic (text and images only.) I thought the phrase meant something less desirable to use, less capable. But this could be genuinely pleasant!
troupo 3y ago
People keep trying to make Gemini happen. I will keep repeating that in its quest for simplicity Gemini threw the baby out of with the water. Basic styling and inline images are invaluable.
- antiframe 3y ago
  I agree that basic styling and in-line images add something, but I like how Gemini strips so much faff out that the prose and links must stand strongly on their own. I've taken to writing my markdown and other documetns simiarly. Cory Doctorow does something similar. https://pluralistic.net/2023/12/20/em-oh-you-ess-ee/#sexytimes https://pluralistic.net/2023/12/20/em-oh-you-ess-ee/#sexytim... We don't need Gemini. We can get the feeling of the Smol Web by just choosing an effective subset of the features of HTML.
  - PurpleRamen 3y ago
    > I like how Gemini strips so much faff out that the prose and links must stand strongly on their own. I don't think that's working at all. Their website is so unapproachable bad, that it fails in selling me reasons why I should even care about this or read further. Letting something standing on its own only really works well if you have a small amount to deliver. Any slightly lengthy text will just bury you in a desert of letters.
    - II2II 3y ago
      > Any slightly lengthy text will just bury you in a desert of letters. Is this a bad thing? Books are a "desert of letters", with little to break the text up outside of chapters, sections, and paragraphs. If you broaden the scope a bit, you can added illustrations and photos. People have been reading books for generations. While many books do break that mould, many books continue to follow that tradition.
      - giantrobot 3y ago
        > Books are a "desert of letters", with little to break the text up outside of chapters, sections, and paragraphs. You're forgetting pages. The fact pages physically limit the visual bounds of all the letters helps people read them. The words from page 6 aren't going to come into view while you're reading page 4 but scrolled down a little too far.
      - PurpleRamen 3y ago
        > Is this a bad thing? Depends on the purpose. For a Webpage, which is a collection of short texts, to sell you on something, it is bad. > Books are a "desert of letters" Depends on the book. A phone book would be a desert of letters, I don't think many would enjoy reading them. Something like a novel, would be a forest of chapters, full of trees with letters arranged in a meaningful way, leading you on a road toward a goal. But a webpage is not a novel, it has an informative purpose, and is full of little small texts of equal value.
        skydhash 3y ago
        I use a rss client which has a content view, and it's barely different than what gemini offers. I also read plenty of epubs and while inline images are possibles, it's often looks really bad. I think there's a value on prioritizing content over forms. Some contents won't fit to these restrictions and that's ok. It's not like it's a web replacement, just an alternative whose restrictions create some kind of exclusivity.
      - troupo 3y ago
        Have you actually seen books? While typographic traditions did go down the drain in the recent decades, it's still hard to find a domain as varied as books. Even the most boring books often have things that Gemini purposefully omits: from styling to inline illustrations to diagrams to insets, asides, footnotes, tables of content, just tables, typographic marks etc. etc. etc.
  - lproven 3y ago
    Entirely this. I feel that if there were some way of promoting a Javascript-free web, everything the world needs is already there in all modern browsers. I suspect that between HTML5, and indeed XHTML, and CSS and all the many modern image formats and so on, everything important that almost any site needs could be done using these tools and no JS at all. And the result could also be interpreted and rendered successfully by much smaller simpler browsers, along the lines of Netsurf and Dillo, which are 10% of the size of a full dynamic-content browser or less. The question is how. A contest? Make the richest website you can that uses no Javascript, Typescript or anything else, and win a prize as well as promotion?
    - sircastor 3y ago
      I think a contest would be antithetical to the idea being circled here. The early web (to me) was filled with content front people who put out there because they wanted to share. Put your page up because here’s something you think is cool. I feel like it loses something with a contest.
      - lproven 3y ago
        I think you misinterpret my motivation here. What I want is less use of Javascript and related scripting languages across the web. I'd like more websites that don't need it. I am merely proposing a mechanism for getting people to explore what can be done without it, and how it might in fact be easier, more fun, and more maintainable.
    - giantrobot 3y ago
      I don't think a contest is needed. Just a desire to participate in the "Smol Web". Back in the day we had the "best viewed in any browser" badge indicating a page worked fine in IE and Navigator and probably also lynx. Personally I hate making web pages that don't work in Dillo or lynx.
      - lproven 3y ago
        I think, as per my comment above, you're misinterpreting me too. I am not trying to create a reason to make script-free websites. I think they are their own justification. I am trying to propose a method for making them visible and showing webdevs what can be done and why it's worth trying.
  - Gualdrapo 3y ago
    Not sure if going back to 2008 or so pre-HTML5 will prevent people from injecting kilobyes and megabytes of JS libraries or nesting dozens of divs because reasons.
- II2II 3y ago
  The problem with most of these discussions is they assume Gemini is a substitute for the web, when it is actually a substitute for Gopher. Gopher never really went away. A few enthusiasts were keeping it alive. Those enthusiasts realized that Gopher had a number of shortcomings, so Gemini was created to address those shortcomings. It was not created to address the shortcomings of the web. (At least not directly. Indirectly one could argue those enthusiasts kept Gopher alive due to the shortcomings of the web.) As for styling and inline images: in a sense, Gemini offers styles to a limited degree. Those styles are tied to the structure of the document, while the appearance is left to the software rendering the document. Even though inline images are considered a faux pas, I seem to recall Lagrange offering that feature. Again, we are dealing with the rendering software making the decision rather than the author. Since the end user chooses and configures the rendering software, it is the end user who has control (rather than the author).
  - classichasclass 3y ago
    I think this is true only up to a point. The sets of Gopher enthusiasts and Gemini enthusiasts surely overlap, but they are not coterminous (disclosure: I administer gopher.floodgap.com). From my view in Gopherspace, Gemini is a better fit for the Gopher+TLS thing people keep trying to do which is both incompatible and doesn't square with Gopher being an ultra-light protocol, and thus most appealing to those people who thought Gopher would be the "new smol web" but found it's more its own thing. In particular, Gopher's signature strong menu-document hierarchy that a lot of us Gopher nerds like doesn't have any true parallel in Gemini. I started gopher.floodgap.com back when it was gopher.ptloma.edu in the late 1990s largely as a historical preservation because I remembered all the cool stuff you could get there. Back then the Web hadn't metastasized to the Tetsuo blob it is today, so that obviously wouldn't have been the reason. I can't speak for the later adopters, but Gemini doesn't scratch my Gopher itches fully (see also https://oldvcr.blogspot.com/2020/11/a-gopher-view-of-gemini.html https://oldvcr.blogspot.com/2020/11/a-gopher-view-of-gemini.... , my notes on this from 2020).
    - II2II 3y ago
      I didn't mean to suggest that Gopher and Gemini enthusiasts are one and the same. Rather, I meant to suggest that it was a subset of Gopher enthusiasts that decided to address some of the shortcomings of Gopher. At least that is the impression that I received while watching from the sidelines. It never really struck me as being a derivative of the web. And thank you for gopher.floodgap.com. While I had some exposure to Gopher in the mid-1990's, most of my exposure was through Floodgap (and SDF) in the early 2000's.
- GuB-42 3y ago
  My feeling is that Gemini's entire purpose is being opinionated and somewhat inconvenient. It purposefully threw the baby out because it didn't want the baby. Everything Gemini does, you can do it in HTTP. In fact, a subset of HTTP+HTML would be even simpler than Gemini (mostly because of a lack of TLS) and still compatible with all the modern web stack. Simplicity isn't the main goal of Gemini. Exclusiveness is.
arkh 3y ago
Gemini protocol using :// again. I feel like more people should watch Dylan Beattle's "The Web that Never Was" presentation. > gem:my.domain.com/hello
- guestbest 3y ago
  Correct me if I’m wrong, but weren’t slashes typically reserved for flags on Unix and back slashes on dos/windows? If that’s the case then surely the vertices line, pipe symbol |, would be the most visually appropriate? Piping remote data over a network connection and through the local client program to parse seems to be the kind of thing bike shedding was intended for, in agreement but focusing on semantics. It just seems new by safe and only requires reprogramming the kernel for the shell semantics. # gemeni:|bike.domain.com|shed Or even better, everything is little endian to allow autocomplete from previous entries by the user in the shell so that once the user hits the domain keyword the shell history parser assumes the user wants to stay on the site # gemeni:|shed|bike|domain|com I’m not sure how feasible this as I’ve never written an internet or shell language
  - WaitWaitWha 3y ago
    > Correct me if I’m wrong, but weren’t slashes typically reserved for flags on Unix and back slashes on dos/windows? Forward slash (/) in DOS, while dash, tack, or hyphen (-) are used for UNIX for command line flags or parameters. The / in UNIX was for the path. The backslash (\) is for DOS paths. (Caveat: I started with UNIX in the mid-80s.)
    - guestbest 3y ago
      Ah, thanks.
- chatmasta 3y ago
  Maybe I'm mis-reading the URI spec [0], but it seems that // is required to separate the scheme from the "authority" part when it's present (as it is for my.domain.com). [0] https://datatracker.ietf.org/doc/html/rfc3986#section-3.2 https://datatracker.ietf.org/doc/html/rfc3986#section-3.2
spiritplumber 3y ago
Love Gemini but it really could use inline images.
- chwilson 3y ago
  Up to the client really, I know that Lagrange[1] supports this. [1]https://github.com/skyjake/lagrange https://github.com/skyjake/lagrange
  - giantrobot 3y ago
    Maybe Gemini could use some lightweight markup language. Shit it could be a form of hypertext. Maybe it could be called something like "hypertext markup language". Then clients could still decide how to render the page but there could be clear links between documents and sites. Nah, that probably will never work.
    - eep_social 3y ago
      Sounds like markdown, where do I sign!
- threatofrain 3y ago
  If Gemini really wanted to embrace a world of writing then they should've had the elements which HTML lacked, such as a table of contents or bibliography.
fabianholzer 3y ago
I am under the impression that an unstated design objective of gemini was preventing eternal september by trading it off for staying an eternal echo chamber tobe inhabited only by people smart enough to write their own client and server software.
- Kamq 3y ago
  Huh? There's plenty of ready clients and servers. It's limited to those that are primarily interested in text (and the occasional image). Which probably prevents it from ever going mainstream (and thus an eternal september). What it's generally sacrificing is interactivity. That seems like it's shooting more for the early internet. Which, yeah, has quite a bit of overlap with people who can code, but not as much as you're implying. And I do remember quite a few echo chambers on the early internet, but I remember a lot less of them without algorithmic feeds. Just by chance, you'd end up running across a bunch of people who disagree with you on a lot of things.
  - fabianholzer 3y ago
    > There's plenty of ready clients and servers. Sure, but being smart enough to implement one usually also means being smart enough to use something of the shelf. But few of the available clients are very accessible to a non-programming audience. It is basically a "natural" filter mechanism. > What it's generally sacrificing is interactivity. That seems like it's shooting more for the early internet. There is a very rudimentary form mechanism, resembling a bit HTTP 0.9, where you could only do get requests with query parameters, which might allow to simulate a bit of interactivity, but I with the protocol being set in stone that would forever remain a hacky simulation (if you only have a hammer...)
    - doublepg23 3y ago
      > But few of the available clients are very accessible to a non-programming audience. It is basically a "natural" filter mechanism. ? There’s a very nice one that runs fine on my iPad, Elaho.
surprisetalk 3y ago
I recently documented some similar thoughts on the “cheap” web: [1] https://potato.cheap https://potato.cheap IMO, protocol is largely a distraction from the good stuff. The question I’ve been wondering lately is, “why don’t most people I meet have personal websites?”
- basscomm 3y ago
  I can think of three reasons that more people don't have personal websites: 1. For non technical users (and there are lots of them) learning enough HTML and figuring out how to put it somewhere accessible is too high of a bar to overcome (forget about updating it ever). It's quicker and easier to just establish a presence on some social media site. 2. For technical users who are capable of setting up a website, it's easier to just go where people already are (i.e. social media). 3. It's anecdotal, but I rarely see anyone browsing the web these days. Most web usage seems to consist of endlessly scrolling through Facebook looking for something to interact with. A link to somewhere else might get clicked on, but the user always goes back to the newsfeed once they're done looking at the link.
  - ugjka 3y ago
    If you host things on a cheap VPS you also have to pay attention how much traffic you are allowed to consume and excess will add up to your bill. Some bad bot or enemy can certainly give a nice surprise at the end of the month if you don't pay attention to alerts on your account
  - ravenstine 3y ago
    > 3. It's anecdotal, but I rarely see anyone browsing the web these days. Exactly. People on HN overestimate today's use of the web. Regular people barely know what a website is at this point or have any idea what a browser is or does. They only know it as a place to ask The Google something, but even that has changed with The Google's app/widget and voice activated assistants. Sure those things open up web views, but that view may not even have a URL bar, so as far as the user knows, they're in "the app", which is kind of true I guess. That "Chrome" or "Safari" app is just kind of a funny thing that opens up sometimes. Most users aren't opening up the browser and typing in URLs. A fraction of them are, but that fraction is diminishing, and I would guess that fraction mostly reflects desktop users. All this is to say that, if you want to create a blog today and not spend effort to heavily promote it and use SEO tricks (that will immediately go out of date), you can basically forget it if you want anyone to read it.
JohnMakin 3y ago
I like the gist and spirit of this, of course, but isn’t this basically impossible in today’s day and age? For instance, if I want to host a small web server running in my closet to serve cat memes to a small number of users, I still am beholden to my ISP who 1) must allow me to host a web server in its TOS and 2) provide me a static address. I don’t really see any way to get around corporations on the internet, unless I’m missing something.
- anamexis 3y ago
  I think that depends on how robust you want to be. I don't have a static IP, but it only changes on the order of weeks or months. So a dynamic DNS configuration with a 5 minute TTL works fine for me. I have no idea if web hosting is in my ISP's TOS, and I also don't care.
- 0xEF 3y ago
  I don't know about elsewhere, but I regularly set up small self-hosted web servers at my home (Midwest US) and have never heard a peep from my ISP about it. Some of the aforementioned servers get 25 - 50 unique visits on a high day, and my ISP couldn't care less. I'm not sure they would unless I was calling a lot of traffic, which is not really a huge concern for Small Webbers. We are few and relatively unnoticed by the larger Internet, which is just the way we like it.
- torh 3y ago
  1) Haven't really checked, but... 2) I have a fixed IPv4 address, and a dynamic IPv6. So far the latter also seems to be fairly stable, and I host my webpages on both IPv4 and IPv6 these days. My server is a ThinkPad X240 with VMware ESXi which in turn host Ubuntu for SSH, Web, etc... And a Pi-hole VM just to block the "wost" of the internet.
- jw_cook 3y ago
  ISPs that allow hosting web services are fairly common, but you're right, if all your local providers forbid it, you're probably out of luck. For your second point, Dynamic DNS is super useful. I've had a good experience with noip.com/ddns.net, but there are plenty of other options out there.
nonrandomstring 3y ago
Nice little article but I could hardly disagree more with the headline. The small web is not for poor men and women. It's for the rich. Those who are rich in culture, intelligence, curiosity, engagement... The existing "web" is for the technologically poor. It's a chavvy, shallow plastic shit-filled trench, a ghetto where thieves and pimps run free, and good ideas die like dogs - to somewhat paraphrase Hunter S. Thompson. I can't see anything of value left there, and cannot imagine how anyone would equate leaving it behind with being poorer.
aleph_minus_one 3y ago
Relevant concerning the given example server typed-hole.org:79: $ echo "feed" | nc typed-hole.org 79 julienxx@nein.club toots: thread 'main' panicked at src/main.rs:52:39: called `Result::unwrap()` on an `Err` value: reqwest::Error { kind: Request, url: Url { scheme: "https", cannot_be_a_base: false, username: "", password: None, host: Some(Domain("nein.club")), port: None, path: "/api/v1/accounts/01C2ETC44GDGZ27S9WKZMSHKWV/statuses", query: Some("limit=5&exclude_reblogs=true&exclude_replies=true"), fragment: None }, source: hyper::Error(Connect, Ssl(Error { code: ErrorCode(1), cause: Some(Ssl(ErrorStack([Error { code: 167772294, library: "SSL routines", function: "tls_post_process_server_certificate", reason: "certificate verify failed", file: "../ssl/statem/statem_clnt.c", line: 1889 }]))) }, X509VerifyResult { code: 18, error: "self-signed certificate" })) } note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace Looks like an error.
mattlondon 3y ago
I think Gemini has missed a trick being being deliberately too simple - e.g. no inline images, no tables, no forms, no basic formatting etc. I know these were deliberately left out for "simplicity" reasons, but then they have a fairly unsimple mechanism for user login/auth that totally undermines their claim for making clients little more than slightly-augmented-terminals, which for me is a smoking gun for these things being missing purely on an ideological basis only and an attempt at controlling how people use it. That's fine, it's their thing and they can do what they want with it. There is a growing trend of people simply choosing to use XHTML Basic (1). This is a "stripped down" version of HTML originally intended for early phones, PDAs, set top boxes etc - it has the critical missing features of Gemini, but forgoes some of the more "advanced" features of the modern web people have issues with. I'd highly recommend people simply target that instead. 1 - https://en.m.wikipedia.org/wiki/XHTML_Basic https://en.m.wikipedia.org/wiki/XHTML_Basic
- krapp 3y ago
  It looks like this still supports the script tag, which seems to be the biggest "modern web" feature people want to abandon.
  - giantrobot 3y ago
    So don't include one. JavaScript isn't fucking mandatory.
    - krapp 3y ago
      The kind of people who tend to want a separate, simpler alternative web are almost universally also the kind of people who are so allergic to javascript that they don't even want it to be an option, on the remote chance they might encounter it in the wild. They're drawn by a combination of nostalgia and spite to manifest an alternate universe where scripting on the web never even happened. Otherwise, sure, they could just write simple HTML and no javascript on the web we have now. I personally would like to retry a version of the script tag (possibly more like the object tag) with native support for WASM that would make it easy to run scripts in any language... which is closer to the original vision before Javascript "won," and possible without requiring browsers to support the languages. Also maybe things like an <include> tag for HTML, and other things that fell by the wayside of early HTML. I'd like to see something like HolyC, with native support for text, hypertext, drawing and code. But 99.9% of people seem to just want a web that's basically whitepapers with links, maybe images.
chrsw 3y ago
I think IRC would fit in here too
1vuio0pswjnm7 3y ago
(2021)

122 points by Looky1173 3y ago

The article title didn't express what this was about, but it seems seems to be a revival of the old, pre-WWW web, via a new protocol Gemini. I'm not sure why it's "poor man's" web, other than being minimalistic (text and images only.) I thought the phrase meant something less desirable to use, less capable. But this could be genuinely pleasant!

troupo 3y ago

People keep trying to make Gemini happen. I will keep repeating that in its quest for simplicity Gemini threw the baby out of with the water. Basic styling and inline images are invaluable.

antiframe 3y ago
I agree that basic styling and in-line images add something, but I like how Gemini strips so much faff out that the prose and links must stand strongly on their own. I've taken to writing my markdown and other documetns simiarly. Cory Doctorow does something similar. https://pluralistic.net/2023/12/20/em-oh-you-ess-ee/#sexytimes https://pluralistic.net/2023/12/20/em-oh-you-ess-ee/#sexytim... We don't need Gemini. We can get the feeling of the Smol Web by just choosing an effective subset of the features of HTML.
- PurpleRamen 3y ago
  > I like how Gemini strips so much faff out that the prose and links must stand strongly on their own. I don't think that's working at all. Their website is so unapproachable bad, that it fails in selling me reasons why I should even care about this or read further. Letting something standing on its own only really works well if you have a small amount to deliver. Any slightly lengthy text will just bury you in a desert of letters.
  - II2II 3y ago
    > Any slightly lengthy text will just bury you in a desert of letters. Is this a bad thing? Books are a "desert of letters", with little to break the text up outside of chapters, sections, and paragraphs. If you broaden the scope a bit, you can added illustrations and photos. People have been reading books for generations. While many books do break that mould, many books continue to follow that tradition.
    - giantrobot 3y ago
      > Books are a "desert of letters", with little to break the text up outside of chapters, sections, and paragraphs. You're forgetting pages. The fact pages physically limit the visual bounds of all the letters helps people read them. The words from page 6 aren't going to come into view while you're reading page 4 but scrolled down a little too far.
    - PurpleRamen 3y ago
      > Is this a bad thing? Depends on the purpose. For a Webpage, which is a collection of short texts, to sell you on something, it is bad. > Books are a "desert of letters" Depends on the book. A phone book would be a desert of letters, I don't think many would enjoy reading them. Something like a novel, would be a forest of chapters, full of trees with letters arranged in a meaningful way, leading you on a road toward a goal. But a webpage is not a novel, it has an informative purpose, and is full of little small texts of equal value.
      - skydhash 3y ago
        I use a rss client which has a content view, and it's barely different than what gemini offers. I also read plenty of epubs and while inline images are possibles, it's often looks really bad. I think there's a value on prioritizing content over forms. Some contents won't fit to these restrictions and that's ok. It's not like it's a web replacement, just an alternative whose restrictions create some kind of exclusivity.
    - troupo 3y ago
      Have you actually seen books? While typographic traditions did go down the drain in the recent decades, it's still hard to find a domain as varied as books. Even the most boring books often have things that Gemini purposefully omits: from styling to inline illustrations to diagrams to insets, asides, footnotes, tables of content, just tables, typographic marks etc. etc. etc.
- lproven 3y ago
  Entirely this. I feel that if there were some way of promoting a Javascript-free web, everything the world needs is already there in all modern browsers. I suspect that between HTML5, and indeed XHTML, and CSS and all the many modern image formats and so on, everything important that almost any site needs could be done using these tools and no JS at all. And the result could also be interpreted and rendered successfully by much smaller simpler browsers, along the lines of Netsurf and Dillo, which are 10% of the size of a full dynamic-content browser or less. The question is how. A contest? Make the richest website you can that uses no Javascript, Typescript or anything else, and win a prize as well as promotion?
  - sircastor 3y ago
    I think a contest would be antithetical to the idea being circled here. The early web (to me) was filled with content front people who put out there because they wanted to share. Put your page up because here’s something you think is cool. I feel like it loses something with a contest.
    - lproven 3y ago
      I think you misinterpret my motivation here. What I want is less use of Javascript and related scripting languages across the web. I'd like more websites that don't need it. I am merely proposing a mechanism for getting people to explore what can be done without it, and how it might in fact be easier, more fun, and more maintainable.
  - giantrobot 3y ago
    I don't think a contest is needed. Just a desire to participate in the "Smol Web". Back in the day we had the "best viewed in any browser" badge indicating a page worked fine in IE and Navigator and probably also lynx. Personally I hate making web pages that don't work in Dillo or lynx.
    - lproven 3y ago
      I think, as per my comment above, you're misinterpreting me too. I am not trying to create a reason to make script-free websites. I think they are their own justification. I am trying to propose a method for making them visible and showing webdevs what can be done and why it's worth trying.
- Gualdrapo 3y ago
  Not sure if going back to 2008 or so pre-HTML5 will prevent people from injecting kilobyes and megabytes of JS libraries or nesting dozens of divs because reasons.
II2II 3y ago
The problem with most of these discussions is they assume Gemini is a substitute for the web, when it is actually a substitute for Gopher. Gopher never really went away. A few enthusiasts were keeping it alive. Those enthusiasts realized that Gopher had a number of shortcomings, so Gemini was created to address those shortcomings. It was not created to address the shortcomings of the web. (At least not directly. Indirectly one could argue those enthusiasts kept Gopher alive due to the shortcomings of the web.) As for styling and inline images: in a sense, Gemini offers styles to a limited degree. Those styles are tied to the structure of the document, while the appearance is left to the software rendering the document. Even though inline images are considered a faux pas, I seem to recall Lagrange offering that feature. Again, we are dealing with the rendering software making the decision rather than the author. Since the end user chooses and configures the rendering software, it is the end user who has control (rather than the author).
- classichasclass 3y ago
  I think this is true only up to a point. The sets of Gopher enthusiasts and Gemini enthusiasts surely overlap, but they are not coterminous (disclosure: I administer gopher.floodgap.com). From my view in Gopherspace, Gemini is a better fit for the Gopher+TLS thing people keep trying to do which is both incompatible and doesn't square with Gopher being an ultra-light protocol, and thus most appealing to those people who thought Gopher would be the "new smol web" but found it's more its own thing. In particular, Gopher's signature strong menu-document hierarchy that a lot of us Gopher nerds like doesn't have any true parallel in Gemini. I started gopher.floodgap.com back when it was gopher.ptloma.edu in the late 1990s largely as a historical preservation because I remembered all the cool stuff you could get there. Back then the Web hadn't metastasized to the Tetsuo blob it is today, so that obviously wouldn't have been the reason. I can't speak for the later adopters, but Gemini doesn't scratch my Gopher itches fully (see also https://oldvcr.blogspot.com/2020/11/a-gopher-view-of-gemini.html https://oldvcr.blogspot.com/2020/11/a-gopher-view-of-gemini.... , my notes on this from 2020).
  - II2II 3y ago
    I didn't mean to suggest that Gopher and Gemini enthusiasts are one and the same. Rather, I meant to suggest that it was a subset of Gopher enthusiasts that decided to address some of the shortcomings of Gopher. At least that is the impression that I received while watching from the sidelines. It never really struck me as being a derivative of the web. And thank you for gopher.floodgap.com. While I had some exposure to Gopher in the mid-1990's, most of my exposure was through Floodgap (and SDF) in the early 2000's.
GuB-42 3y ago
My feeling is that Gemini's entire purpose is being opinionated and somewhat inconvenient. It purposefully threw the baby out because it didn't want the baby. Everything Gemini does, you can do it in HTTP. In fact, a subset of HTTP+HTML would be even simpler than Gemini (mostly because of a lack of TLS) and still compatible with all the modern web stack. Simplicity isn't the main goal of Gemini. Exclusiveness is.

arkh 3y ago

Gemini protocol using :// again. I feel like more people should watch Dylan Beattle's "The Web that Never Was" presentation. > gem:my.domain.com/hello

guestbest 3y ago
Correct me if I’m wrong, but weren’t slashes typically reserved for flags on Unix and back slashes on dos/windows? If that’s the case then surely the vertices line, pipe symbol |, would be the most visually appropriate? Piping remote data over a network connection and through the local client program to parse seems to be the kind of thing bike shedding was intended for, in agreement but focusing on semantics. It just seems new by safe and only requires reprogramming the kernel for the shell semantics. # gemeni:|bike.domain.com|shed Or even better, everything is little endian to allow autocomplete from previous entries by the user in the shell so that once the user hits the domain keyword the shell history parser assumes the user wants to stay on the site # gemeni:|shed|bike|domain|com I’m not sure how feasible this as I’ve never written an internet or shell language
- WaitWaitWha 3y ago
  > Correct me if I’m wrong, but weren’t slashes typically reserved for flags on Unix and back slashes on dos/windows? Forward slash (/) in DOS, while dash, tack, or hyphen (-) are used for UNIX for command line flags or parameters. The / in UNIX was for the path. The backslash (\) is for DOS paths. (Caveat: I started with UNIX in the mid-80s.)
  - guestbest 3y ago
    Ah, thanks.
chatmasta 3y ago
Maybe I'm mis-reading the URI spec [0], but it seems that // is required to separate the scheme from the "authority" part when it's present (as it is for my.domain.com). [0] https://datatracker.ietf.org/doc/html/rfc3986#section-3.2 https://datatracker.ietf.org/doc/html/rfc3986#section-3.2

spiritplumber 3y ago

Love Gemini but it really could use inline images.

chwilson 3y ago
Up to the client really, I know that Lagrange[1] supports this. [1]https://github.com/skyjake/lagrange https://github.com/skyjake/lagrange
- giantrobot 3y ago
  Maybe Gemini could use some lightweight markup language. Shit it could be a form of hypertext. Maybe it could be called something like "hypertext markup language". Then clients could still decide how to render the page but there could be clear links between documents and sites. Nah, that probably will never work.
  - eep_social 3y ago
    Sounds like markdown, where do I sign!
threatofrain 3y ago
If Gemini really wanted to embrace a world of writing then they should've had the elements which HTML lacked, such as a table of contents or bibliography.

fabianholzer 3y ago

I am under the impression that an unstated design objective of gemini was preventing eternal september by trading it off for staying an eternal echo chamber tobe inhabited only by people smart enough to write their own client and server software.

Kamq 3y ago
Huh? There's plenty of ready clients and servers. It's limited to those that are primarily interested in text (and the occasional image). Which probably prevents it from ever going mainstream (and thus an eternal september). What it's generally sacrificing is interactivity. That seems like it's shooting more for the early internet. Which, yeah, has quite a bit of overlap with people who can code, but not as much as you're implying. And I do remember quite a few echo chambers on the early internet, but I remember a lot less of them without algorithmic feeds. Just by chance, you'd end up running across a bunch of people who disagree with you on a lot of things.
- fabianholzer 3y ago
  > There's plenty of ready clients and servers. Sure, but being smart enough to implement one usually also means being smart enough to use something of the shelf. But few of the available clients are very accessible to a non-programming audience. It is basically a "natural" filter mechanism. > What it's generally sacrificing is interactivity. That seems like it's shooting more for the early internet. There is a very rudimentary form mechanism, resembling a bit HTTP 0.9, where you could only do get requests with query parameters, which might allow to simulate a bit of interactivity, but I with the protocol being set in stone that would forever remain a hacky simulation (if you only have a hammer...)
  - doublepg23 3y ago
    > But few of the available clients are very accessible to a non-programming audience. It is basically a "natural" filter mechanism. ? There’s a very nice one that runs fine on my iPad, Elaho.

surprisetalk 3y ago

I recently documented some similar thoughts on the “cheap” web: [1] https://potato.cheap https://potato.cheap IMO, protocol is largely a distraction from the good stuff. The question I’ve been wondering lately is, “why don’t most people I meet have personal websites?”

basscomm 3y ago
I can think of three reasons that more people don't have personal websites: 1. For non technical users (and there are lots of them) learning enough HTML and figuring out how to put it somewhere accessible is too high of a bar to overcome (forget about updating it ever). It's quicker and easier to just establish a presence on some social media site. 2. For technical users who are capable of setting up a website, it's easier to just go where people already are (i.e. social media). 3. It's anecdotal, but I rarely see anyone browsing the web these days. Most web usage seems to consist of endlessly scrolling through Facebook looking for something to interact with. A link to somewhere else might get clicked on, but the user always goes back to the newsfeed once they're done looking at the link.
- ugjka 3y ago
  If you host things on a cheap VPS you also have to pay attention how much traffic you are allowed to consume and excess will add up to your bill. Some bad bot or enemy can certainly give a nice surprise at the end of the month if you don't pay attention to alerts on your account
- ravenstine 3y ago
  > 3. It's anecdotal, but I rarely see anyone browsing the web these days. Exactly. People on HN overestimate today's use of the web. Regular people barely know what a website is at this point or have any idea what a browser is or does. They only know it as a place to ask The Google something, but even that has changed with The Google's app/widget and voice activated assistants. Sure those things open up web views, but that view may not even have a URL bar, so as far as the user knows, they're in "the app", which is kind of true I guess. That "Chrome" or "Safari" app is just kind of a funny thing that opens up sometimes. Most users aren't opening up the browser and typing in URLs. A fraction of them are, but that fraction is diminishing, and I would guess that fraction mostly reflects desktop users. All this is to say that, if you want to create a blog today and not spend effort to heavily promote it and use SEO tricks (that will immediately go out of date), you can basically forget it if you want anyone to read it.

JohnMakin 3y ago

I like the gist and spirit of this, of course, but isn’t this basically impossible in today’s day and age? For instance, if I want to host a small web server running in my closet to serve cat memes to a small number of users, I still am beholden to my ISP who 1) must allow me to host a web server in its TOS and 2) provide me a static address. I don’t really see any way to get around corporations on the internet, unless I’m missing something.

anamexis 3y ago
I think that depends on how robust you want to be. I don't have a static IP, but it only changes on the order of weeks or months. So a dynamic DNS configuration with a 5 minute TTL works fine for me. I have no idea if web hosting is in my ISP's TOS, and I also don't care.
0xEF 3y ago
I don't know about elsewhere, but I regularly set up small self-hosted web servers at my home (Midwest US) and have never heard a peep from my ISP about it. Some of the aforementioned servers get 25 - 50 unique visits on a high day, and my ISP couldn't care less. I'm not sure they would unless I was calling a lot of traffic, which is not really a huge concern for Small Webbers. We are few and relatively unnoticed by the larger Internet, which is just the way we like it.
torh 3y ago
1) Haven't really checked, but... 2) I have a fixed IPv4 address, and a dynamic IPv6. So far the latter also seems to be fairly stable, and I host my webpages on both IPv4 and IPv6 these days. My server is a ThinkPad X240 with VMware ESXi which in turn host Ubuntu for SSH, Web, etc... And a Pi-hole VM just to block the "wost" of the internet.
jw_cook 3y ago
ISPs that allow hosting web services are fairly common, but you're right, if all your local providers forbid it, you're probably out of luck. For your second point, Dynamic DNS is super useful. I've had a good experience with noip.com/ddns.net, but there are plenty of other options out there.

nonrandomstring 3y ago

Nice little article but I could hardly disagree more with the headline. The small web is not for poor men and women. It's for the rich. Those who are rich in culture, intelligence, curiosity, engagement... The existing "web" is for the technologically poor. It's a chavvy, shallow plastic shit-filled trench, a ghetto where thieves and pimps run free, and good ideas die like dogs - to somewhat paraphrase Hunter S. Thompson. I can't see anything of value left there, and cannot imagine how anyone would equate leaving it behind with being poorer.

aleph_minus_one 3y ago

Relevant concerning the given example server typed-hole.org:79: $ echo "feed" | nc typed-hole.org 79 julienxx@nein.club toots: thread 'main' panicked at src/main.rs:52:39: called `Result::unwrap()` on an `Err` value: reqwest::Error { kind: Request, url: Url { scheme: "https", cannot_be_a_base: false, username: "", password: None, host: Some(Domain("nein.club")), port: None, path: "/api/v1/accounts/01C2ETC44GDGZ27S9WKZMSHKWV/statuses", query: Some("limit=5&exclude_reblogs=true&exclude_replies=true"), fragment: None }, source: hyper::Error(Connect, Ssl(Error { code: ErrorCode(1), cause: Some(Ssl(ErrorStack([Error { code: 167772294, library: "SSL routines", function: "tls_post_process_server_certificate", reason: "certificate verify failed", file: "../ssl/statem/statem_clnt.c", line: 1889 }]))) }, X509VerifyResult { code: 18, error: "self-signed certificate" })) } note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace Looks like an error.

mattlondon 3y ago

I think Gemini has missed a trick being being deliberately too simple - e.g. no inline images, no tables, no forms, no basic formatting etc. I know these were deliberately left out for "simplicity" reasons, but then they have a fairly unsimple mechanism for user login/auth that totally undermines their claim for making clients little more than slightly-augmented-terminals, which for me is a smoking gun for these things being missing purely on an ideological basis only and an attempt at controlling how people use it. That's fine, it's their thing and they can do what they want with it. There is a growing trend of people simply choosing to use XHTML Basic (1). This is a "stripped down" version of HTML originally intended for early phones, PDAs, set top boxes etc - it has the critical missing features of Gemini, but forgoes some of the more "advanced" features of the modern web people have issues with. I'd highly recommend people simply target that instead. 1 - https://en.m.wikipedia.org/wiki/XHTML_Basic https://en.m.wikipedia.org/wiki/XHTML_Basic

krapp 3y ago
It looks like this still supports the script tag, which seems to be the biggest "modern web" feature people want to abandon.
- giantrobot 3y ago
  So don't include one. JavaScript isn't fucking mandatory.
  - krapp 3y ago
    The kind of people who tend to want a separate, simpler alternative web are almost universally also the kind of people who are so allergic to javascript that they don't even want it to be an option, on the remote chance they might encounter it in the wild. They're drawn by a combination of nostalgia and spite to manifest an alternate universe where scripting on the web never even happened. Otherwise, sure, they could just write simple HTML and no javascript on the web we have now. I personally would like to retry a version of the script tag (possibly more like the object tag) with native support for WASM that would make it easy to run scripts in any language... which is closer to the original vision before Javascript "won," and possible without requiring browsers to support the languages. Also maybe things like an <include> tag for HTML, and other things that fell by the wayside of early HTML. I'd like to see something like HolyC, with native support for text, hypertext, drawing and code. But 99.9% of people seem to just want a web that's basically whitepapers with links, maybe images.

chrsw 3y ago

I think IRC would fit in here too

1vuio0pswjnm7 3y ago

(2021)