There’s something cool about having a whole community on one machine. It’s like we’re all on one space-ship. We share the same silicon. The bits stay on the mothership.
The beautiful thing about the tilde-verse, which you’d never know until you have an account there and use it, is that it’s like a giant art experiment.
People build things out of sheer whimsy, and other people use them and share in that whimsy. The result can often be quite beautiful in ways that are hard to find on the larger intertubes.
This is a fine example.
Could you use plan and finger? Kind of, but that wouldn’t create a timeline like this does, or collect all the .plans together in a ‘feed’.
I disagree, but we’re getting into my personal ideas about what these things “should” be used for.
In my experience, your .plan file is where you advertised who you were and what you worked on in a general sense. While you could make your .plan this incredibly long rambling affair, I feel like there are better mechanisms for doing this sort of thing, and Dustin has written one.
This has fun privacy implications too. I’d like my social network to better model my actual meatspace social network. The problem, of course, is that people have many different meatspace social networks.
Except the physics of computer security, esp covert/side channels, make that impossible. You’ll always be more private if your friends are in a room together instead of on a server. Especially if there’s obstacles to seeing or hearing things that people can optionally use.
Me too. There’s something beautiful about creating a real honest to god community based around UNIX, plain text, and nothing else.
And some of the things people build there are just awesome. Are they technological marvels for all time? Often probably not, but when you can, for instance, take part in a giant text based planting simulator where you can grow cool plants of your own and water and feed others, that’s not something you see every day.
So when you follow someone .. does it just poll the side like with RSS? There isn’t a lot of implementation details that I can find. It seems simple enough that I doubt it implements anything like ActivityPub.
It doesn’t poll or implement anything like activity pub. It’s just a set structure of files that each user on a computer has. When commands are run (like “social timeline”) it parses your following file, then parses the posts files of all of those users, compiles a list, and outputs it. So everything is read every time on the fly.
I do have ideas of potentially implementing a “remote follow” feature, where the account files are published somewhere publicly accessible, and then users on other tilde sites could follow you, but that may be over complicating something that was built intentionally simple.
Oh, It was a long-ago fad (early aughts, iirc) among a very few people. It was best suited to the “one website per person” model, which never really materialized, and RDF itself never caught on very much.
But the idea of a flat file social network goes back at least that far. FOAF was, at least, designed for the web rather than a multiaccount local machine.
As for my specific experience, I had a FOAF file once, a very long time ago. I never updated it, and I don’t think it made it into any of the remaining collections.
This is really neat. The decentralized social network idea has been attempted for many years (one implementation was Diaspora in 2010), but a simple git backed version seems to be the way to go.
Reminds me of the original incarnation of UNaXcess (~1992, one machine, telnet in) except with person-stored posts instead of subject-stored posts. Truly we can never leave the past behind.
This does seem like an off-topic comment, especially given that the submission doesn’t seem to have been made by the author of the blog post :)
I feel like bikeshedding about how blog posts are implemented (“why did they use JS when they could have …”) is happening a lot and is generally off-topic and not constructive.
In particular I don’t think people should upvote comments like this. Maybe it is useful to make this comment when the author is the submitter, so they can be aware of how their implementation choices affect their users. But I don’t think we should collectively be upvoting it to the point that a completely off-topic comment chain is the highest-voted thing in the discussion.
Especially the perennial complaining about Javascript.
I’m sorry that this comment also doesn’t contribute anything to the discussion of the blog post content. If that makes me a hypocrite I would totally understand getting downvoted for it.
In particular I don’t think people should upvote comments like this.
I heartily disagree, because JavaScript abuse is technologically wrong, and technological forums should call out technological articles which make technological mistakes.
As far as I see it’s a static document that uses the URL to generate the page on the client side. It seems to be parsing the markup string in this file to generate the pages (the parsing seems to be done by these two files: (1), (2) – sadly I can’t find any published projects under the names of lukin or lipu…).
I’m not quite sure what the benefit it as compared to author- or server-side generated content, but maybe it’s just a project in progress or an experiment – tilde users tend to like to experiment and rethink mainstream or established ways of working with the WWW.
I agree with epilys that in this particular case it seems rather juxtaposing that they are talking aboet the virtues of a flat file blogging system and then insist on using javascript to render the page.
zge correctly guessed both reasons that the site is set up the way it is:
It’s a WIP. I have plans to create a build system that renders out a static version of the website. When you visit any of the static pages, it would link out to other static pages. If js is enabled, it’ll hijack the page and insert js links so the user can browse around without making a new request.
The site was also built this way as an experiment. I am all for websites that don’t require javascript. This was more of an experiment to see how I could structure things outside of the norm.
I do also plan to write a page describing the site and how it’s built. THe reason you couldn’t find lukin or lipu is because those are both custom scripts I wrote. I named them after Toki Pona words.
Already having lots of fun with it! Unfortunately tilde.town is down for sign-ups right now, they’re in the middle of writing a new tool to help handle sign-ups. :(
There’s something cool about having a whole community on one machine. It’s like we’re all on one space-ship. We share the same silicon. The bits stay on the mothership.
If you are on the same machine, finger and .plan work just as well, no need to reinvent them. ;)
The beautiful thing about the tilde-verse, which you’d never know until you have an account there and use it, is that it’s like a giant art experiment.
People build things out of sheer whimsy, and other people use them and share in that whimsy. The result can often be quite beautiful in ways that are hard to find on the larger intertubes.
This is a fine example.
Could you use plan and finger? Kind of, but that wouldn’t create a timeline like this does, or collect all the .plans together in a ‘feed’.
There would definitely be value in building those things on top of
.plan/finger.I disagree, but we’re getting into my personal ideas about what these things “should” be used for.
In my experience, your .plan file is where you advertised who you were and what you worked on in a general sense. While you could make your .plan this incredibly long rambling affair, I feel like there are better mechanisms for doing this sort of thing, and Dustin has written one.
You should check out cosmic.voyage. It’s quite literally what you’re describing.
I’ve never heard of this, it sound amazing. Thanks for sharing :)
This has fun privacy implications too. I’d like my social network to better model my actual meatspace social network. The problem, of course, is that people have many different meatspace social networks.
Except the physics of computer security, esp covert/side channels, make that impossible. You’ll always be more private if your friends are in a room together instead of on a server. Especially if there’s obstacles to seeing or hearing things that people can optionally use.
I really like the tilde verse!
Me too. There’s something beautiful about creating a real honest to god community based around UNIX, plain text, and nothing else.
And some of the things people build there are just awesome. Are they technological marvels for all time? Often probably not, but when you can, for instance, take part in a giant text based planting simulator where you can grow cool plants of your own and water and feed others, that’s not something you see every day.
So when you follow someone .. does it just poll the side like with RSS? There isn’t a lot of implementation details that I can find. It seems simple enough that I doubt it implements anything like ActivityPub.
I suspect it collates files just in time when you run
social timeline.It doesn’t poll or implement anything like activity pub. It’s just a set structure of files that each user on a computer has. When commands are run (like “social timeline”) it parses your following file, then parses the posts files of all of those users, compiles a list, and outputs it. So everything is read every time on the fly.
I do have ideas of potentially implementing a “remote follow” feature, where the account files are published somewhere publicly accessible, and then users on other tilde sites could follow you, but that may be over complicating something that was built intentionally simple.
Any chance this is based on FOAF?
Nope, I had never heard of that, but it looks interesting. Thanks for sharing :)
No, in typical Tilde style, it’s plain text, that’s all.
Oh, man, I want the GGG so badly now. You’ve really opened a can of worms for me. What’s your experience with FOAF?
Oh, It was a long-ago fad (early aughts, iirc) among a very few people. It was best suited to the “one website per person” model, which never really materialized, and RDF itself never caught on very much.
But the idea of a flat file social network goes back at least that far. FOAF was, at least, designed for the web rather than a multiaccount local machine.
As for my specific experience, I had a FOAF file once, a very long time ago. I never updated it, and I don’t think it made it into any of the remaining collections.
This is really neat. The decentralized social network idea has been attempted for many years (one implementation was Diaspora in 2010), but a simple git backed version seems to be the way to go.
Reminds me of the original incarnation of UNaXcess (~1992, one machine, telnet in) except with person-stored posts instead of subject-stored posts. Truly we can never leave the past behind.
A bit off-topic, but why does this page require javascript to render? It’s plain text.
This does seem like an off-topic comment, especially given that the submission doesn’t seem to have been made by the author of the blog post :)
I feel like bikeshedding about how blog posts are implemented (“why did they use JS when they could have …”) is happening a lot and is generally off-topic and not constructive.
In particular I don’t think people should upvote comments like this. Maybe it is useful to make this comment when the author is the submitter, so they can be aware of how their implementation choices affect their users. But I don’t think we should collectively be upvoting it to the point that a completely off-topic comment chain is the highest-voted thing in the discussion.
Especially the perennial complaining about Javascript.
I’m sorry that this comment also doesn’t contribute anything to the discussion of the blog post content. If that makes me a hypocrite I would totally understand getting downvoted for it.
I heartily disagree, because JavaScript abuse is technologically wrong, and technological forums should call out technological articles which make technological mistakes.
There are lots of things that are “wrong”, technologically or otherwise. That doesn’t mean that every thread needs to be derailed by discussing them.
As far as I see it’s a static document that uses the URL to generate the page on the client side. It seems to be parsing the markup string in this file to generate the pages (the parsing seems to be done by these two files: (1), (2) – sadly I can’t find any published projects under the names of lukin or lipu…).
I’m not quite sure what the benefit it as compared to author- or server-side generated content, but maybe it’s just a project in progress or an experiment – tilde users tend to like to experiment and rethink mainstream or established ways of working with the WWW.
I agree with epilys that in this particular case it seems rather juxtaposing that they are talking aboet the virtues of a flat file blogging system and then insist on using javascript to render the page.
Hello everyone, I’m the owner of the website.
zge correctly guessed both reasons that the site is set up the way it is:
It’s a WIP. I have plans to create a build system that renders out a static version of the website. When you visit any of the static pages, it would link out to other static pages. If js is enabled, it’ll hijack the page and insert js links so the user can browse around without making a new request.
The site was also built this way as an experiment. I am all for websites that don’t require javascript. This was more of an experiment to see how I could structure things outside of the norm.
I do also plan to write a page describing the site and how it’s built. THe reason you couldn’t find lukin or lipu is because those are both custom scripts I wrote. I named them after Toki Pona words.
Already having lots of fun with it! Unfortunately tilde.town is down for sign-ups right now, they’re in the middle of writing a new tool to help handle sign-ups. :(
Pic of it being used on ~town
Related for people not all on one computer: twtxt
It could have served the file system via 9p, with a proper auth server. Sounds like a dream of plan9 dominance.