Skip Navigation

Could We Build a Decentralised Social Platform Rooted in Place?

carlnewton.github.io Could We Build a Decentralised Social Platform Rooted in Place?

Exploring the idea of a decentralised social platform based on location.

Over the past year or so I’ve been playing with the idea of a decentralised social platform based on your location. By putting physical location at the centre of the experience, such a platform could be used to bring communities together and provide a source of local information when travelling. Please let me know what you guys think.

53
53 comments
  • So basically "NextDoor: Fediverse Edition".

    I'm not against but I think uptake will be difficult because, depending on your starting region you may be competing with an established product and because there is a heavy venn overlap between people who care about decentralization and those who care about privacy and wouldn't want an app tracking and to some degree giving away their location.

    (Clarity Edit: I have a cold, my ending was guessable , but technically gibberish.)

    51
  • I spent several weeks thinking about this exact idea.

    Federation is cool. You could set up each instance to only federate with instances for nearby towns and cities. Maybe a “2 district” radius. Users would only see content for their local communities. Local news stays local. Local government could officially participate if they wish. People you talk to are actually neighbors you might see in person. Larger regions like counties, states, provinces, or even countries, could also have dedicated instances and federate similarly. I think this is the big appeal and it sounds awesome!

    There are a few problems 🙂

    First is a little bit of confusion with posting. Let’s say that I see a post about a cool new restaurant in my town. I share it with a friend who lives a few towns away and that’s outside the “federation radius”. I can’t share the post with that friend very easily. Maybe the tools could be enhanced to make this viable?

    Second is a matter of privacy. How do you know that new accounts belong to people associated with the geographic location of each instance? If you don’t validate, the system will certainly be abused. If you do validate, then users need to supply some real info! Home address, ID, etc. that’s a big deal for users and instance admins.

    Third. What happens if you move? Do you have to abandon your old account and start over? Again, the system itself can be developed further to solve this. But that’ll take time and money.

    Next is the operating costs. You would need to build thousands of instances to build this system up. And each one would have to be tied to a geographic region. You need new features to handle signups this way. You have the simple cost of running these servers. You probably need a lot of staff to manage it all. This is an expensive platform for one party to run. Alternatively…

    It doesn’t have to be one party running this entire system. That’s the point of the Fediverse, right? The operational costs go way down if anyone can run their own instance. But how do you enforce the rules of federating with instances for geographically nearby locations? I don’t see a reasonable way to solve this one.

    I could probably keep listing issues. But these are the big ones IMO. If you solve these, the system is viable and could be amazing.

    34
  • If there were more like-minded people around my local area IRL, I wouldn't be on the internet looking for like-minded people; I'd be at a bar.

    16
  • A federated Nextdoor or local Facebook group seems a great idea as it is clearly popular. My only issue is with dividing Home and Nearby by distance. That works well in ideal cases but can get weird in others.

    So I live on the coast at the mouth of a big river. If I look up events near me they are done as the crow flies and it can offer up locations on the other side of the estuary. However, to get their by land involves a long U shaped journey through a tunnel, so what appears to be 5 minutes away is 30+ and of little interest to me.

    Another example might be city vs country. In the city 5 miles would drag in a large population, in the country it might not even get you to the next village.

    A better solution might be postcodes/zip codes (or equivalent) - they're usually designed to encompass similar population numbers, so change in size depending on population density. The data is also freely available (it is on OpenStreetMap, for example) and it should be easy enough to crunch through the data and create a database that defines the adjacent areas for a specific postcode (looking them over, it tends to be 5 or 6).

    Other than that, I think the main issue would be getting enough people involved as a quiet feed would kill it dead.

    12
  • Loving this concept. May I make a suggestion? Show this to and discuss this with your local library. That strikes me as a good potential partner, and a model that can be replicated in most places to potentially help with everything from hosting to community resources access.

    9
    • Thanks! I'll take all the suggestions I can get! This is interesting, and something I've never really considered for any local project. Is it common for libraries to take an interest in online platforms like this? Which country do you live in by the way? I'm not sure if it's a knowledge gap on my part or just something that libraries in the UK wouldn't get involved with.

      3
      • I live in the UK, but am from Norway. I know a few librarians though, and I know that community libraries are usually (or at least often) interested in projects that can connect their communities and help them with outreach. Something like this certainly could do that, and with libraries existing in most communities there is a built in network for broader proliferation there.

        I’m also just very keen on the idea of libraries having a central role to play in the future of the broader fediverse ecosystem.

        Edit: It may be key to pitch this to them not as a platform, but as a decentralised community network.

        5
  • Build it yourself. Stash several mini computers connected to Lora radios around town. Connect to form a mesh all focused on the same fediverse page. Connect wifi hotspots to those servers, connecting phones to the hosted page, not to the internet.
    You'd get a localized community site that by design can't be infiltrated by people from across the country hating on you (e.g., San Francisco and Texas).
    No idea on feasibility or how much work, but seems to address your question at least.

    8
    • This is where I thought it was going and, the ideas needn't be incompatible. If you went local (down to say a zip code/postcode) the individual resources required would be minimal. I wonder if people would be able to contribute a smidge of storage/bandwidth like they do with things like SETI. The local instance/mesh would essentially know your location and sign you up to the right one.

      I think something like Nostr would work, as each user would be a node (some allowing broadcast and networking), so moving house or even going on holiday wouldn't be an issue.

      5
  • you could just design an alogorithm that heavily favors any posts physically close to you. the closer they are, the higher in your feed.

    6
    • I'd like to see this experiment carried out at a sufficient scale. I feel like there would be a benefit to a gravity like component that takes density and distance into account so that people in sparsely "populated" regions aren't just effectively seeing an unprioritized feed of the entire network.

      5
      • I like this! This is an interesting and probably an effective idea for addressing the issue of desolate communities. It does introduce a new issue though. Let's say for the sake of argument, I created an instance where I lived, and it engulfed the British isles, and then after some time, a new instance was created in Scotland. The posts that were previously made in my (now England) instance that are geographically located in Scotland would need to be transferred to the Scottish instance, because otherwise, they would never be found using the relevant instance finding technique that I've described. This doesn't sound like a terrible technical issue (though probably not an entirely trivial one given that you don't want to have all of your posts hijacked by a bad actor), but it does sound like a data/privacy issue, in that the users who posted that data have essentially had it transferred to the ownership (and moderation decisions) of somebody else without permission. An interesting thought none the less. I'd prefer the platform to be as simple as possible but if it can't be picked up due to a lack of interest then that would be a design flaw. On the other hand, I suppose it'll be in the interest of each owner to foster their own community.

        3
  • I have been thinking about this for a while. I want an online community that encourages meet-ups and face-to-face time. No so much twitter-esq, but more event based. Maybe with a feed that shows small announcements, news and reports in a magazine style?

    It would be super cool if many towns and cities have their own online meeting place, that can also interact with neighbouring places!

    I haven't look to much into it, but maybe @bonfire@indieweb.social can provide this?

    EDIT: Their webpage: https://bonfirenetworks.org/

    6
  • There was a nonfederated one. Elk talk, or Oxes. Anyway, most useful on college campuses. Pointing this out for reseach.

    Sounds great!

    Edit: It was yikyak!

    6
  • Yes, I would use this a lot! That's an incredibly exciting idea and something that I think is desperately needed for the fediverse. I hate to say it, but I think this is something that other (currently more popular) platforms still have an advantage over.

    As someone that travels for work, it's not always easy for me to learn where and what my resources are. I could see a platform like this being able to help out immensely.

    Buying/selling/donating groups are something that immediately comes to mind too.

    6
  • @carlnewton@feddit.uk one advantage to a system like this, is that you don't need it to catch on in a bunch of different communities for it to work. It can catch on even in one place, and it will already be useful for the people there.

    5
  • Lots of people with similar ideas so there is probably demand. When I was thinking of ways to bring people together I had considered a platform so to speak that lets you join different communities, and over time or once verified you gain voting weight. Like lemmy or other website have communities and you can join or watch without playing and they can hose meetups, except the system I had in mind doesn't have the same style of 'voting' on things that sites your used to are. There are chains of trust to build voting weight or reputation.

    I'll post my rambling thoughts here:

    The idea here is for a 'consensus engine' which results in many applications and a universal voting method. As the current repository of these posts is about to be deleted I am copying them here, separated by three dots such as ... June 7th 2023 From small clubs to nations to solar systems, we can bring people together through similarities of feeling and thought. Share ideas and information how modern social media does, bounded by whatever group you chose, which can be rebuffed if needed. If you managed to read the sticky post on the universal voting system then it will help to see how this Consensus Engine can be used.

    I think it could supplement a future social media if need be in the sense that you can connect with your friends, or anyone you know or don't and place whoever in whatever groups you chose and then interact with just that group. From sharing 3D home videos to posts about their feelings it allows open communication. However the communication goes both ways. When anything is posted then every aspect of it can be judged by others. From how truthful it is, to how it stays on topic or does not, to how poetic it sounds, as well as how it might appear like a paid ad, and so on. Individual sections of posts can be judged separately.

    part of the goal of this is to use it for scientific research and to establish a chain of trust. This is done by having experiments that are reproducible and thus increasing the weight of that users locally, in the related groups to that subject, a weight that diminishes over time but can increase as reproductions and positive judgements on their works are had. The goal is to allow experiments to documents every step of the process and to do so openly so that humans as a whole can learn. Also to allow others to chime in and offer their ideas.

    The system could store parts of the comment or post, then the weights of those posts could be changed by others. When the OP goes offline their system saves a chunk of data regarding the last status of the thing, and when they come back they ask the network what that status was so the weight of their comment and themselves in that group can be updated. This is done by saving some data of other users when they save their own data, when they come back online they are now the network which can update the other data to the latest of their download. The bleeding edge of consensus will be fuzzy, but overtime it can emerge and even change when new data comes in.     It can allow for people to find things in common like 'I'd like to have all my needs met,' to 'I don't like it when people steal from me.' and working up one common thing at a time we can find where some people are okay with pulling a gun on a 6 year old looking for a lost kitten and how some people can justify giving money to know grifters. We want to make the truth provable and connect it clearly to actions and ideas. If some people who are religious or very us right wing 'wake up' and see what they have done, maybe they can change and help everyone.

    Another thing it can do it to lay out future plans and past history for tv show writers, town charters, and whole civilizations. People can decide they want to stop the growth of their city or increase housing density like normal voting, but they can plan for things father out like working together as one world to build colonies on Mars or something. They can do that by outlining all the steps needs to do something, each step broken down by other steps, and each step facing a many potential 0poor outcomes and the steps needed to provide against them.

    Elections on things from single and specific issues to worldwide issues can be voted on by anyone, and those that have the most positive involvement based on what has consensus as fact have high weight to their own votes than others who, when asked questions about the subject they are voting on, disagree with the consensus.

    When a comment is made others can vote on it, to vote on something you must vote on a second thing as well. You have the option of going deeper than the question you are asked. Things like, do you agree with this, are these two things similar/the same. All sorts of classifying questions and web of relation questions are asked. when someone answers they are asked if they think their answer is similar to another one posted, they can also contribute more like why it is, what parts are different, or anything, and then others can vote of how they feel about these answers. People are asked questions both within their subjectgroup collection as well as outside and they can adjust how the ratio to an extent.

    By using the CE anyone can contribute to any subject, little heard of ideas can gain traction from locally recognized experts and false hoods can be more readily proven false by those with the most recognized experience. Repeat allegations of the same falsehoods can easily be debunked. Definitions of semantics, words and phrases can be decided upon so no one can try to change what a word means partway into a discussion.


    I don't know how to create this, I do think this might help in moving to a post-scarcity world if we could find a way to implement it. ... Oct 24, 2022 the Consensus Engine - a universal 'voting' mechanism that can be used by anyone for everything. A protocol meant to bond people together.

    A way to do things including:

    • accelerate scientific progress
    • reduce misinformation
    • Plan & manage groups (company, city, society, book club, etc).
    • share personal opinions
    • post interesting things
    • translate ancient text
    • interpret art
    • find specific variations (product development, preference discovery)
    • document history
    • clarifying definitions
    • expose hypocrites
    • A/B/etC alphabet testing between strategies

    Through the system a user can rate things using a variety of 'rating angles' which are specific aspects. A user can judge a phrase as being things (or not being those things) such as:

    *relevant to the discussion       *factually accurate       *interesting


    using the system in small settings at first would enable it to show that it works which can help in adoption.       This would also allow each user to become verified in person.

    If the system could be used for actual voting the the registration could be made by sending a code to the users registered voting address.


    Users contribute their opinion or posts which can be judged by others.

    Other users can judge: the entire comment, parts like specific phrases or sentences, even words or letters. By rating any comment or part of one with one or more of the multiple voting categories we can get an idea of why people like things.


    As users use the system they are asked to vote on other people's opinions and posts. Typically a person has to answer a consensus related question at least 1 out of 3 posts but they have the option to answer more questions.     They can get more in depth with specific topics but are every once in a while shown questions from outside their topic focus, in a sort of 'more closely related- more often, less related topic - questions about these topic appear less often.'


    As a user you can make posts or comment on topics. The users can determine the topic and subtopics if the post is open to all.

    When you make a comment about a subject you are shown a similar comment, The system might say your post and the shown post are the same intent. You can dispute this and highlight why your post is different. As always, other people will see the two posts and asked to judge or explain the similarities or differences.       We want this to be a way to find intricate but existing differences as well as find the motivation for each decision to see how the same motivations end up differently in other circumstances.

    We can't put the small pieces at the end back together but we can discover through analysis what the commonalities are and find what the building blocks are before they crumble into actions.


    Topics that are deemed more important by the group can carry more weight to be given more visibility.


    Viewing random or less connected topics can help expand the sphere of knowledge for users. They have the chance learn more about a subject they find interested but have not encountered before. As a user is exposed to more about a subject and gives more input or opinions then they have more chances to gain weight to their opinions in that subject.

    A topic is defined as different from another similar topic by the users so that the thresholds of things can be found and if there is a difference in threshold values from one grouping of people or another we can better understand it


    The web of connections can be used to view common traits and similar interests. If users who consent decide to they can use the platform to find people who like similar things, like matching people who have similar taste in music or weird movies to be friends or whatever the people are looking for.       This can help create friendships and alleviate loneliness if users want to find other users with similar tastes.


    users can chose to give their info out or not, and by info I mean everything. Privacy should be paramount. Th

    2
  • How to keep it from becoming disgusting like Topix? Vile things were said on those forums.

    1
  • Invite only app maybe? If you get the sign up process to be an NFC or QR code type thing you could effectively limit it to people in that area that know each other or that see the qr code. You could even have the signin do imprecise GPS checking if you wanted to limit it further (not fool proof, but does it have to be?)

    You could have something like the described in the digital bonfires idea and have a regularly scheduled means of moving things from local to people just passing through.

    1
You've viewed 53 comments.