well it works like this:

basically it is a decentralized search engine that uses activity pub (so probably it is technically in the fediverse)

how??

well instead of the thing scraping the entire web (which is probably by what you guess has many implications and is probably impossible)

it is a human made database of websites that syncs with other databases of websites

so if there for example are 2 databases in 2 servers each with 50 websites: you can get 100 websites

and to make it easier for links and more to be added

i make requests: which are requests for a specific websites or URL to be added, which can be approved or denied

how do you fix such idea

    • Adniem@thelemmy.clubOP
      3·
      20 hours ago

      well the ui is more like duckduckgo or any other search engine

      and its more like a search engine than a social network

      and its more optimized for its purpose to be a search engine

      • thisbenzingring@lemmy.todayEnglish
        5·
        20 hours ago

        Sounds like what Angie’s List started out as. The Internet was basically this before Yahoo invented the search engine.

        It’s a good idea but it would take a lot of effort to keep it from becoming overwhelmed with spam or exploits

        • KoboldCoterie@pawb.socialEnglish
          1·
          19 hours ago

          A search engine that functioned similar to old web rings of the 90s, where each website effectively endorses a collection of other websites and the search starts with a list of user-approved trunk sites and gives results that branch out from those sites, could be an interesting version of this. Would require some initial work on the user’s part (flagging sites they trusted and liked to use as the trunk of the tree) and on websites’ part (each including a list of other sites that they endorse), but it could be a much nicer way to browse. Obviously search results would be very limited, it wouldn’t be good for looking up specific information on new topics, but if you had, for example, a tree of sites with gaming information and news, and you were looking for articles about a new console release, presumably it could help you find it. Think of it like current web search but which will only provide results from sites that exist in your tree.

          Show how many steps each result is removed from the trunk, show which site each referral came from, and give the option to blacklist sites that you don’t want to see (and in turn, all of their referrals), and over time you could end up with something approaching the old internet, at least in so far as you saw.

            • KoboldCoterie@pawb.socialEnglish
              3·
              18 hours ago

              I think the difference is that, if I recall correctly, StumbleUpon was more or less a recommendation engine, where anyone could submit interesting websites to be added to the list, whereas what I’m proposing is more akin to a search engine curated through individual websites’ endorsements. The primary difference being that if you get bad endorsements, you can blacklist the source, so sites have a vested interest in providing good recommendations (and in turn, in theory, receiving recommendations back). It’s basically how the web worked in the 1990s… you found sites via groups of links on other sites, not through centralized search engines, and those links were curated by other sites’ owners.

              • thisbenzingring@lemmy.todayEnglish
                2·
                15 hours ago

                StumbleUpon was a good time, you had to give submissions a few categories. So you would set your categories and the button would give you whatever page was next in the list that fits your combined categories. You could blacklist sites, categories and users that submitted something. That’s how it could be user controlled

                I enjoyed it back in the day. Made short stints on the web a little more random. It was a great way to find small and interesting pages