cross-posted from: https://thelemmy.club/post/59156075

Here is a brief concept: It is the federated/fediverse Web search engine (this may have been based on a older post i made)

But how is it federated, that part i hope i can explain it well,

How It May Work:

1: Instead of a crawler, it uses the older method of A Hand curated database of links and websites (By the instance mods), And no it is not like a lemmy link aggregation community, it is designed to replace DuckDuckGo, Google, Startpage, Kagi and more,

1.1: it is that the server admins add a entry to a database with that entry having a Link and a Name and other properties, which turns into search result, you see, and the database of Entries federate so you have more entries

2: The instances federate, so you have more possible search results

3: you can upvote or downvote specific search results (so if a page was SEO spam you get less of it)

4: it has integration with the rest of the fediverse

A Why:

Web Index’s, And Crawlers, could be locked down so privacy first search engines would be locked down and have issues

which would cause panic, unlike social media, you cannot easily have an alternative

A Notice:

THIS CONCEPT HAS NO AI INVOLVED IN ITS MAKING WHATSOEVER

i hope you understood the concept

Tags For Mastodonians:

#technology #fediverse #foss #opensource #freesoftware #ideas #concepts

Edits To The Concept:

1: A Karma System for instances that get a lot of downvotes in their links

2: a Vote log seen by admins with tools to help detect vote manipulation

3: A Rate limiter for downvoting and upvoting

  • ShellMonkey@piefed.socdojo.com
    link
    fedilink
    English
    arrow-up
    2
    ·
    58 minutes ago

    Sounds like either it would depend on a huge number of volunteer hours to maintain and get any reasonable size database, or you would end up with multitudes of small crawlers active to accomplish the same and would probably get blocked off as a user agent from most largish sites pretty quick.

    • Adniem@thelemmy.clubOP
      link
      fedilink
      arrow-up
      2
      ·
      52 minutes ago

      i think to solve such issue we need some things, i can imagine:

      1: oEmbed Support ()

      2: Regex, Regex and Regex

      3: A Browser Extension for adding pages if you are an admin as you go

      4: If i could lower the amount of clicks to add something it would greatly help

      • lntl
        link
        fedilink
        arrow-up
        2
        ·
        38 minutes ago

        I like the idea of a toggle in the webrowser that indexes and shares history to build indexes which could then be federated

  • MrWrinkles@leminal.space
    link
    fedilink
    arrow-up
    1
    ·
    57 minutes ago

    This was on my mind last week, yeah. Same idea. I could maybe code a quick server in TCL or something. Just a really small, simple server to return a list of curated links and metadata. Not html, maybe json. Then code a simple client to display links and ratings/stars/“credibility” which I could click to open in my browser. Sort of an old-school Yahoo or gopher or something.