❄️Faerie❄️Jun 23, 2023
Amelia

Internet search sucks now.

Making a classic "search engine" requires obscene amounts of money and resources, so creating direct competition is not a realistic option.

But what if we had a way to each take our own personal bookmarks and link collections, and share them, using a federated protocol?

My fifty sites, your sixty-five links, some friends and their contributions... All hand-selected and vetted, manually tagged and organized by real people... Hosted on small, volunteer-run servers...

What if web search results were organized not by "how many ad dollars did this site generate for the search company" but rather something like "someone you know directly tagged this site, and someone twice-removed tagged this other site so it's lower priority"?

#DisabilityDrivenDevelopment

Jun 23, 2023 at 3:08pmWeb
Show threadajsWritesThingsJun 23, 2023
@mordremoth 100% believe that this is the huge thing that's missing. I keep saying it, but 'federated stumbleupon' wouldn't be a bad place to start
Show threadajsWritesThingsJun 23, 2023
@mordremoth edit - *huge thing, not high lol
Show threadJune BlenderJun 23, 2023
@mordremoth Amelia, where can I learn about disability-driven development? Is there a seminal book or online class for developers?
Show threadAmeliaJun 23, 2023
@juneb Check out https://spoonstack.org/ and particularly the Essays section!
SpoonStack

Show threadJune BlenderJun 24, 2023
@mordremoth Read it. Are you developing on GitHub? Can you please link?
Show threadAmeliaJun 24, 2023
@juneb Check on that site, under the Projects section.
Show threadThe DoctorJun 23, 2023
@mordremoth https://drwho.virtadpt.net/archive/2019-02-04/sometimes-the-old-ways-may-be-best/
Antarctica Starts Here. | Sometimes the old ways may be best.

Show threadAmeliaJun 23, 2023

@drwho Yep, exactly. Except this process needs to be far, far more accessible so that the barriers to participate don't just create a content bubble of "stuff highly techy people have collected."

The raw ability has existed for years. What's missing is a good experience.

Show threadArunmozhiJun 24, 2023
@mordremoth running something YaCY which can crawl user submitted sites should offer such a setup. It already supports setting up a cluster of nodes and sharing index and crawling workload IIRC.

https://yacy.net/
Home - YaCy

YaCy P2P - Decentralized Search Engine

Show threadAmeliaJun 24, 2023
@tecoholic As I've noted to half a dozen other people in this thread, YaCy does not federate over an open protocol. This is not the same thing at all.
Show threadArunmozhiJun 24, 2023
@mordremoth Ah. Sorry. I saw just one reply in my instance and didn’t realise there were so many other replies.
Show threadCyber YukiJun 23, 2023

@mordremoth We need to bring back internet directory sites.

EDIT:

Remember Yahoo! pages?

Entertainment / Games / Roleplaying / Dungeons & Dragons

Adding your page was TRIVIAL.

Show threadmnl mnl mnl mnl mnlJun 23, 2023
@mordremoth very interested to talk about this! Have been thinking about this for quite a while! Here are some of my notes:
- https://publish.obsidian.md/manuel/Projects/Ideas/IDEA+-+A+federated+custom+search+engine
IDEA - A federated custom search engine - mnml's vault - Obsidian Publish

# A federated, convivial search engine I am thinking of either piggybacking off something like [https://lieu.cblgh.org/](https://lieu.cblgh.org/ "https://lieu.cblgh.org/") or build my own to build li…

mnml's vault
Show threadAmeliaJun 23, 2023

@mnl Love seeing how many people are already contemplating ways to do this!

To be honest I think as long as we lean towards maximum inter-operability and keep a focus on personal control, this has real potential to reshape Internet use entirely.

Show threadCyberpunk LibrarianJun 23, 2023
@mordremoth Federated del.icio.us? I'm on board for that.
Show threadprolixiousJun 23, 2023
@CyberpunkLibrarian @mordremoth or Federated pinboard.in
Show threadTindraJun 23, 2023
@mordremoth @CyberpunkLibrarian im old enough to remember when collections of human-curated links sorted by topic was the state of the art.
Show threadCyberpunk LibrarianJun 23, 2023
@TindrasGrove @mordremoth Same here and I honestly miss that. Google has so badly played itself with SEO insanity and, while projects like Searx are great, you're still just using it to search many of the same broken search engines.
Show threadAmeliaJun 23, 2023

@CyberpunkLibrarian When I first started using the internet (1996 or so) everyone found things by sharing URLs manually. Curated directories and link aggregators were the natural evolutions.

I think "search engines" may well end up becoming a historical anomaly someday. The only real failing of ideas like webrings is they couldn't talk to each other to expand.

Show threadFrank SkorniaJun 23, 2023
@mordremoth @CyberpunkLibrarian The final season of 'Halt and Catch Fire' had a really good plotline about this, with the daughter of 2 main characters making a web directory on her own and how they tried to scale it up.
Show threadPicks the EvangelistJun 23, 2023
@mordremoth Ratings, rather than simple tags, with an aggregate score generated by social proximity? Maybe ratings could decay over time. Would you want to, idk, rate your friends in terms of expertise on given subjects to trust their ratings more for a given subject?

Might be something to build into Searx, or at least use Searx as a base for ...
Show threadEarthshine --> movedJun 23, 2023
@mordremoth This is exactly what OpenCola was meant to provide. https://www.opencola.io/
OpenCola

OpenCola, the antisocial network

Show threadAmeliaJun 23, 2023

@earthshine Cool!

I'd love to see their work after some heavy engagement with some accessibility advocates.

Show threadchrisJun 23, 2023
@mordremoth
Related: https://idiomdrottning.org/folksonomy

@Sandra
I Want a New Folksonomy

Show threadNimaJun 23, 2023
@mordremoth the other day I searched for "timer for windows 10" and Google gave me blue light filter websites!!
Show threadNimaJun 23, 2023

@mordremoth there was a similar service back in the web2 era.

https://en.m.wikipedia.org/wiki/Delicious_(website)

Delicious (website) - Wikipedia

Show threadAmeliaJun 23, 2023

@point5a Link aggregators were huge for a while, yes.

The main failing they all shared was being centralized services. Their growth and success inherently required increasing costs and complexity - a "vertical scaling" problem, by nature.

Federated, decentralized solutions face other challenges, but by nature are horizonal scaling models, which means they are less likely to require a single entity to spend huge amounts of money keeping things on.

Show threadNimaJun 23, 2023
@mordremoth I believe future is decentralized. Social decentralized bookmarking is sounds really good idea.
Show threadMarcus MüllerJun 23, 2023

@mordremoth @point5a that sounds a bit overly idealistic. We have decades of experience searching decentralized databases now - remember late 90s/early 00s and their p2p file sharing networks? The search traffic gets massive, unless you start inserting centralized nodes.

Just from a purely theoretical view, the overall effort of asking everyone who takes part about everything must be larger then asking one central database. I honestly fail to see how the net effort for individual con-
1/2

Show threadMarcus MüllerJun 23, 2023

@mordremoth @point5a
-tributors could be lower in a decentralized scenario than if the same people just pushed their local knowledge to a permanently available central service (and paid their diminutive share of upkeep for that).

Decentralization is not a panacea for computational load. It's the worst thing for overall traffic volume. Don't romanticize decentralization for something it isn't or something it's the crass opposite of.

2/2

Show threadDexterJun 23, 2023

@mordremoth I thought this was the intention of what link aggregator sites like Lemmy/Kbin are designed to do.

EDIT: My apologies, I reread this and it comes off hella aggressive, which absolutely isn't my intent.

More just I'm curious if Lemmy/Kbin fit what you're describing, and what would need to be different if they currently don't.

Show threadAmeliaJun 23, 2023

@wheeljack They are designed to be link aggregation systems, yes, but I'm not talking about replacing link aggregation - I'm talking about replacing web search.

My point here is that decentralized link aggregation could be a great starting point for replacing the classic "crawl and index" systems that make centralized search engines extremely costly.

Show threadmiraculous enbybugJun 23, 2023
@mordremoth I had this exact idea a few months ago… a federated website directory
Show threadWilling to cut a deal. . .Jun 23, 2023

@melunaka @mordremoth

... because the best way to find stuff is not with a search engine, but asking your online connections.

Show threadHaelwenn /элвэн/ Jun 23, 2023
@mordremoth sixty-five?

I have 224 public bookmarks and counting: https://hacktivis.me/bookmarks

And then there's link-sharing dedicated platforms like lemmy, shaarli, lobsters, … that probably go well in the thousands.
Bookmarks - lanodan's cyber-home

Show threadLatte macchiato  Jun 23, 2023
@mordremoth look into yacy
it goes into that direction
Show threadAmeliaJun 23, 2023
@privateger If YaCy used an open-standard protocol that could interact with other software implementation, then yes, it would. From what I can tell, it's just the distributed part, not a genuinely federation-oriented approach.
Show threadAJ (Rising Soul)Jun 23, 2023
@mordremoth Yesss! I want something like this to exist. I would totally curate a small directory of links on certain topics I'm knowledgeable about.
Show threadUrsidinoj/The BjornsdottirsJun 23, 2023
@mordremoth Web Directory Service?
Show threadSolarpunk StationJun 23, 2023

@mordremoth

This seems a lot more useful than my idea of bringing back web rings.

Show threadAmeliaJun 23, 2023
@solarpunkgnome The only real failure of web rings was not having tools to automatically share, connect, and grow them!
Show threadvriska, thief of light Jun 23, 2023
@mordremoth i feel like search engines are used for both "recommending resources" and "finding answers to very specific questions"
i think this would only really be practical for the former
Show threadAmeliaJun 23, 2023

@leo That's fair, but part of the issue right now is that using search engines to answer questions is increasingly ill-advised, due to the growth of pure junk masquerading as content for SEO/advertising purposes.

I'm thinking about a world where we talk to people to answer questions instead.

Show threadvriska, thief of light Jun 23, 2023
@mordremoth i dont think we should just accept that problem as unsolvable
needing to find an expert to talk to to answer every question i have would not be a good thing
Show threadvriska, thief of light Jun 23, 2023
@mordremoth (....especially considering that in a lot of fields even experts heavily use search engines to answer questions nowadays!)
Show threadAmeliaJun 23, 2023

@leo I said nothing about accepting anything as unsolvable, nor did I suggest any such extreme as "every single question." You seem to be objecting to your own partially unconscious assumptions, not my ideas.

You might benefit from taking some time to understand why the current setup of technology is designed to condition you to shut down your imagination, and who actually benefits from that status quo.

Show threadTravis KJ5DCLJun 23, 2023
@mordremoth have you played with https://yacy.net/ ?
Home - YaCy

YaCy P2P - Decentralized Search Engine

Show threadAmeliaJun 23, 2023
@travis Non-centralized is not the same as federated. YaCy does not interoperate with other software over a shared, open protocol, as far as I am aware.
Show threadTravis KJ5DCLJun 23, 2023
@mordremoth thats valid, and i do agree. its too bad that, at least so far, a diy, self-hostable, practical, search thing doesnt exist.
Show threadsarah tonin Jun 23, 2023
@mordremoth i wonder how well link aggregators like that work for specific researches (debugging issues, needing precise info on niche subjects)
Show threadAmeliaJun 23, 2023

@SRAZKVT Nobody knows because this does not yet exist!

Only one way to find out...

Show threadsarah tonin Jun 23, 2023
@mordremoth if anything, both of these cases could in easy manner just go back to a single centralised website, which is essentially just a search engine, but for a specific topic, which is slightly better, but still not good
Show threadcontrôlisteJun 23, 2023
@mordremoth It's still corporare silo hell, but I've had a great experience with are.na. Being able to see who saved the same links as you really helps discovering stuff
Show threadAndrew CusworthJun 23, 2023

@mordremoth Seconding several of the other responses, there were genuinely interesting routes through the web courtesy of web rings and directories, not to mention pointers from usenet groups.

There was something appealing about the chain-linked hypertext flow of web rings, albeit that (and I was very young, so my memory is hazy and probably misinformed) there must have been a substantial element of gatekeeping involved in the membership of those rings. I guess that, by nature, is one of the ways in which 'hand-vetting' works, and it is no more curated (and possibly rather more coherently curated) than a search engine built on unknown and largely inscrutable algorithmic processes.

And, do I miss that long-lost constellatory phase of the internet? Absolutely.

Show threadJon RJun 23, 2023
@mordremoth This strikes me as a brilliant solution to a lot of problems that are already bad and are only going to get worse!
Show threadJNP-3RJun 23, 2023
@mordremoth Heck yeah, federated DMOZ. You could set up a suggestion system for each instance too, so anyone on that instance could send a URL to admin for approval.
Show threadJamey SharpJun 23, 2023

@mordremoth I've wanted this kind of "search my friends' bookmarks" tool for years, both for the reasons you mention as well as to avoid needing centralized DNS. If you have URLs that are not human-memorable, perhaps because they contain content hashes or public key IDs, then having the ability to find those URLs by some other means is even more important.

I kind of want such a search engine to also have a reputation system of some sort to answer questions like "did this link come from someone I trust to have checked its authenticity?" But I'm not sure anybody has yet worked out how such reputation systems can work in practice.

Show threadRob ParsonsJun 23, 2023
@mordremoth this makes me think instantly of Diigo.
Show threadI am Jack's Lost 404Jun 23, 2023

@mordremoth

This could also extend to product ratings/reviews...