Technology

Research journals along with a university based Library search engine to find them.

And everyone gave me shit for keeping my feedly account.

The Reader died, but the feeds do live on, between mastodon, lemmy and feedly I got plenty to read.

Have good filters for all the crap and use search engines with modifiers. What’s a subject or thing you’ve struggled to research so I can see if I have the same issue?

The internet used to be about people sharing what they know for free to help others and it became a WINNER TAKE ALL kind of internet. There are no blog, article, reviews, that are not fake anymore, you can buy each one of these services, even search results can be bought. Google, Duckduckgo, Bing, Kagi, are all the same shit different smell, the results are not relevant anymore, the only thing that makes them different is the browser extension you run to block the spyware, tracking, and surveillance on your every click, data that gets sold by your ISP, Social media, and every site you visit, not unless you are blocking that info between your browser and the site you visit - which is doable with a lot of browser extensions.

Honestly, niche YouTube channels. The problem is sometimes you don't want to sit through a 30-45 minute video to find the information you're after.

Interesting, was the timing and community chosen for this query better than mine a couple days ago? Regardless, this post provides me more responses to a similar question to sort through, so no complaints here!

I used to frequent searchlores.org and fravia.com back in the day, they were a treasure trove of specialised web search and data mining techniques.

UNSW maintain a mirror of the old websites, last updated 14 years ago, worth a look if you have some time on your hands.

http://biostatisticien.eu/www.searchlores.org/indexo.htm

Brave search seems marginally better than others.

This is exactly the reason I've been considering if it's possibly the time to start and launch a brand new search engine, especially now subscription based systems are so common.

With at the core a pledge to not record and/or share any user data or interaction and supported by a subscription service for who wants to pay and really oldschool tier selfhosted "sidebar" ads for the rest.

None of this "insert ads into content" shite.

For the algo, also far more oldschool "less intelligent", where keywords and content matter (backed by a curation of good/bad sites) and options for users to report sites, that will then be re-curated.

For adding sites, allow subscribers to suggest sites that then get listed to other subscribers (or if it grows large enough to support employees, subscribers AND employees) for validation.

If a site is then later found to be questionable, everyone that suggested and validated it can get a negative validation score, which will be used for future reference when selecting users to validate new sites.

Something like they get +1 for every validation they do.

But -1 for 1 bad validation, -11 for 2, -31 for 3, -61 for 4, -101 for 5, etc, so if they validate 100 sites and validate 5 incorrectly, they are no longer allowed to validate new sites.

And for validation, once there are enough subscribers, you take 100+ random subscribers, of which 50% needs to respond to validate and if 90% of responders validate positively, it passes. If less than 90% validate positively, it goes for manual review by the administration.

Etc etc.