OSDev.org

bzt wrote:ps: if your websearch engine did not show this link, then you've faced the deepest hell of modern internet plagued with destructive information bubbles. Try deleting all cookies, use private mode, change browser identification string and change IP if you can. Or try to train the search AI that you need technical documents, good luck with that.

I'm sorry to hear it takes so much effort to defeat search enging bubbling these days. Years ago, I could change bubbles just by using a different profile of my browser. It was good to have different bubbles for different activities, but not so comfortable changing profiles all the time. I use privacy-focused search engines instead -- no bubbling.

I mostly use Duck Duck Go. (Short domain: ddg.gg) In my experience, it's very good when searching for technical things. I don't search for other things much, but seems all right there too. Only once, I couldn't find a comic I knew existed. It was because gocomics has a broken robots.txt; DDG was overly cautious and didn't index it.

That one time, I used Startpage instead. When I first came across it, Startpage stated they paid Google for search data without bubbling. I don't see a mention of Google now, so I don't know what the back end is. EDIT: Startpage has been aquired by an ad company. Judge that however you will.

Note that I don't particularly use these for actual privacy any more; I'm not evaluating that. It's just more comfortable to not be bubbled.


As an example of how bad bubbling can get, a guy I know with a highly technical job searched for a wing mirror for his motorbike once, and ever afterward Google would only show him Harley Davidson wing mirrors when his work-related searches included part numbers beginning with HD. (The "ever" bit might be an exaggeration, but it was annoying.)

eekee wrote:I'm sorry to hear it takes so much effort to defeat search enging bubbling these days.

Well, the point is, remove everything that could be used to track you. Filter bubble is impossible if the search engine doesn't know who you are. Besides of cookies, websites with JS enabled could use local storage to identify you. Using private windows should clear that too, but I've found that Firefox only clears that if you restart the browser; and Chromium doesn't clear it at all, not even in private mode. To check this, type "localStorage" in the Developer Console on a particular site. So I also turned JS off, and simply don't use search engines which are heavily loaded with JS bullshit (I look at you google).

BTW, duckduckgo is a frontend to bing, and startpage is nothing more than a google frontend. This means they proxy your search query to these engines, and in turn the engines use these frontend's bubble (which hopefully is wide enough because many people are using them for many different searches). Fair warning, startpage has recently been acquired by an advertisement company, so I guess it's not safe to be used any more.

My favourite is Qwant (html-only version), but you can find more alternatives in this ITFOSS article.

eekee wrote:Note that I don't particularly use these for actual privacy any more; I'm not evaluating that. It's just more comfortable to not be bubbled.

Agreed! Privacy is very important, but not serving the desired results actually renders a search engine completely useless! Not to mention you can't use "let me google for you" links any more, because the recipient is not guaranteed to see the same results!

Cheers,
bzt

bzt wrote:My favourite is Qwant (html-only version), but you can find more alternatives in this ITFOSS article.

Thanks! It's great to know there are usable html-only search engines, Duck Duck Go never fully implemented their html-only page. HTML-only is good for OS devs who don't want to support or develop a modern web browser. Personally, I'd put developing a modern web browser up there with developing an OS, compiler, or A-list game: a whole other 10 years work for an individual. 9front got around it by implementing just enough of a virtualizer to run OpenBSD and maybe Linux. I think that's great, but now I'm going for hardware without virtualization extensions.

DDG is backed by Bing? That's a surprise to me. Early on, they were saying they started the project when they saw storage costs had fallen to the point it was reasonable. But maybe they changed backends or maybe they made a half-truth. If there's any Big Duck Bubble, it's been okay for me. I'm not a deep searcher, but DDG turns up enough of the results I want close enough to the top.

I probably shouldn't be, but I'm highly amused by Chromium not clearing local storage in private mode. I may have a bit of troll in my DNA. Is it the type of storage the user has to approve before use, or can pages store a certain amount without approval?

eekee wrote:Thanks! It's great to know there are usable html-only search engines, Duck Duck Go never fully implemented their html-only page.

The URL https://html.duckduckgo.com/html works for me without JS pretty well.

eekee wrote:DDG is backed by Bing? That's a surprise to me.

I first read about it here. And indeed, they are admittedly use the Bing API, see here

We also of course have more traditional links in the search results, which we also source from multiple partners, though most commonly from Bing (and none from Google).

eekee wrote:Is it the type of storage the user has to approve before use, or can pages store a certain amount without approval?

No, there's no user approval needed for localStore, it just works, and it's not configurable under "Privacy & Security" / "Tracking Protection" either (but you can explicitly disable site data elsewhere). But because of a bug, Firefox did not respect the security settings for localStore for more than 6 years... Read this, that page has some useful Firefox and Chrome specific links too. For Chrome, you can also use an addon, like this one for example.

Cheers,
bzt

bzt wrote:The URL https://html.duckduckgo.com/html works for me without JS pretty well.

I just tried it. It would work for me almost always, I think. I tested its pagination; the 4th page had completely random results not fitting the query at all. However, I almost never get to the bottom of the first page before I get what I want, so I could use it. Overall, it's in about the same state it was... ooh... 5 or 6 years ago when my friend had trouble with the pagination, emailed them, and was told, "We don't support /html."

bzt wrote:they admittedly use the Bing API

I see, it's a blend; thanks. Sounds like a good idea, but it explains why tripadvisor results are so prominent whenever I search for places.

bzt wrote:No, there's no user approval needed for localStore, it just works, and it's not configurable under "Privacy & Security" / "Tracking Protection" either

Oh dear! Well, it's good there's an extension, at least.

Cheers!

EDIT: I just searched Qwant for "non-uniform rational b-splines" and it reported, "No results found". m) I got 3 big pages of results in DDG HTML and more in DDG regular. I also got results in Qwant regular. I've reported it to them.

I use Firefox, which has built in tracker-blockers, is that effective?

rizxt wrote:I use Firefox, which has built in tracker-blockers, is that effective?

That's a good question. Mozilla have pages on what's blocked, but those pages don't say what it doesn't block. Mozilla gets its block lists from disconnect.me:
https://disconnect.me/trackerprotection

If you figure out how to navigate the disconnect.me page, let me know. :s