Hacker News Reader

Comment by Spivak | Hacker News Reader

The one add-on I'm going to cling to for the rest of forever is the Google search link fixer. It makes clicking on a result instant.

View on HN

I think the reason Google does this is to prevent the site you're landing on from knowing what you searched for through the HTTP Referer header.

That header is the user-agent's decision, I consider it an anti-user feature. Google (like most search engines) wants to know what you clicked on.

That's correct. Google uses Referrer-Policy [1] to prevent the full referrer from being sent to the destination website. The transition page is purely for tracking. The ping attribute [2] is meant to replace that transition page eventually, but it needs to be ubiquitous and equivalent in terms of features before Google switches to that method.

[1]: https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Re...

[2]: https://developer.mozilla.org/en-US/docs/Web/HTML/Element/a#...

I vaguely recall the `rel` attribute value "noreferrer" on the anchor element producing this behavior; is that not the case (any more)?

That's graduated into Referrer Policy, more or less, but it should still work.

Have you considered using DuckDuckGo or do you just prefer the results Google provides?

Duck duck go is terrible. I give it a try frequently, and it is the default search provider on my phone, but most of the time I get useless results and have to either do it again with google or give up and try searching the history on my desktop for what I needed. I don't know why PLAIN TEXT SUBSTRING MATCHING is so flipping hard. Yes, I know spam exists, but I would prefer a dumb search to a smart but useless one.

Looking at my own webserver logs, I often wonder why ahrefs or majestic do not start a search engine effort. They crawl me as often as bing or google do, and could easily get a foothold with a "less is more" approach to search.

Haven't used GS in years. In many cases, substring-matching is a far superior search tool. Multiple reasons it might not be available (cycles for one), but it'd be great to have a second, 'advanced search' layer (like e.g. archive.org and worldcat have) available.

Agreed that DDG's 'helpful guesses about what you're looking for' generate about 50% bad results when you know EXACTLY what you want, and so you're stuck shaping your terms with pluses, quotes and minuses. Price to pay to avoid Google worth it to me.

You know what, you nailed it. I do not want helpful guesses from a machine. :) I liked them better when garbage in made garbage out and people knew not to trust the machine blindly.

Wow. I didn't notice it happening, but I've become previous-gen tech...

There's more to a search engine than crawling. Actually storing that data in a system that permits real-time searches based on arbitrary keywords is fairly expensive. Those services are only interested in creating a page link graph, which is a lot cheaper to operate.

Reply on news.ycombinator.com