Are they passing on search terms for previous searches or just the current search?

In your (frankly terrifying) example, would the site have had to have ranked for "naked bieber pics", showing up in the results and then you click it to pull the search terms? Because that would seem to be pretty innocuous.

Just the current search, but the problem is millions of pages run third-party ads, and most are run by very few networks. These ad networks can aggregate the terms and then they have previous searches as well.

Doesn't Incognito browsing, like present in Chrome, remedy all these worries?

People find ways around it, for example for a while Flash didn't respect incognito so ad networks just used their cookies instead (I think at one point html5 storage was being used too, but could be wrong).

There's been various HN posts about how people circumvent browser privacy settings to track you across websites and even tricks such as generating an almost unique key by hashing together browser, fonts installed, IP address, etc.. I'm not sure of the current state of play in that war though.

I guess if you really don't want people to track private searches with your accounts, use a whole other browser where you're not logged into Google in incognito mode.

not if you are logged in using your google account

Pedophilia is illegal in most countries, I'd be happy if ad networks gather data about your interest on young boys and pass it to the feds! Beasts!

Edit: judging by the downvotes, I guess a good bunch of hackers like to live in atrocity!

I think the downvotes are because you're missing the point so hard that they're assuming you must be trolling.

He who would trade liberty for some temporary security, deserves neither liberty nor security (via Ben Franklin)

You got the oft-excluded part of the quote right -- temporary security -- but the actual quote is, "he who would trade Essential Liberty for some Temporary Security, deserves neither Liberty or Security."

Many people omit the word, "temporary". The suspicious part of me says they do it so their (mis-quoted) statement appears more bold and uncompromising.

When you click a link, you usually send the refferrer url to the site you go to. Naturally google's search terms are a very easy to parse and a well known url. So when you click a site's link they know that you came from google and that you searched for xyz.

Duck duck go actually redirects you in a way so that the referrer is not known to the site. They all think you came from duckduckgo.com. DDG also keeps no logs. Basically your own activity online cannot be used against you.

Now, privacy aside, I really like DDG for other reasons, like search relevance, continuous scrolling, zero-click info, etc. So its kind of win-win for me.

I am fairly sure that the issue is the browser passing on the referrer, not that Google is making websites aware of any additional information. I might be wrong, of course; there's a lot of FUD about this floating around.

The browser does it, and did it long before Google existed.

That doesn't mean Google couldn't or shouldn't prevent the browser from doing it.

When this all started, no one envisioned ad networks and data companies aggregating all this personal information, or even that search terms would be a central part of the Internet.

So now that we know what is going on, why allow this personal information to leak? As far as I can tell, the only reason is so Webmasters can do better at Google SEO. And that reason can be wholly mitigated through the use of Google's Webmaster Tools.

"And that reason can be wholly mitigated through the use of Google's Webmaster Tools."

No, Google's webmaster tools only provide a sampling of the data. We used to provide info for only 100 queries. Now we provide it for more queries, but it's still a sample: http://googlewebmastercentral.blogspot.com/2010/04/more-data...

Please don't make the argument that the data in our webmaster console is equivalent to the data that websites can currently find in their server logs, because that's not the case.

Hi Matt. Is there any technical reason why that tool can't provide full information? Clearly Google isn't opposed to not sending referrer data, given the existence of https search, but doesn't DDG have a point about third parties having access to the referrer data? (Full disclosure: I wrote this article)

Maybe not "when it all started" but we envisioned massive data gathering and analyzing pretty early on.

Yep - if you don't want people seeing your bieberqueries, you should use SSL Google:

https://encrypted.google.com/

... or if you're super paranoid, disable sending referrer in your browser.

Firefox (at least) will still send referrer info when you navigate to a https: site from encrypted Google. You can change this in about:config by changing network.http.sendSecureXSiteReferrer to false.

Source: http://kb.mozillazine.org/Network.http.sendSecureXSiteReferr...

This.

Its silly to rely on trusting the search engine provider to be secure, better be secure on the client side.

I'm curious if it's possible to write a Firefox plugin that would block the HTTP-REFERRER header for certain sites.

RefControl https://addons.mozilla.org/en-US/firefox/addon/refcontrol/

I'm using it since years and I don't understand all that referrer-privacy hype. Just don't send the referrer.

I recently started using RefControl to get around the nytimes.com block which comes up if the referer isn't google.

For privacy reasons, I set it so it forges the referrer to be the root of the destination site if the domain is different too. So when I go to google.com and click on http://www.example.com/foo, the referer it sends is http://www.example.com/ instead of http://www.google.com/searchterm.

I haven't found any sites which this breaks yet...

You do realize this breaks the ability of www.example.com to track its traffic? They have now lost the information as to how their search rank is performing on Google, and have acquired incorrect information that they think you typed in www.example.com directly.

You're entirely within legal rights to do this, of course. But if everyone did, then sites would have no idea where their traffic is coming from. Morally it is a bit questionable: your browser is blatantly and directly lying to www.example.com about how you got there.

You think a website has a right to track where I came from. I think a user has a right to privacy.

If it is from an adwords campaign, you get the search terms.

Even if the site owner doesn't know that the referring url contained the search term "naked beiber pics" they still know you visited their page which most likely contains "naked beiber pics".

Exactly, the privacy issue around http referrers is a red herring.

You're free to opt out at any time:

http://www.google.com/privacy/ads/

AFAIK this doesn't shut off the search leakage in the referrer header.

That's a feature of how the internet works at an architectural level. How exactly is that Google's fault?

If you don't want referral headers to be sent from one website to another, use HTTPS, or turn it off in your browser.

Unfortunately, normal people don't know how to turn them off or use the https version. Google could do a lot of things to make this easier.

The best would be to just prevent ad networks from collecting this personal information altogether. The only reason I've heard against doing so is that webmasters need search terms for SEO, but they can get them via Webmaster tools.

Failing that, they could make a setting in the http version that turns off the headers. They could also make a setting that defaults to https. And they could make https easier to get to. Currently you can't even do https://google.com/ (you have to type in www.)

> Google could do a lot of things to make this easier.

And so could Microsoft, Mozilla, Apple, and Opera.

As others have mentioned, this is a good marketing move, but it's not clear to me that you actually care that much about this issue.

Having spoken to Gabriel offline, because I was building my own search engine, he cares quite a bit about this issue. A ton of us in tech clain to care about privacy, but very few of us put their balls (and actions) where their beliefs are. DuckDuckGo's privacy policy has been in place for quite some time. And, it's only been recently that he's advertised it.

Anyone who really cares can - and will - use HTTPS.

EDIT: It appears that at least one person disagrees with me. So, a quick question: Which takes more effort for the end user, switching to a new search engine or switching to https?

In the last few weeks, I've talked to a lot of people that care. And in general people do care: http://blogs.wsj.com/digits/2010/12/21/web-surfers-troubled-...

I can tell you that normal people who care also don't know about HTTPS or how to use it. I'm all for Google making it easier to do so, or otherwise stop this personal information from leaking. See http://news.ycombinator.com/item?id=2122431 for some ideas.

Also, FWIW, HTTPS alone doesn't solve the issue completely, as HTTPS->HTTPS traffic still leaks.

Actually, with Google Instant, the search terms aren't in the URL anymore - does it still leak?

Yes. See http://news.ycombinator.com/item?id=2063961

Not if you use Opera.