Live data from Hacker News

How Google Chrome’s ad blocker works

ctrl.blog

71–80 of 326 posts

Re: How Google Chrome’s ad blocker works

#71

Earlier quoted context omitted.

Remember when Google claimed that paid inclusion was one of the worst sins a search engine could commit? One so bad that it made it claim other search engines were evil?

Now we have google shopping search entirely pay to play.

I've attempted to use that a few times, I won't even waste my time on that garbage search anymore

Re: How Google Chrome’s ad blocker works

#73
post #31

Earlier quoted context omitted.

Google may have pushed for this, but the standards were agreed upon by the ad industry: https://marketingland.com/industry-groups-self-regulatory-pr...

The standards should be set by independent consumer watchdogs or the competent (one would hope) authorities, depending on the industry to self regulate has not worked in over 25 years, I see no reason why it would magically start to work now.

The ad industry had no means or incentive to self regulate. That has changed.

The incentive is ad blockers. The means is "Coalition for Better Ads."

Can't trust them to fight against tracking. Can't trust them to eliminate malvertising. Can't trust them to unblock your website in a timely manner or give good feedback about what is wrong with it. Can expect them to keep obnoxious use of advertising down to stop the whole industry from collapsing.

Re: How Google Chrome’s ad blocker works

#74
post #37

If you want to go read the source yourself, I believe this is the implementation: https://cs.chromium.org/chromium/src/extensions/browser/api/... And the code reviews cover the feature are tagged with DNR: https://chromium-review.googlesource.com/?polygerrit=0#/q/Dn... Edit: this may just be the net blocking code.

Now I'm half tempted to serve up ads from a path in the form '/a{n}b/\d+' with sufficiently large n to bring the O(mn) std::string::find subpattern matcher to a crawl when it encounters my harmless 'src="/a{m}"' iframes. Preferably n > 32K/2 to blow through L1 cache.

Re: How Google Chrome’s ad blocker works

#75

Google's incentives here are still antithetical to my own, so I'm going to stick with uBlock. For example, on Google search, ads masquerade as legitimate search results. These used to be highlighted in light yellow, clearly marking them as different from organic search results. Now, they only have the word "Ad" in light grey text. This notice is on the right, intentionally where it will be overlooked when reading thr…

> I'm going to stick with uBlock

Yeah, me too, but I don't think Google hopes to convince adblock users to revert to an adblock-free experience.

They're probably hoping (as is indeed stated in the article) that by removing the biggest pain points they will stop new users from seeking adblocking solutions.

We'll see what happens, but it doesn't seem an unreasonable thing to hope for.

Re: How Google Chrome’s ad blocker works

#76
post #74
post #37

If you want to go read the source yourself, I believe this is the implementation: https://cs.chromium.org/chromium/src/extensions/browser/api/... And the code reviews cover the feature are tagged with DNR: https://chromium-review.googlesource.com/?polygerrit=0#/q/Dn... Edit: this may just be the net blocking code.

Now I'm half tempted to serve up ads from a path in the form '/a{n}b/\d+' with sufficiently large n to bring the O(mn) std::string::find subpattern matcher to a crawl when it encounters my harmless 'src="/a{m}"' iframes. Preferably n > 32K/2 to blow through L1 cache.

Which code are you looking at? I see it uses https://cs.chromium.org/chromium/src/components/url_pattern_....

Also, from what I gather from their docs, this is only for the ad-block-filter formatted lists. For their own super-secret-better-ads list, they use safe browsing lists which use a hash-some-then-phone-home approach IIRC [0]. I mean, even Mozilla that uses the safe browsing lists says at [1] that the internal documentation [2] is only available under NDA.

0 - https://developers.google.com/safe-browsing/v4/local-databas...

1 - https://wiki.mozilla.org/Security/Safe_Browsing

2 - https://mana.mozilla.org/wiki/display/FIREFOX/Safe+Browsing

Re: How Google Chrome’s ad blocker works

#77
post #20

I wouldn't trust a burglar to install my security system. I won't trust Google to block ads.

You can trust a burglar to install a security system that will keep other burglars out, but you can be 100% sure that you will not be able to keep the installer out, in fact you are just about giving them an incentive to rob you blind at the first opportunity. Google will do an amazing job at blocking competitor ads.

But when you're robbed you'll know exactly who did it. I expect Google will try very hard to appear to treat all ads equally, hence the focus on all ads on particular sites, instead of trying to make judgements on individual ads. If they don't, they will quickly suffer massive fines that actually hurt their bottom line every year, as well as lawsuits from competitors that they stand a chance of actually losing.

Re: How Google Chrome’s ad blocker works

#78

Earlier quoted context omitted.

Latest Firefox Quantum release available with faster, always-on privacy with opt-in Tracking Protection and new features[1]- 23 jan 2018 [1] https://blog.mozilla.org/blog/2018/01/23/latest-firefox-quan...

I'm aware of this, but it seems to be only basic tracker blocking and not even near as flexible and comprehensive as say uBlock. They need to expand this significantly to offer something equivalent to this Chrome initiative. At the very least, offer the ability to add custom blacklists and per domain rules.

Pretty sure they use the same list as uBlock (and pretty much everything else now, including Disconnect and EFF's privacy badger). My understanding is it's based on Disconnect.

If you want that kind of flexibility it's easy to add, which means there is no reason for Firefox to have that by default.

Re: How Google Chrome’s ad blocker works

#79
post #74

Earlier quoted context omitted.

Now I'm half tempted to serve up ads from a path in the form '/a{n}b/\d+' with sufficiently large n to bring the O(mn) std::string::find subpattern matcher to a crawl when it encounters my harmless 'src="/a{m}"' iframes. Preferably n > 32K/2 to blow through L1 cache.

Which code are you looking at? I see it uses https://cs.chromium.org/chromium/src/components/url_pattern_... . Also, from what I gather from their docs, this is only for the ad-block-filter formatted lists. For their own super-secret-better-ads list, they use safe browsing lists which use a hash-some-then-phone-home approach IIRC [0]. I mean, even Mozilla that uses the safe browsing lists says at [1] that the interna…

This is the call stack I'm looking at:

#0 https://cs.chromium.org/chromium/src/components/url_pattern_...

#1 https://cs.chromium.org/chromium/src/components/url_pattern_...

#2 https://cs.chromium.org/chromium/src/components/url_pattern_...

#3 https://cs.chromium.org/chromium/src/components/url_pattern_...

Re: How Google Chrome’s ad blocker works

#80
post #65

Earlier quoted context omitted.

>Which is it, the super secret consortium that won't even let you download the full list of bad sites, or Easy List/Privacy? Why don't they just use EasyList and report offenders there instead of this new consortium? The lists serve different purposes. EasyList is a lit of ad-serving URLs to block. The uh, consortium list is a black/white list of which sites don't follow the standards. The standards are not about bad…

EasyList is also a list of ad-serving sites to block and a list of elements to hide. Can you help me understand why the blacklist can't be just placed in a normal ad blocking list like EasyList? If it's by site anyways, I'm not sure the actual difference for the blocker itself. It's not like the EasyList standards can be automatically enforced by software either. I think "serve different purposes" could be rephrased…

They don't want to block ads on sites, unless the site is on the other list.

That is, if site foo.com and bar.com both serve ads, but foo.com is on the "blacklist", the ad blocking will be enabled on foo.com. However, ad blocking will remain disabled on bar.com.

It's not Google's intention to block all ads on all sites - only ads on "bad" sites. If they can reduce the number of adblock installs by reducing the overall number of invasive ads, they can keep their own ad business from folding.

Post reply on HN