Slashdot Mirror
Google Backtracks on Chrome Modifications That Would Have Crippled Ad Blockers (zdnet.com)
Google has changed its stance on upcoming Chrome Manifest V3 changes as benchmark shows they lied about performance hit. Catalin Cimpanu, writing for ZDNet: A study analyzing the performance of Chrome ad blocker extensions published on Friday has proven wrong claims made by Google developers last month, when a controversy broke out surrounding their decision to modify the Chrome browser in such a way that would have eventually killed off ad blockers and many other extensions. The study, carried out by the team behind the Ghostery ad blocker, found that ad blockers had sub-millisecond impact on Chrome's network requests that could hardly be called a performance hit. Hours after the Ghostery team published its study and benchmark results, the Chrome team backtracked on their planned modifications. At the root of Ghostery's benchmark into ad blocker performance stands Manifest V3, a new standard for developing Chrome extensions that Google announced last October.
I'd still want an ad blocker. It's optional anyways. Don't like the performance? Don't install the extension.
It should be pointed out that the new API being proposed will actually make adblocking more efficient. The new API allows extensions to configure Chrome with a set of rules that specify what URLs to block. These rules would be evaluated by Chrome, in native code. In contrast, the method currently used by adblockers is an API that simply calls a snippet of extension-provided Javascript on each network request. While this can be fast if the adblock extension author codes it well, it clearly will never be as efficient as matching rules implemented in native code.
The source of the controversy is that the new API has an upper bound on the number of rules that can be provided: 30K So an adblocker could not add more than 30K adblocking rules. The message from Devlin Cronin (lead engineer for the extensions team) implies that they think that should be enough for good adblocking, if rules are managed properly. He cites research which finds that 90% of the 60K EasyList rules aren't useful, which implies that the same adblocking benefit would be achieved with 6K rules. Clearly, huge rule lists will be slower than small rule lists. Cronin says the rule list size limit will be increased, though there will still be a limit.
For anyone interested in the technical details, here's the documentation of the new API: https://developer.chrome.com/e... and of the existing API: https://developer.chrome.com/e...
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
Really, the goal here seems to be to make adblocking safer and more efficient, better for users, not to kill it.
That presupposes that adblockers are inefficient now. But, as linked to in the summary, simple measurements show that isn't true.
And I use an adblocker, advise everyone I know to use adblockers, and would switch to a different browser if Chrome were to block adblockers.
Adblockers work better in Firefox. You should advise everyone to switch to Firefox now instead of waiting. uBlock Origin uses WebAssembly in Firefox for better performance, but Chrome does not allow this yet.
So don't use Chrome. Problem solved.
Good luck with that. Native applications built with Electron, such as Discord, Slack, Skype, Atom, and Visual Studio Code, all bundle a copy of Chromium. This is Google Chrome with a handful of non-free parts cut out, mostly related to video DRM and Adobe Flash Player. The use of Chromium in Electron encourages development of web applications that work only with Chromium. For example, the owner of a Discord server can upload images to that server for use as emoji, but clicking the upload button does absolutely nothing in Firefox. This encourages Firefox users to switch to Google Chrome (on Windows or macOS) or Chromium (on X11/Linux) to make web applications work again.
In addition, on a phone or tablet running Android 4 or later, Google Chrome has a RAM use advantage over Firefox because its HTML engine is always loaded. This is analogous to IE's advantage on Windows 98 through 8.1.