Live data from Hacker News

The Vivaldi browser will not support Google’s FLoC

vivaldi.com

211–216 of 216 posts

Re: The Vivaldi browser will not support Google’s FLoC

#211
post #114

Earlier quoted context omitted.

Actually, the lack of multi-account containers probably kills it for me. They've become such an important part of my workflow these days.

There's the profiles in Vivaldi. That's what I use and I found it to be better than Firefox's multi account containers because I don't have to choose the container each time I open the site.

you dont need to choose a container everytime. i have rules that certain sites always open on specific containers..

work stuff open on a specific work container..

facebook, twiter, youtube and fastmail each have their individual container.

there are even addons that allow you to have a individual container to be created for each site you visit..

Re: The Vivaldi browser will not support Google’s FLoC

#212
post #205

Earlier quoted context omitted.

There's also the EFF website, "Am I FLoCed?" https://news.ycombinator.com/item?id=26755313 - April 2021 (33 comments, 144 points)

Thanks! Added. I want to write software to let the community build these lists. Doing things like this is good practice for figuring out how the software can work.

That's a really cool idea!

Re: The Vivaldi browser will not support Google’s FLoC

#213
I just cannot see how this is anything other than a naked attempt by Google to get itself broken up. There's literally no reason someone writing a web-browser should be building spyware for advertisers. It's like if Microsoft had decided to start building those scummy Internet Explorer Toolbars in the 2000s.

Re: The Vivaldi browser will not support Google’s FLoC

#215

Earlier quoted context omitted.

This would be boring if it weren't so serious. It's damn time that the Googles of this world were brought to heel with legislation that has real teeth — i.e.: the levying of fines worth 20% upward of annual turnover as well as heavy fines levied on employees who concoct, design and or deliberately participate in introducing or maintaining these scams. Once employees are also targeted and held responsible for their ac…

If I do a search on https://www.google.com without sending SNI, I get a terse message placed in the CN X.509 attribute of the returned certificate: "Fix your client". It feels like arrogance and it seems like they have acquired a remarkable sense of entitlement. (The search still works.) As for speed, it would be interesting to have a "surfing contest", i.e., a web surfing contest that is a "browser challenge", pitti…

"As for speed, it would be interesting to have a "surfing contest"

Indeed it would be and as far as I know it's not been done recently either (anyway not with openly published results). Nevertheless, I would like to see say the top six browsers tested on heavy JavaScript websites that use lots of advertising. I'd reckon 100 sites would be plenty to get a definitive outcome of the speed differences.

Of course, it's unlikely happen given the fact that JavaScript is now so ubiquitous along with the all-too-common assumption that the web cannot cope (or would cease to work) without it—the logic being that there would be little point to the exercise. Moreover, if a such a test were conducted and nuking JS showed that most sites would be much faster without it then it'd still receive negative publicity as the many naysayers would point to website features that didn't work. And that's absolutely true some features will not—and with some, albeit still fewer again, will not work at all.

(Remember, this used not to be the case 20 years ago. So it's time we examined why JS is so critical now when it once wasn't. I'd make a substantial bet that the real or most substantive reason has little to do with improving the experience of website visitors (that is unless one considers that slowing websites down as a benefit) - and it's much more to do with benefiting website owners and their advertisers. One thing is for certain, we really do need to put a proper measure on this. Then, like limiting the advertising on television (as most counties already do with (essentially) the exception of the US), with accurate stats legislators could apply appropriate limits on advertising and data collection, etc. to web services. As things stand at present, it's nothing other than the Wild West.)

Of course, the test sites would have to be selected fairly and with care to provide a representative sample. I tend to gravitate to sites where JS is minimal and where killing JS makes little or no difference to viewing the core material (many technical sites fall into this category). Therefore, taking say the last 100 websites that I've visited would not be a representative sample. Incidentally, of the sites I visit about 95% work very well without JavaScript—so my default experience is to browse without JS enabled. Of the remaining 5% I've a choice of what to do, I can either toggle JS on and refresh the browser or instantly get out of there. On average I'd back out of about two-thirds of that 5% and toggle JS on for the remainder (all up, that's around 2% with JS).

This is where the naysayers miss the point. Whilst it's possible to default JS to off at the beginning of a new session I usually don't bother, so it's not unusual for me to start a new day's browsing with JS still toggled on from the previous session without me initially realizing it. However within a second or so I know that JavaScript has been left on by the way the browser responds (anyone who does what I do will know this experience well). Even if no JS content is obvious or visible on the page, if it's running tasks in the background then I will almost certainly notice it immediately as the browser is less responsive, paging becomes slower and there's a jerkiness to it that's not dissimilar to running programs under a Virtual Machine). I know JS aficionados either won't believe me or that they'll claim I'm exaggerating. Moreover, I know that me saying that I'm not exaggerating or attempting to prove it won't convince them one iota whatsoever, so there's no point in me trying—and then I don't have to prove anything anyway! If anyone really needs to know then he/she needs to do a quick survey of others who don't use JS and they'll soon find that those non-JS users will support what I'm saying.

"they have acquired a remarkable sense of entitlement. (The search still works.)"

Right, this is a good indication that the code that's actually been killed is the monitoring or spyware stuff and the coders/website owners are indignant about the fact that it isn't working (after all, given various error messages we get in these circumstances, we know they're checking to see whether the code in this section is running as intended or not). I regularly get messages such as 'Please turn on JavaScript to get the full benefit of this site [and or] certain features of this site will not work without JS' - and that's true—but they are the very ones that I do not want to work!

You often see the same nonsense with Android apps. Like the PC, I turn off as much Google as I can. I do not use any of Google's services, so on my rooted smartphones this means removing all Google apps (Google Play Services, Play Store, etc.). When the Google apps are removed many Android apps will complain to the effect: 'XYZ app will not run until you update Google Play Services' but they still work anyway (there are many exceptions that won't run and how to get around the problems is too off-topic to discuss here). Anyway, when apps still work but complain about Google Play Services it's almost a certain bet that it's the data gathering aspect of the Play Services that's relevant here—so I just ignore it (most people, even diehard Google fanatics, would prefer that data-gathering didn't work). FYI, always keep your old APK app files before you upgrade (and don't overwrite APK backups, as often that message doesn't pop up in older versions). That's to say, often Android updates contain nothing other than more highly refined code for spying, ipso facto, with Google Play Services and Play Store updates. (Of course, you'll get the security spiel/reason—as that's always the mandatory excuse/justification. Let me put it this way, you'll likely get both security updates and told so, what you won't be told is that Google's Android's Google Play Services 'receivers' schema has also been revamped to collect even more user data than previously (Android experts, I know that's a shortcut explanation/oversimplification but it'll have to do).

In my opinion a better and much more relevant and informative test would be to do a survey of a larger number of websites, say perhaps in excess of a 1000 or more of the most popular ones. Then analyze exactly what JavaScript is being used for on each site—that being to do a detailed statistical breakdown of JS's various functions on a per site basis then average this breakdown analysis across the whole web.

I've not conducted a rigorous test like this myself as it's a great deal of work to get meaningful statistics but I know from the number of times I've looked at the code from heavy JS sites that a very large percentage of the scripting goes into alerting every dog and sundry about what users are viewing and or doing. Whether one analyzes the website code or logs the destination IP addresses these pages call home to, one will see the huge amount of feedback that's being sent to many, many dozens of data-collecting websites. It's quite horrifying to watch it happening (I've been so alarmed that it's not the first time I've pulled the RJ-45 plug out of the PC socket to stop it continuing).

Re: The Vivaldi browser will not support Google’s FLoC

#216

Earlier quoted context omitted.

Why would that need ML? PageRank has worked for ages for this problem.

I would consider PageRank as a quite successful example of machine learning. Am I wrong?

The definition of what counts as ML/AI can be fuzzy, but I don't think the original PageRank algorithm would qualify as ML/AI under any definition: http://infolab.stanford.edu/~backrub/google.html
Post reply on HN