Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Each Firefox download has a unique identifier

ghacks.net

Post reply on HN
Print view View original

51–60 of 480 posts

Re: Each Firefox download has a unique identifier

#51

Post by SahAssar »

bayindirh wrote: post #38

> A quick check of Chrome installers returned identical hashes each time. OK, however, are we completely sure that Chrome installer doesn't generate this token on launch and talk with the mothership? This sounds like whitewashing Chrome just to increase the impact of the article or push Chrome or both. Like Chrome is not tracking me in and out of the internet and in the kitchen making tea and noting its brand and rep…

The way firefox does it can connect the downloading session with the running session. You can argue with the value or validity of that, but it seems like the chrome installer cant do that, which is nice.

As for why it's in the article I think it's valuable to include it since if chrome was doing it too it might be seen as just "normal", but now it seems even more weird that firefox which is supposed to be the privacy alternative is tracking something that google is not.

Re: Each Firefox download has a unique identifier

#52

Post by Barrin92 »

klntsky wrote: post #17

> This will allow us to track which installs result from which downloads to determine the answers to questions like, "Why do we see so many installs per day, but not that many downloads per day?" What value does Mozilla see in being able to do that?

the reverse case might be more interesting. Many downloads but few install follow throughs may suggest the installation process is to cumbersome or something along those lines.

Re: Each Firefox download has a unique identifier

#53

Post by LordDragonfang »

bayindirh wrote: post #38

> A quick check of Chrome installers returned identical hashes each time. OK, however, are we completely sure that Chrome installer doesn't generate this token on launch and talk with the mothership? This sounds like whitewashing Chrome just to increase the impact of the article or push Chrome or both. Like Chrome is not tracking me in and out of the internet and in the kitchen making tea and noting its brand and rep…

>OK, however, are we completely sure that Chrome installer doesn't generate this token on launch and talk with the mothership?

That wouldn't give any information about where/when you got the installer from, which is the topic of this article. Doing so would be impossible without embedding information in the exe (which would change the hash).

While I agree that it's a little weird to specifically note it for Google of all companies, the relevance to the article is that Chrome isn't engaging in this specific type of tracking.

Re: Each Firefox download has a unique identifier

#54

Post by LordDragonfang »

causality0 wrote: post #48
ordu wrote: post #14

They seem to be trying to gather a lot of telemetry to measure how they can boost popularity of Firefox. I wonder did they tried to measure how the measurement itself influences popularity? Social measurements are like quantum ones, they change reality. There was a funny story of a Hawthorn Experiment[1], which tried to find ways to boost productivity but at the end managed to state just that the very attempt to cond…

Well they've spent a decade trying the "be more like Chrome" method. I suggest they try the "be more like Firefox from when Firefox was successful" method.

Firefox was successful when it was the alternative, better, option to the dominant Internet Explorer. Now the dominant browser is Chrom(e|ium). The two scenarios are very different.

Re: Each Firefox download has a unique identifier

#55

Post by estaseuropano »

ordu wrote: post #14

They seem to be trying to gather a lot of telemetry to measure how they can boost popularity of Firefox. I wonder did they tried to measure how the measurement itself influences popularity? Social measurements are like quantum ones, they change reality. There was a funny story of a Hawthorn Experiment[1], which tried to find ways to boost productivity but at the end managed to state just that the very attempt to cond…

That seems like a very harsh interpretation. Very few people will care whether their specific download is tracked. I do honestly wonder how that adds vakuento Mozilla, but no one will not use Firefox due to this- especially as every single alternative is much worse than Firefox on such metrics.

Re: Each Firefox download has a unique identifier

#56

Post by oh_sigh »

gjsman-1000 wrote: post #13

How much R&D do you think Firefox squandered on making a custom installer generator for every download and being unable to cache the files on a cheap CDN?

Not much? I did exactly this when I worked on a really popular P2P file sharing client (at one point estimated to be installed on >15% of all PCs worldwide). It even improved our actual installs, but that is probably about just using an ultralight weight installer rather than having a tracking ID integrated into it. It literally took me a week. Granted, things were really fast and loose back then. It would probably take me 2 years and a team of engineers to do a similar thing at my current FANG job.

Re: Each Firefox download has a unique identifier

#57

Post by delusional »

ordu wrote: post #14

They seem to be trying to gather a lot of telemetry to measure how they can boost popularity of Firefox. I wonder did they tried to measure how the measurement itself influences popularity? Social measurements are like quantum ones, they change reality. There was a funny story of a Hawthorn Experiment[1], which tried to find ways to boost productivity but at the end managed to state just that the very attempt to cond…

I think the physical analogy you're thinking about is "the observer effect"[1]. And it's actually a pretty much universal problem in physics, not just quantum mechanics.

[1]: https://en.wikipedia.org/wiki/Observer_effect_(physics)

Re: Each Firefox download has a unique identifier

#58

Post by Tijdreiziger »

whatshisface wrote: post #2

Package manager installations wouldn't have this problem because everybody gets the same copy of the same binary and associated files, rights?

Chocolatey and WinGet don’t.

Looks like Chocolatey gets the binary from download.mozilla.org [1], while WinGet gets it from download-installer.cdn.mozilla.net [2] (which looks to be the HTTPS repository mentioned in the article, thus being exempt from tracking?)

[1] https://community.chocolatey.org/packages/Firefox#files

[2] https://github.com/microsoft/winget-pkgs/blob/master/manifes...

Re: Each Firefox download has a unique identifier

#59

Post by causality0 »

LordDragonfang wrote: post #54
causality0 wrote: post #48

Earlier quoted context omitted.

Well they've spent a decade trying the "be more like Chrome" method. I suggest they try the "be more like Firefox from when Firefox was successful" method.

Firefox was successful when it was the alternative, better, option to the dominant Internet Explorer. Now the dominant browser is Chrom(e|ium). The two scenarios are very different.

Precisely. Firefox is never going to defeat Chrome in the "being Chrome" category. If it wants to exist as more than a tool for Google to avoid antitrust lawsuits, it can't keep playing that game. It has to differentiate. Privacy is not differentiation because it's invisible and HN commentators are 90% of the people who care about it. I want the sense of power back. I want the feeling that Firefox gave me a decade ago that my browser behaved exactly the way I wanted it to and nothing about it ticked me off because if I didn't like it I could just change it.

Nowadays using Firefox feels more like holding a political demonstration in an empty room than using the finely-tuned instrument I once had.

Re: Each Firefox download has a unique identifier

#60

Post by Tijdreiziger »

bayindirh wrote: post #38

> A quick check of Chrome installers returned identical hashes each time. OK, however, are we completely sure that Chrome installer doesn't generate this token on launch and talk with the mothership? This sounds like whitewashing Chrome just to increase the impact of the article or push Chrome or both. Like Chrome is not tracking me in and out of the internet and in the kitchen making tea and noting its brand and rep…

Chrome has the X-Client-Data header: https://github.com/bromite/bromite/issues/480
Post reply on HN