How we count our users
Starting with the upcoming Stable release, we will change the way we count our users, gradually dropping the use of the unique ID for user counting.
Click here to see the full blog post
Really nice insight into how it all works. Thanks Julien!
Thanks for the "backstage" information
QuHno last edited by
Are tripple boot (3 OS on the same computer) users counted 3 times?
Interesting insights. Thanks for the write-up, and I look forward to a more privacy centric counter.
If one of the copies is missing, it is set up again using the other available copy.
Does that mean if I was using some kind of standalone vivaldi on a USB stick*, and I plugged it in to someone else's computer, then some files would be written to their device, despite the fact that I'm running off a USB drive?
*I know, there's no official "portable" version yet, but hypothetically...
Also, do you think there would be any way to use an installation that is logged in to sync as a way to distinguish from repeated installs?
luetage last edited by
@julien_picalausa Since you're directly at the source now, what was our count for last week?
A couple of questions :
Could those number be more or less regularly published ?
I think it's more or less evident from the post, but can you confirm that no IP address is logged when the daily/weekly ping is done ?
I’m so happy you built this.
@quhno: Triple boot would be counted 3 times. From my point of view, that's pretty much the same scenario as if you were running 3 separate physical computers.
It would be nice to be able to detect that scenario to refine the numbers further, but the methods needed to do that would become quite intrusive and not really what we want to be doing.
@cqoicebordel: The official number is available at https://vivaldi.com/company/ .
As for IP addresses, we have never kept any permanent record of full IP addresses that we receive as part of the user counting request. We currently only store a truncated address that allows us to determine which region of the world our users are coming from. It would not allow us in any way to pinpoint a specific user. Since we are going to upgrade the server-side as part of this, I'll investigate with Claudia's help whether keeping this truncated IP address is even necessary.
@lonm: Currently, no: it is actually only done if the LocalState copy goes missing but not the other way around. This is one of the things that is subject to change though as we move forward. This is because when the server counts by unique id, it is really important to not have a single user of standalone go and "mark" a bunch of computers as their own. In situations where the server only counts requests, it is more important to consistently remember all the data associated with the last request time to be preserved more reliably and having two installations of vivaldi avoiding to interfere with each other on reporting because they have the same ID is not as big of a deal.
@julien_picalausa: Nice !
For the number on that page, do you have any idea how often it's updated ?
For the IP addresses, just a side suggestion : please make sure it's IPv4 and IPv6 reachable. We are starting to see endpoints without any IPv4 addresses.
NSANE last edited by
Over time, tinfoil enthusiasts have heavily critiziced Vivaldi due to this tracking ID. Hopefully this change will put that to rest, but I'm sure they'll come up with new ideas anyways...
luetage last edited by
@NSANE FOSS devotees will never fully embrace Vivaldi. Those are exactly the ones, who couldn't stand Vivaldi "phoning home". It's simply not the target audience.
lamarca last edited by
Apart from being an end user, what I like most about Sync is the background part. Thanks.
Last but not least, I can confirm the IPv4 issue mentioned in one comment above. We both are on Linux
Agreed, great insight.
Another point, I use almost always the Snapshot, sometimes one instance in Desktop1 and another in Desktop2. How is that handled? Or do you filter IP addresses.
Just thinking it might skew your results if there are many others like me.
QuHno last edited by
Over time, tinfoil enthusiasts have heavily critiziced Vivaldi due to this tracking ID.
No tinfoil hat needed.
A unique tracking ID is always problematic because someone else could misuse it. Definitely not implying that Vivaldi does it, but errors / bugs happen and someone might get hold of the database etc.
If server, databases and (web)firewalls are "free from bugs" it only means that nobody has found a previously unknown flaw. You can find only what you can imagine during security testing - and nobody has a perfect imagination.
i_ri last edited by i_ri
Distro. hopper installs new every few days. Reinstalling a windows machine creating first instance install happens so often these days. Desktop plus notebook user; android version will complicate this multiple-device single user. The nerd that runs ten machines. virtualbox users.
Most errors would tend to inflate rather than decrease the outcome?
You use the term target audience. The users in this forum are different than the main target audience if you ask me; the vivaldi users not in forum might be the ones who would object to id. Those here at forum are more willing to participate in a status quo for benefit vivaldi. So I think you speak to a biased group here. I would like to keep my unique id. I am not usual?
Can an old-fashion approach be considered? how about this?: beta [snapshot] Requires participation in sending information to vivaldi of use and crash and such... unique id with solid reporting. Stable users are Not required to participate.
In your current approach to backing out or recreating the id system of current, interim, interim, later ... Can that be set to the interim steps permanently with this differential for beta and Stable?
Securing vivaldi helps secure the future of vivaldi with safe acceptance of the suitability of terms. Thank you for modernizing and adjusting the safe promotion of vivaldi. and sharing the minutiae is fabulous. Thank you.
@julien_picalausa Thank you, more nice article to share.
RogerWilco last edited by
I am glad to see that the privacy related issues apparently are receiving a greater focus (again) and the current processes are being scrutinized. A thank you for that, I'll keep an eye out for further changes leading to/restoring more anonymity, which is sort of limited by other realities.
I'd still be willing to pay an annual license fee for the browser and be given the assurance that no phoning-home at all will occur. Yes, it's weird and sad at the same time - we are willing to pay to be left alone.
And although the GDPR isn't perfect, it'd be nice if this were to apply here as well, like it would have at the location of Tetzchner's former HQ in the EU.
Anyway, thanks for putting a stronger emphasis on privacy again.