Counting number of users while preserving privacy
ryan at thunderbird.net
Fri Jan 18 03:57:24 UTC 2019
Well, I think that well-described anonymized collection with an opt-out
would make a majority of our users happy. Obviously we have to ensure
that our userbase understands we are not collecting any personal data.
I would be happy to explain this and ensure that folks understand
precisely what we are doing, and the insights gained would be well worth it.
On 1/17/19 4:54 AM, Ben Bucksch wrote:
> Yes. I've made such proposals, 7 years ago:
> (They were rejected by the Mozilla Metrics team, despite fulfilling
> all their stated requirements.)
> Eric Moore wrote on 16.01.19 06:31:
>> https://lwn.net/SubscriberLink/776327/e0cf49b9b9976c5a/ (some useful
>> background/history of the proposal)
>> https://fedoraproject.org/wiki/Changes/DNF_Better_Counting (the
>> current proposal)
>> Fedora appears to have spent a lot of thought about a new way to
>> count the number of their installations that does a much better job
>> of preserving privacy, and makes it hard for the data it returns to
>> be misused. It would avoid using a unique id or the ip address. It
>> includes a VARIANT_ID (an example would be "Workstation Edition"). In
>> our case the "count me" variable might have a very low cap such as 6
>> (implying 6 weeks or older) to minimize the risk of uniqueness.
>> Could something similar be done in Thunderbird when it requests data
>> from one of the Thunderbird servers?
>> We don't appear to have reliable data on the number of Thunderbird
>> users, did they explicitly install any add-ons (boolean, ignore whats
>> bundled with Thunderbird), and what version they are using. Something
>> modeled on Fedora's approach might be a useful way to get that while
>> avoiding most users (such as myself) instinctive wariness of
>> telemetry such as the toolkit telemetry. If we do this I suggest the
>> community manager explain beforehand (preferably before the first
>> beta that supports it) how it works, how it deals with privacy issues
>> and how to opt-out, in a post on Planet Thunderbird.
>> tb-planning mailing list
>> tb-planning at mozilla.org
> tb-planning mailing list
> tb-planning at mozilla.org
More information about the tb-planning