Live data from Hacker News

Can Canonical count users without uniquely identifying them?

theravingrick.blogspot.com

1–6 of 6 posts

Re: Can Canonical count users without uniquely identifying them?

#3

Model numbers? Isn't that why UUIDs were created? Unique identifier, and anonymous: https://secure.wikimedia.org/wikipedia/en/wiki/UUID

"The customer didn't really want to use a unique identifier though, because though it was anonymous, the customer wanted to count computers, but unique identifiers are for tracking (following a user over time)."

Re: Can Canonical count users without uniquely identifying them?

#4
post #3

Model numbers? Isn't that why UUIDs were created? Unique identifier, and anonymous: https://secure.wikimedia.org/wikipedia/en/wiki/UUID

"The customer didn't really want to use a unique identifier though, because though it was anonymous, the customer wanted to count computers, but unique identifiers are for tracking (following a user over time)."

That seems to be a misunderstanding of the nature of a UUID. Simply replacing the model with the UUID provides no tracking in this use case, as it's a single purpose reference.

Re: Can Canonical count users without uniquely identifying them?

#5
post #3

Earlier quoted context omitted.

"The customer didn't really want to use a unique identifier though, because though it was anonymous, the customer wanted to count computers, but unique identifiers are for tracking (following a user over time)."

That seems to be a misunderstanding of the nature of a UUID. Simply replacing the model with the UUID provides no tracking in this use case, as it's a single purpose reference.

I can't figure out what exactly you are proposing here..

If you propose that they generate a new UUID on every ping the OS makes, then you'd be able to tell how many unique installations are active on any given day, but you wouldn't know how long they had been active, nor would you know what model of computer they were, nor would you be able to know how many were active for a longer time range such as a week or a month.

Re: Can Canonical count users without uniquely identifying them?

#6

Earlier quoted context omitted.

That seems to be a misunderstanding of the nature of a UUID. Simply replacing the model with the UUID provides no tracking in this use case, as it's a single purpose reference.

I can't figure out what exactly you are proposing here.. If you propose that they generate a new UUID on every ping the OS makes, then you'd be able to tell how many unique installations are active on any given day, but you wouldn't know how long they had been active, nor would you know what model of computer they were, nor would you be able to know how many were active for a longer time range such as a week or a mon…

Fedora, for example, already does this with UUIDs and the smolt project. A UUID is generated on installation and used for the 'ping'. The user opts-in at first boot, and the software runs periodically:

http://en.wikipedia.org/wiki/Smolt_(Linux)

The data is aggregated for reporting:

http://www.smolts.org/