Mozilla Data Project Is Not a Good Idea
I'm a fan of Michael Arrington and his work at TechCrunch, though I disagree with his assessment of Mozilla's new secret 'Data' effort.
The plan is basically to collect data from Firefox users (who opt in) in order to provide a data set on site popularity and user trends. It's an interesting idea and one that might help Mozilla, but IMHO it's not a good one for the broader marketplace for a few reasons.
1) The data will always be biased because it will only be for Firefox users
2) 'Hackers' will try to do 'bad things' with the data which could well provide personally identifiable information (sure Mozilla would do its best to secure users, but the point is they would be providing a new potential attack vector).
3) More data isn't always better. Every web server in existence has some form of log system which accurately measures real traffic. Adding yet another new statistics system only confuses an already confused marketplace.
4) A users 'History' file already tracks the data (though it doesn't publish it publicly...).
I personally like what Red Hat's Fedora project is doing with users statistics. Fedora (by way of its Smolt technology) tracks how many IP addresses actually connect to Fedora Update servers. With that data Fedora know how many 'active' Fedora installations it has.
How many active Firefox installations are there? Sure we know how many downloads, but wouldn't it be great to have real number on users too?
**UPDATE 5:41 PM EDT - I'm wrong on the Firefox installations issue. Mozilla's Asa Dotzler commented below (thanks Asa!!) that Mozilla does have stats on this now and that current users number about 170 million **
So YES, getting stats is a good thing. And YES Mozilla Data will be a solid effort at understanding what Firefox users may be doing. But NO I will not personally participate myself and while I'll comment on their Data (when it's available) I'll always take it with a grain of salt.