[LINK] BlueCava tracks individual devices, assigns creditworthiness

[rene]

On Thu, 2 Dec 2010 16:30:52 +1100, Birch, Jim wrote:

> Rene wrote:
>
>> My head hurts too, but imo it's the test that's arguably flaky/not
>> very helpful. ...
[...]
> The site grabs everything it that it can (that isn't blocked by
> NoScript, etc) like OS, browser version, fonts, plugins, and matches
> it to its DB.  There are about 1.2 million records in the db.  If
> you've got a locked down system or very standard system you might be
> quite common.   If you are unique, you get one in 1.2 million.  

Well, if that is the case, then I still reckon it's flaky because I fully 
expected to get told my browser 'footprint' is unique. I expected that in 
part because the site claims to account for installed fonts in determining 
what is a 'unique' footprint. Imo the probability of anyone else on the 
planet having exactly the same fonts installed as I do is zero or extremely 
close to zero. IOW, imo the chances of anyone else having just happened to 
have installed a couple of particular free fonts from the multitude 
available, and also still having a few particular (but not all) fonts 
installed that came with very old software programs, is almost impossible. 
Neverthless, I did not get told one in 1.2 million.

> If
> there's another matching record you get one on 640K, 1 in 3: 430K,
> etc.   If the site isn't tracking users in some other way, eg a login
> or a cookie, it can't tell if you are a repeat or a new user so it
> must treat you as a new user.  So the numbers go down each time.
>
> This is either a bad design fault, or a respectful security feature,
> ie, they're not keeping identifying info.

I suspect it's intended to be a respectful security feature, combined with 
the impossibility of keeping 'identifying info' about people who have run 
the test (given the site does not require login and heaps of people likely 
to run the test probably have cookies turned off - unless visitors have a 
static IP address and they record that - which even if they did would not 
prevent repeat visits from people who don't).

All up, I still think the test is flaky, not helpful.

Irene