Which scientific discoveries led to the creation of fingerprintjs?

[shenaseh]

I'm going to review a free* software for my university thesis, and the open source fingerprintjs looks appealing.
Is it documented how it works to generate a scored fingerprint id?
If its algorithm(s) are based on scientific articles, could you please provide those references? I would appreciate any advice in this matter.

---

* Free as freedom, not a free beer!

[makma]

Hey @shenaseh,

Is it documented how it works to generate a scored fingerprint id?

• All code is well commented.
• What do you mean by the scored fingerprint?

If its algorithm(s) are based on scientific articles, could you please provide those references? I would appreciate any advice in this matter.

A good scientific content for start might be EssentialFP: Exposing the Essence of Browser Fingerprinting

[shenaseh]

What do you mean by the scored fingerprint?

The degree of uniqueness of the identifier (fingerprint) generated from the user's device

A good scientific content for start might be EssentialFP: Exposing the Essence of Browser Fingerprinting

❤️

All code is well commented.

Yes, I am deciphering the code (:
Agreed. Well-commented. I just thought you might have some advice to assist me forward.

[makma]

The degree of uniqueness of the identifier (fingerprint) generated from the user's device

The library does not provide or consider such a metric. It just generates a hash from the provided signals without grading.

[shenaseh]

If not, then what is meant by Accuracy: up to 60% in the README page description?

[makma]

That's an estimation based on our internal data and might differ on the specific environment.
What we mean by the accuracy is explained in this article: https://fingerprint.com/blog/device-fingerprinting-accuracy/

[perpi]

That's an estimation based on our internal data

Do you mean the formula below?

might differ on the specific environment.

Would you explain more about how the environment has a role?

The library does not provide or consider such a metric.

Here, in the open source version, you have something called the Confidence. By choosing this word, how do you differentiate it from the word "accuracy"? And, what purpose can it be used for?
Also, why is it measured so simply and in terms of only one general variable (the type of user's operating system)?