[Discussion-Improvement] JA4H: take into the values of the headers too?

[unixfox]

Hello,

This is a discussion/potential improvement about the implementation for JA4H.

I read the paper and the python implementation, I found out that the values of the headers are not taken into account. Except for cookies and accept-language.

Is there a particular why it was discarded in the implementation?

A fingerprint should be as precise as possible.
For example, if you launch two instances of Google Chrome, one in version 131 and one in version 132. You still get the same JA4H hash even though the user agents are different, it's two different Chrome instances. That's the same outcome with running Google Chrome on Windows and Linux.

Would it be interesting to improve the current implementation of JA4H in order to add a new section with a sha256 of the header values? Named JA4H_e

Or maybe have a separate JA4H for this new implementation, like JA4Hv2 or JA4H2?

I'm really interested to hear your thoughts.

Thank you!