On Monday, 3 July 2017 23:05:53 UTC+1, Jeremy Rowley  wrote:
> And it's hardly fair to deride my lack of understanding on what transitive
> trust entails in the digital certificate space considering it's outside of
> the usual trust paths, not defined in the standard RFCs, and not the same as
> the mathematical expression.

Perhaps you have confused transitivity with commutativity or one of the other 
simple properties. Transitivity is the property whereby if F(A,B) and F(B,C) 
then F(A,C), for example the "greater than" binary relation is transitive.

The previously undisclosed certificate chains to a certificate which chains to 
another certificate and so on until you reach one that is included in the 
programme.  The Mozilla policy isn't trying to do anything very fancy here, 
it's just spelling out how chains actually work, which is why it was a concern 
that you seem so surprised this policy applies.

It seems _especially_ strange to object for this certificate that you had never 
imagined it was necessary to disclose it, while perhaps half a dozen similar 
certificates in the same family were previously disclosed. I would suggest that 
rather than demonstrating DigiCert had no idea it was supposed to disclose 
them, it instead indicates that DigiCert's processes for actually ensuring this 
gets done are inadequate and so some random proportion of sub-CAs may go 
undisclosed entirely by accident. That's not good news.
_______________________________________________
dev-security-policy mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-security-policy

Reply via email to