On Fri, Nov 11, 2016 at 6:03 AM, Dimitris Zacharopoulos <[email protected]> wrote: > (something weird happened in the reply all. Re-sending). > > On 11/11/2016 3:45 μμ, Gervase Markham wrote: >> >> On 11/11/16 13:26, Dimitris Zacharopoulos wrote: >>> >>> Although this is very helpful so that people can better understand the >>> CA/B Forum's expectations and BR scope by introducing fresh language and >>> definitions (and make some definitions clearer), I believe that >>> proposing new definitions is a very radical move that would ultimately >>> cause even more confusion to CAs, auditors and relying parties. The best >>> place to discuss these definitions would be the CA/B Forum. >> >> Hi Dimitris, >> >> Can you explain what about Peter's proposals are radical? It seems to me >> to be mostly a clarification of existing intent, at least as far as >> Mozilla policy is concerned. >> >> Gerv >> >> > > I was referring to not changing the definitions for which most people are > currently familiar with and are included in the BRs. IMHO defining "CA" as a > Certificate with CA:true when the BRs define a "CA" as "an organization" is > a big step to make. I thought Peter's proposal included introducing these > definitions in the Mozilla policy to clarify the proper BR scope. My point > is that we need to be as aligned with existing BR definitions as possible to > avoid more confusion. If these definitions are to be improved/changed > (something we're already working on in the Forum's policy review WG), I > believe the CA/B Forum is the best place to discuss these changes.
Dimitris, I agree that the definitions I wrote may not align perfectly with those that the CA/Browser Forum Policy Review WG has been using. However, after discussions at the F2F, discussions about why disclosure of all CA Certificates in SalesForce (vs. all unique Issuer DNs) is important, and discussions with others experienced in running PKIs, I believe that it it critical we get this right. CA Operators, CA Key Pairs, Certification Authorities, CA Distinguished Names, and CA Certificates are all distinct things. Each of which certain relationships to the others, which are frequently one-to-many or many-to-one, so it is important that we get things clearly defined. This text is my attempt to start documenting those relationships. It it clear to me that we don't have a common understanding of the "WebPKI" and what organizations and people wishing to participate in the WebPKI should expect. It is also clear that we have an ever increasing number of organizations and people wishing to either bring their existing PKI into the WebPKI or start a new PKI to join the WebPKI. I think it is critical we help them understand how the WebPKI works and what to expect. In virtually every conversation I've had, someone gets surprised by how the WebPKI works. We need to fix this so there is a common understanding. Only then can we write effective requirements. Thanks, Peter P.S. I disagree that the CA/Browser Forum is the best place for this discussion. The WebPKI is larger than the CA/BF and members of the CA/BF have clearly said that many parts of the WebPKI are out of scope for the CA/BF, as it exists today. _______________________________________________ dev-security-policy mailing list [email protected] https://lists.mozilla.org/listinfo/dev-security-policy

