Gee, this list hasn’t been this active for a while — glad I can help get the blood flowing :-)
Re. all the banter on what certs should/should not be enforced, some of us really old people remember PGP (before GPG)…
One used to generate a PGP PKI cert independent (or mostly independent) of e-mail address (one had many accounts on many systems so pretty much ignored the “From:” and validated purely on the cert signature).
Then came “FirstName.LastName@blah” e-mail accounts which didn’t fit into the 8-character Unix username restriction. AnD uNiX pEoPlE dId NoT lIKe CAPS — so always typed in lower-case with no punctuation and run-on sentences that never ever ended…
Later came “FirstName.MiddleInitial.LastName.some-number.civ.US@blah”. The Exchange folks liked camel-case but the PKI folks all had broken CAPS LOCK keys so either always typed in lower-case OR ALWAYS SHOUTED IN UPPERCASE. (Try changing a Unix username to “FirstName.MiddleInitial.LastName.some-number.civ.US” — then the person becomes a contractor and their e-mail address changes!)
Then DISA took over the world and removed all association with one’s organization and issued everyone a “@mail.mil” address — and users never bothered getting a new DoD CAC or generating a new cert based on their new address (let’s just wait the three years until I need a new CAC).
So… There has been a real e-mail address vs. cert e-mail address mismatch for years due to more than one cause.
And, of course, now it doesn’t matter because there are so many cert mismatches on sending/receiving e-mails and accessing web servers that all one does is hit “ignore, accept risk” to get on with their lives…
Thank goodness I’m not cynical…
Basil
On Apr 25, 2017, at 12:25 PM, Paul Nelson < email@hidden> wrote: I agree with this. Mail clients should not prevent users from viewing or decrypting messages but should be able to present validation evidence.
On Apr 25, 2017, at 11:07 AM, Blumenthal, Uri - 0553 - MITLL < email@hidden> wrote:
So presenting the user with the facts and letting him decide seems the only *reasonable* alternative.
|