And what about sites not using SSL, plenty of their owners are sued etc without the need to point the finger using any sort of SSL... I don't know if I should be shocked or not, but a reasonably large electricity company in Australia is collecting credit card details without using SSL, I don't know if this was a badly pointed URL or what I haven't really looked into it terribly hard yet... Who gets the blame then?
Well, I wouldn't login to their site for sure, and I would drop them a note telling them why I'm costing them more money by sending the payment by check in the mail.
I think the bigger question is, any form with any personally identifiable information should be protected, after all a lot of countries are building up privacy laws and such and I'm pretty sure a finger can be easily pointed if when they stuff up and get dragged through court without needing SSL to identify the company in question...
SSL is not well suited to court disputes since it doesn't really leave traces. Try S/MIME and then the situation is different. In many countries, some digital signature are already as good as your personal signature. But again, it is all a matter of trust in the CA, so don't count on your self signed cert having any value or legal standing.
In order for PKI to work certs must have some value (or we may as well just exchange public keys -- and let the governments of the world read all our SSL traffic through their proxied firewalls).
Lets face it, if some government body really wanted to do that they could already, they setup a Public Certificate Authority, get audited, get included in all browsers and such, then issue dual certificates and boom what good is PKI then? Well the easier option would be to get an existing CA to issue them certificates and we're all back to square one that SSL isn't going to protect from man in the middle attacks of that kind. For all we know they already are employing either or a combination of both methods already.
That's a good reason why you should be careful about which CA you trust!
If some government was doing what you propose, it would be possible to notice it - if anyone checked the cert chain during SSL connections manually (or even programmatically, with a bot). I'm not saying it would instantly get noticed, but it would get noticed eventually.
If it was a rogue CA, there should be a process to remove it. Hopefully it should lose its certification and simply be removed. If not, it would be easy to prove by collecting a number of the "proxy" certs under false identities, and contact the owner in the subject certs to see if they actually requested the cert and have the private key.
If a CA made an honest mistake and got taken advantage of, issuing certs to the wrong people, there is always revocation to take care of the problem after it is discovered (that would, of course, require that revocation be enabled for most people to work).
However, under your proposal, if everybody uses self-signed certs, there is never any way to do any meaningless identity check at all, so there is never any way to tell, now or ever, if your party is who they claim.
Note that nobody precludes you from using your own self-signed certs today. You just get a warning in the browser about the authentication, which I think is very fair. You can still use encryption if you are brave enough to ignore that warning.
_______________________________________________
mozilla-crypto mailing list
[EMAIL PROTECTED]
http://mail.mozilla.org/listinfo/mozilla-crypto
