On 09/27/2016 12:17 AM, Hugo Landau wrote:
> But right now you can use Let's Encrypt without providing any means of
> out-of-band communication, so it's not safe to assume that such a
> mechanism — e. mail or otherwise — will be available.

This is exactly why we chose to write our Subscriber agreement such that
we can update without explicit opt-in.

> You mention RSS, but that requires a web interface to have some way of
> tying assent to a registration, which would require users to do special
> and unusual things with their registration private key to prove they
> hold it.

RSS is an extreme edge case of an edge case. We're imagining a CA that
require explicit assent to changes in the ToS, and additionally does not
require an email address for its customers, but *does* offer an RSS feed
for notifications about updates to their ToS. Even in such an edge case
of an edge case, they could create an individualized RSS URL for each
customer.

On 09/27/2016 10:00 AM, Jeremy Rowley wrote:
>> a) Click a link in the email, read the new ToS, and click "Agree" at
the bottom.
>> b) Click a link in the email, read the new ToS, and copy and paste
the new ToS URL into the customer's ACME client config.
> Why not allow both? We'd use (a) over (b), but if email isn't
required, then
> (b) is certainly an acceptable method of confirming ToS acceptance.

Because no one will use (b), and it adds unnecessary complexity and
fragility to the protocol.

To further emphasize my point that "it will never be used," here are
some other examples of CA ToS:

-------------------------
https://www.symantec.com/content/en/us/about/media/repository/ssl-subscriber-agreement.pdf

Company may revise the terms of this Agreement at any time for the
following reasons: (i) it becomes necessary due to applicable laws or
industry standards, including, without limitation, any change of the
foregoing; (ii) it becomes necessary for technological reasons when any
changes is made without materially degrading the Service functionality;
(iii) it becomes necessary to maintain the operation of the Service when
any change is made without materially degrading the Service
functionality; or (iv) changes are in favor of the Subscriber. Any such
change will be binding and effective thirty (30) days after publication
of the change on Symantec’s website, or upon notification to Subscriber
by email.


-------------------------
https://www.digicert.com/docs/agreements/DigiCert_SA.pdf

DigiCert may amend any of its (i) website and any documents listed
thereon, (ii) CPS, (iii) fees, (iv) privacy policy, or (iv) the
conditions under which you receive a Certificate.  Amendments are
effective upon the earlier of DigiCert’s posting the amendment on its
website or your receipt of the amendment.


-------------------------
https://www.comodo.com/repository/docs/ssl_certificate_subscriber_agreement.pdf

Comodo may amend this agreement, the CPS, the Relying Party Agreement,
the Relying Party Warranty, its website, and any documents listed in its
Repository at any time by posting either the amendment or the amended
document in the Repository.  Subscriber shall periodically review the
Repository to be aware of any changes. Subscriber may terminate the
agreement if Subscriber does not agree to the amendment. Subscriber’s
continued use of the Services after an amendment is posted constitutes
Subscriber’s acceptance of the amendment.
-------------------------

It should be clear from these examples that effectively the entire CA
industry, like the entire Web industry, and indeed corporations in
general, rely on auto-updating terms of service. As I said in my initial
post on this topic, I think that's not a great. But pretending it's not
so, or trying to litigate that broad policy problem in a spec for
automating certificate issuance is both ineffective and worsens the
protocol.

_______________________________________________
Acme mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/acme

Reply via email to