Has anything happened on the CMS critical flag issue since our last
discussion? I was intending to discuss the issue on the s/mime list, but
somehow I do not think our discussion went onto the list.
I still think the best answer is define an ASN.1 attribute type that has the
capability to include a criticality flag per attribute. I understand that
such a solution would not be acceptable for attributes already defined by
s/mime v1 and v2, but would it be acceptable for new attributes with new
OIDs in v3? There could be a default rule that indicates absence of the
criticality flag means non-critical. If existing attributes definitions
need to be critical then a new OID could be allocated for them with include
the optional criticality flag indication as part of the attribute
definition.
I am forwarding some of our previous discussion for information to list
subscribers.
On Tuesday, January 06, 1998 6:15 AM, dpkemp(_at_)missi(_dot_)ncsc(_dot_)mil
[SMTP:dpkemp(_at_)missi(_dot_)ncsc(_dot_)mil] wrote:
This would have no mandatory effect on v2 user agents, but does enable
the enforcement of criticality in v3, and appears to be the simplest
fully compatible method of doing so.
I agree that this is fully compatible -- the question is whether or not
v2 user agents ignoring it is acceptable. If there aren't any
complaints about v2 clients ignoring it, then we can go ahead and draft
syntax for it:
CriticalAttribute ::= OBJECT IDENTIFIER
CriticalAttributes ::= SEQUENCE OF CriticalAttribute
Or more simply:
CriticalAttributes ::= SEQUENCE OF OBJECT IDENTIFIER
And the OID:
id-aa-criticalAttributes OBJECT IDENTIFIER ::= { id-aa 5 }
The proposed use that Phill had is interesting, but I wonder if there
are implications in the event that an authenticated attribute in a
message is in conflict with the capabilities of the signing certificate,
both of which may be marked critical. For instance:
Certificate has a critical extension that says "This certificate cannot
be used to sign contracts"
Message has a critical attribute that says "This is a signed contract"
Is this something to be concerned about?
Blake
--
Blake C. Ramsdell
Worldtalk Corporation
For current info, check http://www.deming.com/users/blaker
Voice +1 425 882 8861 x103 Fax +1 425 882 8060
On Thursday, January 08, 1998 4:51 PM, Ross
[SMTP:ross(_at_)jgross(_dot_)demon(_dot_)co(_dot_)uk] wrote:
Providing the sequence of CriticalAttributes can be used to point to the
object IDs of the critical attributes. But can there be more than one
occurrence of an attribute in the signature sequence? If so, can one
occurrence be critical and the other occurrence not?
That's a very good observation. I will think about this.
Perhaps another way we can do this is by partitioning the OIDs so that
all critical ones are under a particular arc? If there is a noncritical
version of the same attribute, it can go under the noncritical arc
also...
Blake
--
Blake C. Ramsdell
Worldtalk Corporation
For current info, check http://www.deming.com/users/blaker
Voice +1 425 882 8861 x103 Fax +1 425 882 8060
-----Original Message-----
From: Blake Ramsdell <BlakeR(_at_)deming(_dot_)com>
To: 'Ross' <ross(_at_)jgross(_dot_)demon(_dot_)co(_dot_)uk>
Date: Friday, January 09, 1998 12:23 PM
Subject: RE: CMS Critical flag for signed attributes?
On Friday, January 09, 1998 8:16 AM, Ross
[SMTP:ross(_at_)jgross(_dot_)demon(_dot_)co(_dot_)uk]
wrote:
That solution would work, although it would mean that many attributes
would
have two OIDs.
You're right, and there may be some confusion as to which is the "right"
one to use. This needs some more thought.
Blake
--
Blake C. Ramsdell
Worldtalk Corporation
For current info, check http://www.deming.com/users/blaker
Voice +1 425 882 8861 x103 Fax +1 425 882 8060