- From: Tom Gindin <tgindin@us.ibm.com>
- Date: Tue, 2 Oct 2001 07:54:49 -0400
- To: TAMURA Kent <kent@trl.ibm.co.jp>
- Cc: harada@prs.cs.fujitsu.co.jp, w3c-ietf-xmldsig@w3.org, toriumi@sysrap.cs.fujitsu.co.jp
I don't know if X509CRL's would be useful if attached to a real-time transaction or not, but they would certainly never be used for non-repudiation - you want the first one with effective time AFTER the transaction. You're quite right that there is no valid reason to use KeyValue if the key is certified, but somebody might use it if they were using a trust model similar to the original PGP one. Tom Gindin TAMURA Kent <kent@trl.ibm.co.jp>@w3.org on 10/02/2001 03:59:21 AM Sent by: w3c-ietf-xmldsig-request@w3.org To: harada@prs.cs.fujitsu.co.jp, w3c-ietf-xmldsig@w3.org, toriumi@sysrap.cs.fujitsu.co.jp cc: Subject: Re: Fw: Re:Call for Review: XML Digital Signature is a W3C Proposed Recommendation In message "Fw: Re:Call for Review: XML Digital Signature is a W3C Proposed Recommendation" on 01/09/18, "Harada" <harada@prs.cs.fujitsu.co.jp> writes: > In verifying, do you use X509CRLs which is created before verifying? X.509 CRL has information about "updated date" and "next update date". So we can assume the CRL attached to a signature is the latest until "next update date". In my opinion, we would use neither X509CRL elements nor KeyValue elements with signatures in practical systems. X509CRLs with signatures might be old, and we should not trust key information not in X.509 certificates. A signature should have an X.509 certificate or a key name, and verifier retrieve CRL from a local XKMS service. -- TAMURA Kent @ Tokyo Research Laboratory, IBM
Received on Tuesday, 2 October 2001 07:56:30 UTC