RE: [dss] Timestamping

[Trevor Perrin <trevp@trevp.net>]

At 03:38 PM 4/3/2003 -0500, Robert Zuccherato wrote:

>Actually, I would prefer to just leave a general extension mechanism that
>could be used by later work.  For example, in Entrust's timestamping
>submission we included an <Extensions> element:
>
>   <element name="Extensions">
>     <complexType>
>       <sequence>
>         <any namespace="##any" minOccurs="0" maxOccurs="unbounded"
>processContents="lax" />
>       </sequence>
>     </complexType>
>   </element>
>
>I don't think there is any need to include any tags at this time unless we
>have a clear use for them.

Makes sense, but we need to consider how adding linking data into the 
<Extensions> element will interact with the timestamp's signature.

 From our earlier discussion, we decided (I think) that something like a 
<TimeStampInfo> in the Entrust proposal should be able to be inside an 
enveloping signature *or* inside a signed attribute.

But with a linking scheme, a signature may not be used at all.  Or there 
may be a signature *and* linking data - in Dimitri's explanation, it seemed 
like you'd want the signature to cover some of the linking data but not all 
of it (because creating the timestamp might be a 2-phase process, where you 
first get some linking data with a signature, and then later, at the end of 
a round, get the rest of the linking data).

So I guess as long as we allow the signature to be optional, and to cover 
some of the data in the <Extensions> element or not, we retain the ability 
to support linking schemes.


Trevor