RE: [dss] Re: TSTInfoType

["Dimitri Andivahis" <dimitri@surety.com>]

I guess it's standard practice to have a catch-all element 
for schema extensibility, so something like:
  <xs:element name="Extensions" type=dss:AnyType minOccurs="0"/>
is a good idea, and could be used in different ways by 
different profiles.

Having said that, I would prefer <LinkingInfo> in its 
own separate element of type "any"

<xs:element name="Tst">
     <xs:complexType>
       <xs:sequence>
         <xs:element name="TstInfo" type="tst:TstInfoType" />
         <xs:element name="Extensions" type=dss:AnyType minOccurs="0"/>
         <xs:element name="LinkingInfo" type=dss:AnyType minOccurs="0"/>
         <xs:element ref="ds:Signature" minOccurs="0"/>
       </xs:sequence>
     </xs:complexType>
</xs:element>

so that humans can easily see that either the <ds:Signature> or 
the <LinkingInfo> element or both (if both are present) 
provide a crypto-binding between the timestamped data 
and the token.

Dimitri

> -----Original Message-----
> From: Trevor Perrin [mailto:trevp@trevp.net]
> Sent: Friday, November 07, 2003 6:12 PM
> To: Dimitri Andivahis; dss@lists.oasis-open.org
> Subject: RE: [dss] Re: TSTInfoType
> 
> 
> 
> 
> At 05:04 PM 11/7/2003 -0500, Dimitri Andivahis wrote:
> [...]
> >2. Proposed definition for the time stamp token <Tst>
> >
> >In place of the one in WD-04
> >   <xs:element name="Tst" type="ds:SignatureType">
> >that comes with the added requirement that a <TstInfo>
> >element MUST be included in the <Tst> element as
> >a <ds:Signature/Object> element, I propose the following:
> >
> ><xs:element name="Tst">
> >     <xs:complexType>
> >       <xs:sequence>
> >         <xs:element name="TstInfo" type="tst:TstInfoType" />
> >         <xs:element name="LinkingInfo" type=dss:AnyType minOccurs="0"/>
> >         <xs:element ref="ds:Signature" minOccurs="0"/>
> >       </xs:sequence>
> >     </xs:complexType>
> ></xs:element>
> 
> Could <LinkingInfo> be renamed something more generic? (I dunno, like 
> <Extensions>?).
> 
> It would seem better to me if this format was generically extensible, 
> without being tied to linking info.
> 
> 
> Trevor 
>