RE: Parameterization of Transforms

At 09:46 99/10/29 -0700, John Boyer wrote:
 >What you're proposing for function parameters is analogous to wrapping all
 >of the subelements in the item above in 'field' elements, as follows:

I think one of the things being confused here is the necessity of parameters
for Transforms [1], for .*Method Algorithms, and as to whether they must be
the same thing. But I'm going to assume they are not at the start of this
email and try to focus on what we are trying to say with respect to Digests.

[1] http://lists.w3.org/Archives/Public/w3c-ietf-xmldsig/1999OctDec/0203.html
[2]
http://lists.w3.org/Archives/Public/w3c-ietf-xmldsig/1999OctDec/0214.html 

 >Don:
 ><item>
 >	<field type="name">Television, 19 inch, Color</field>
 >	<field type="serialno">123456</field>
 >	<field type="unitcost">100</field>
 >	<field type="price">300</field>
 >	...
 ></item>

Yes, I definitely wouldn't like the above. I'd rather someone stick in their
own XML (which I'm advocating here where I can.) 

 >Joseph (sans xmlns):
 ><item>
 >	<field>
 >		<name>Television, 19 inch, Color</name>
 >	</field>
 >	...
 ></item>

Actually, I haven't tried to argue we need parameter in this thread, I just
want to avoid unqualified type/values. Otherwise, I've been trying to
_understand_ the parameter issue.

Forget the syntax for a moment, again what is it that we are saying?

The (object) pointed to by ObjectReference has a location of
"http://ietf.org"
The (object) pointed to by ObjectReference has a Digest (object)
   Digest (object) has an Method (object)
        Method (object) has a literal value of "urn:nist-gov:sha1"
        Method (object) has a Parameter (object) with an open content model.
   Digest (object) has a Value (object)
        Value (object) has an encoding of "Base64"
        Value (object) has a literal value of "a23bcd43"

I am not fairly confident this is what we are trying to say (see [1] for the
DLG).

This could be represented as follows

<ObjectReference Location="http://www.ietf.org"> 
   <Digest>
        <Algorithm="urn:nist-gov:sha1">
            <Parameter Type="urn:ietf:length">32</Parameter>
         </Algorithm>
         <Value Encoding="urn:ietf-org:base64">a23bcd43</DigestValue>
   </Digest>
</ObjectReference> 

or to collapse it a bit more.

<ObjectReference Location="http://www.ietf.org"> 
   <Digest Algorithm="urn:nist-gov:sha1">
        <Parameter Type="urn:ietf:length">32</Parameter>
         <Value Encoding="urn:ietf-org:base64">a23bcd43</DigestValue>
   </Digest>
</ObjectReference> 

It is interesting to note that this is completely consistent with that
offered in [1] for transforms except it has an additional Value element
which don't need to be represented in transforms -- it is implicit -- though
some could conceivable want to represent the intermediate results of
transforms! John, what I believe you are disputing in either instance is the
use of the Parameter element (not the structure itself) because the above
could be skipped, or done use XML elements, I agree!

<ObjectReference Location="http://www.ietf.org"> 
   <Digest Method="urn:nist-gov:sha1">
        <length xmlsl="http://www.w3.org/1999/sig/length">32</length>
   </Digest>
   <Value Encoding="urn:ietf-org:base64">a23bcd43</DigestValue> 
</ObjectReference> 

[1] http://www.w3.org/Signature/Drafts/xmldsig-datamodel-19991029.gif

_________________________________________________________
Joseph Reagle Jr.   
Policy Analyst           mailto:reagle@w3.org
XML-Signature Co-Chair   http://w3.org/People/Reagle/

Received on Friday, 29 October 1999 16:09:49 UTC