Re: summary of soapbuilders discussion about inlining multirefs

 Asir,
 To clarify, I don't feel XMLDSIG need use SOAP Encoding or any
mechanisms thereof, what I meant was that XMLDSIG may need
forward references. I don't know whether XMLDSIG uses SOAP
Encoding right now, and I did suggest that other referencing
mechanism could be used instead of SOAP Encoding in XMLDSIG.
 As I indicated in other mails, I believe forward references are
not necessary for serialization inside one serialization tree.
But others said that for example the members of a struct may be
reordered and such a reordering could create forward references.
 Personally, I can't see any scenario where the members would be
reordered without recomputing the serialization, which in turn
(if we forbade forward refs) would serialize with backward refs
only. So I myself would not object to forbidding forward
references, although the text would have to be very clear on its
meaning. Also, it must be noted that this change would require
references between headers and body to be serialized in headers,
for example in a non-mU header targeted at .../none, _and_ this
header would have to precede all the other headers that reference
the data. This might be hard to implement. 8-)
 This is related to problem 2) in my summary. If we accept the
solution 2b), forward references may be forbidden easily, IMHO.
If we accept solution 2c), forward references may go if we add to
2c that the referenced data must appear before the references.
Hard to implement, I think.
 Best regards,

                   Jacek Kopecky

                   Senior Architect, Systinet (formerly Idoox)
                   http://www.systinet.com/



On Fri, 26 Oct 2001, Asir S Vedamuthu wrote:

 > <snip> [1]
 > "There were various points that people were pointing out:
 >  1) maybe we should also disallow forward references
 > ..
 > Now let me detail the points.
 >  1) Some people felt forward references might be bad, other felt
 > my original proposal disallowed forward references. I propose to
 > keep forward references because they allow references from
 > headers to body, which might be necessary for things like
 > XMLDSIG, although any other referencing mechanism (most probably
 > XML IDREF) could be used instead of SOAP Encoding referencing."
 > </snip>
 >
 > I guess you are referring to id and href mechanism for representing
 > multi-references.
 >
 > In your proposal for Issue # 30 [2], you said that
 >
 > "remove the mentions of the attribute information items 'id'
 > and 'href' from sections 2 of both parts of the spec, for these
 > are encoding-specific attributes,"
 >
 > OK, so id and href are encoding-specific attributes.
 >
 > First, a question of clarification - you used the word necessary; why would
 > XMLDSIG use SOAP Encoding specific attributes for referencing? Are they
 > currently using it?
 >
 > Second, like others, I also believe that forward references are bad. 'Cos,
 > there is very limited payoff for a big cost.
 >
 > [1] http://lists.w3.org/Archives/Public/xml-dist-app/2001Oct/0231.html
 > [2] http://lists.w3.org/Archives/Public/xml-dist-app/2001Oct/0244.html
 >
 > Regards,
 >
 > Asir S Vedamuthu
 >
 > webMethods, Inc.
 > 703-460-2513 or asirv@webmethods.com
 > http://www.webmethods.com/
 >

Received on Friday, 26 October 2001 08:58:39 UTC