Re: XOP/MTOM specs remaining problems

> 4.2 Interpreting XOP Packages
> 
> Issue: Nothing is said about what to do when the
> reconstruction of the XML Infoset fails, or if the XOP
> Package is incorrect while still allowing a
> reconstruction of the XML Infoset (e.g., xop:Include
> has a child element).
> 
> Proposal: States that any failure is to be handled in
> an implementation dependant way.  States that any non
> fatal error may be ignored at the discretion of the
> implementation.

I have some concerns about this philosphically, in that I think it takes 
us closer than necessary to implying that we are specifying the behavior 
of processors as opposed to declaratively specifying an encoding. 

How about:

--------
4.2:  Interpreting XOP Packages

This section specifies the means by which the original Infoset can be 
reconstructed from a XOP package that has been prepared according to the 
rules of 4.1 Creating XOP Packages. 

Note:  conventions or error reporting mechanisms to  be used in processing 
 multipart packages that incorrectly purport to be XOP Packages are beyond 
the scope of this specification. 
--------

The essential difference here is that instead of normatively specifying 
that processors handle this in an implementation dependent way, we make 
clear that anything beyond the processing of correct XOP packages is 
beyond the scope of our specification.

Noah

[1] 
http://www.w3.org/2000/xp/Group/3/06/Attachments/XOP.html#interpreting_xop_packages

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------