Latest revisions

[Greg Pavlik <greg.pavlik@oracle.com>]

Gentlemen,

I have gone through to resolve my open issues and also to attempt to 
tidy up the spec by flagging redundancy and content we agreed to remove 
at the F2F (for example, the rationalizations for nesting belong in an 
FAQ or the Primer).

Please read this revision and check to make sure you agree with the 
changes I have made. I have modified the context structure and added 
text on both addressing/service references and on getting the value of a 
context that is passed by reference. Make sure that you agree with what 
is in there.

However, I have some action items for you all:

1) Eric, Mark: please make you agree with the defintion(s) of 
activities. Make sure that all explanations of activities are consistent 
with reference to the execution environment and contexts.
2) Eric, Mark: Please look for redundancy: it's annoying and it makes 
the spec ultimately harder to read and maintain. I tried to eliminate 
things that I thought were repetitive and unhelpful, let's talk before 
we undo the deletes.
3) Simeon: I have made several changes to the schemas and XML instances 
-- the changes aren't hard to find -- they need a) to be backported to 
the real schema/wsdl that you maintain and b) validated. Note that the 
SOAP example is currently incorrect, as the proper namespaces are not 
imported. Also, let's revist whether we actually need the "generic" 
service-ref element. I'm starting to think not.
4) Speaking of namespaces, we no longer have a section that says what 
namespace prefixes refer to. Is this an oversight? Let's give this one 
to Mark.
5) Mark, can you just remove the getContents method from ContextService?
6) What exactly is returned when a URL in a pass-by-reference context is 
dereferenced? An XML document that contains the context structure as 
understood by the issuing authority? We should spell this out; if I 
recall, the ContextManager responds with a requested-context message.
7) Does anyone have an action to map the request-reply messages into a 
(normative) table as per Peter's F2F request? This is actually important 
as it allows us to avoid by normative rules requirements in the callback 
pattern when using WS-MD.

By what date are we shooting for a new draft spec?