What changed in 1.08?

[Doug Bunting <Doug.Bunting@Sun.COM>]

A bit earlier than I expected, I am going to call it a night.  So far, I am 
done except for one important editorial (I think) issue identified below 
and the CF4 work remaining.  I can do the remainder as soon as I hear 
whether my proposals below and in the "CF4 detailed proposal for V1.07" 
thread work for everyone.  On the second set, I think we need more text but 
am too tired to make suggestions.  I also think Jacques is travelling 
tomorrow.  Help, please.

What follows is a pretty detailed list of changes made since the 
teleconference.

Specification:
- a couple of very minor editorial things (likely all my fault in 1.07)
==> May to MAY in section 4.4
==> "reply and a new Reliable" to "the reply and new Reliable" in same section
==> "This enable the combination" to "This enables the combination" in same 
section!
==> remove a single space before a period in 4.4.2.1

- (MP8)
   remove B.6 and all mention of "other" or "client side" parameters
==> good news, I found no references to B.6 elsewhere in the specification 
text.

- (CF4)
   make a proposal to the group covering the "failure" case for group 
termination, finish CF4 base issue as well (do not resend case)
   update document to include these proposed changes
==> Jacques sent out a good proposal but I asked question I would prefer to 
have answered before I start work.  Please comment on that thread.

- (Technical issues as I go through action items...)
   remove Response@replyPattern attribute
==> I diligently removed the correct number of octets (characters) from the 
Content-length header of the affected examples, then realized the values 
were incorrect in any case.  Does anyone want to count the HTTP body part 
characters in these HTTP examples to set the Content-length headers 
properly?  No, not me though I can update the examples if someone cares 
enough to provide the counts.

==> One important issue: The removal of Section 4.4.1 means we no longer 
say *anything* about the first SequenceReplies or NonSequenceReply element 
within a Response element in the Response RM-Reply Pattern case.  I think 
we need to add

	"When the Response occurs under the Response RM-Reply Pattern, the first 
element in this list describes the status of the received Reliable Message. 
  In this case and the SequenceReplies element is used, the first contained 
ReplyRange element will include the received Reliable Message within its 
range."

after the bullets in Section 4.4.  This will allow a larger ReplyRange than 
just a single message but does place the Reliable Message's status in the 
first such range.  If someone in this group prefers a single-message range 
for this case or some other handling, speak up please.

- (Technical issues as I go through action items...)
   change schema to make SequenceReplies and NonSequenceReply into a 
single, mixed list of length 1 or more
==> Though the wording was already technically correct, I described the 
schema in a bit more detail.  Have a look at Section 4.4 for those minor 
changes.  I also mixed up the list in example 6 to show what is possible. 
Then again, I noticed example 9 was invalid under the old schema ;)

- (Re: [wsrm] Full agenda for WSRM TC teleconf 8/3/04)
   rename No[nt]SupportedFeature fault to FeatureNotSupported in both 
schema and specification
==> done


wsrmfp-1.1.xsd:
- (MP8)
   remove B.6 and all mention of "other" or "client side" parameters
==> Sunil has uploaded improved schema


WS-Reliability-1.1.xsd:
- (Technical issues as I go through action items...)
   remove Response@replyPattern attribute
==> Sunil has uploaded improved schema

- (Technical issues as I go through action items...)
   change schema to make SequenceReplies and NonSequenceReply into a 
single, mixed list of length 1 or more
==> Sunil has uploaded improved schema

- (Re: [wsrm] Full agenda for WSRM TC teleconf 8/3/04)
   remove InvalidMessageHeader fault from schema
==> Sunil has uploaded improved schema

- (Re: [wsrm] Full agenda for WSRM TC teleconf 8/3/04)
   rename No[nt]SupportedFeature fault to FeatureNotSupported in both 
schema and specification
==> Sunil has uploaded improved schema

- (Re: [wsrm] Full agenda for WSRM TC teleconf 8/3/04)
   remove global ReplyTo element and change references to that element to 
instead be local elements referring to the appropriate type (a schema-only 
editorial change)
==> Sunil's most recent schema improves on this but leaves ReplyTo global

- (Re: [wsrm] Full agenda for WSRM TC teleconf 8/3/04)
   editorial license to highlight BareURI element (presently at bottom)
==> Sunil has uploaded improved schema that makes BareURI more visible

thanx,
	doug