review of 2.1, comments

Title: review of 2.1, comments

Following up with the review of 2.1, tagged from Matt:

(pdf, line1535 ) Propose a requirement a little stronger, that the MSH actually does not choke on this:

Replace:

"It is recommended that sending MSHs act liberally in allowing..."

with:

"The presence of a SyncReply element in synchronous responses is meaningless for the MSH, although

it is not an error case: an MSH MUST process such a response as if the element was absent. It is RECOMMENDED 

for an MSH to not include SyncReply when sending a synchronous response."

(pdf, line 2675) add the requirement that "the SOAPAction header field value (which MAY be "ebXML") MUST be

a quoted string in all cases, even if an empty string." (otherwise that may hurt interoperability).

(pdf, line 2677) Propose a requirement a little stronger, that the MSH actually does not choke on this:

(

Replace:

"Further, it is recommended that sending MSHs act liberally... SOAPACtion..."

with:

"Synchronous responses over HTTP may contain a SOAPAction header field.  The receiving MSH MUST process 

such a response as if the header field was absent."  (assuming the behavior of proxy/firewall in such case

would not affect the MSH receiving this ...)

Jacques