RE: Comment on the TRP overview and Requirements documents

["Dick Brooks" <dick@8760.com>]

If the FROM field is not usable you can still send an error response
indicating "Missing From data" on the socket used to send the message
(assuming a HTTP session with keep-alive). So it is possible to inform the
sender, even when there is no FROM value, of an error. Another possibility,
if trading partners use basic authentication for access control, it's
possible to identify the sender via the login username and this can help
determine where error messages should be sent. Either technique will work
(at least this is how we've been doing it in the gas industry for ~ 4
years).

The only time you cannot send an error message is (with HTTP anyway) when no
basic authentication is used and the socket closes before an
"acknowledgement" is returned.

Dick Brooks
http://www.8760.com/


-----Original Message-----
From: owner-ebxml-transport@lists.oasis-open.org
[mailto:owner-ebxml-transport@lists.oasis-open.org]On Behalf Of David
Burdett
Sent: Tuesday, March 07, 2000 6:43 PM
To: 'Duane Nickull'; ebXML-Transport@lists. oasis-open. org
Subject: RE: Comment on the TRP overview and Requirements documents


>>>I agree, however,  in the case where the document could not be parsed
(specifically the return address in the header) is this possible?<<<

Maybe if the transport protocol supports it or implies, e.g. HTTP or the
"From" in SMTP, but if you can't read the transport header then ....

I agree though that if you can't read a message sufficiently to understand
where to send the reply, then the sender behavior should be to infer that
the message has not been delivered. This makes sense, because, once you've
sorted out teething problems, most errors will be caused by data corruption
during transport (do you agree?)

David

-----Original Message-----
From: Duane Nickull [mailto:duane@xmlglobal.com]
Sent: Tuesday, March 07, 2000 4:17 PM
To: ebXML-Transport@lists. oasis-open. org
Subject: Comment on the TRP overview and Requirements documents


THis is awesome.  I see you have all been busy.  I have a couple of requests
and comments.

1.  Can you please place a large version number on the front page so we
don;t confuse this with future issues of the document.  (ie Version 0.2)

2.  Message 4.2 1(b) states "failure to deliver a document should be
notified to the party that sent the document"

I agree, however,  in the case where the document could not be parsed
(specifically the return address in the header) is this possible?

Maybe the methodoloy should be that a message that is not acknowledged
should be considered not delivered.  The use models for business may be
stricter than that with email, http etc.

Just "foo" for thought.

Duane