[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Subject: RE: Comments on 0.21
Forwarding to the Transport list David -----Original Message----- From: Eckenfels. Bernd [mailto:B.Eckenfels@seeburger.de] Sent: Monday, October 09, 2000 12:58 PM To: knaujok@pacbell.net; Burdett, David Subject: FW: Comments on 0.21 Hello Klaus or David, attached is my Comment to the Messaging Service Specification. I was unable to send to the List Address cause lists.oasis-open.org does not exist (as proposed on the web page) and lists.ebxml.org does not allow for non-subscribers to mail to the transport list (only subscribed to the architecture list). I tried to send my comments to Rik, but somehow i did not get back a response. So can u make sure the text is reviewed by the transport guys? Greetings Bernd -----Original Message----- From: Eckenfels. Bernd Sent: Monday, October 02, 2000 6:25 PM To: 'ebxml-transport@lists.ebxml.org' Subject: Comments on 0.21 Hello, i have some comments on 0.21 to make the document a bit more clean and make the specification less troublesome for ad-hoc communication. And even add some more flexibility if we want to use mailbox/VAN routing. 412-414: i would allow the PartyID Element to occur multiple times, so you can give all the known contexts. This will enable us to se a DUNS Number, a ILN Number, a TAX Code a Company Name, a e-mail Address, a van mailbox id and so on. This is especially good for routing. 451-453: if there is no earlier message we should require the element to be not present, not to be empty. Or if we want to have it empty but present, then we should tell this explicitely: if there was no former message the ELEMENT has to be present and be empty </ RefToMessageID> or <RefToMessageID></RefToMEssageID>. Personally I think we should require the element to be skiped if it is empty. We should also define what kind of message we are referencing to: a order change can reference to the order send or it can reference to the order-dis-acknowledgement of the partner. PErhaps we should allow the RefToMessageID to be repeatable and add an additional Attribute: context= answer: the Referenced Message ID is a message ID generated by you and we answer to it prev: the Referenced Message ID is a message id from us and we send the next message to is root: the Reference Message ID is the first message of a new business process and we reference it for better threading or something like this. 460: we should state who will interpret the ReliableMesssageInfo field and what he should do with it. In the enumeration we should also add the "AtLeastOnce" or something like this. Cause this is the normal way transmission protocols without duplicate checking will work (SMTP for example). Filtering duplicates should be responsebility of the endpoints. 551-591: i would allow the following elements to be optional: (the reason for this is it is better to allow them to be removed than everybody is making something "up" to fill in them. TPAInfo, ReliableMessageInfo, ServiceInterface , Action, ConversationID, make RefToMessageID (0-n)=* In the DTD you are using fixed attributes to describe the data types of the elements e-dtype... if you do so, you should mention where this is defined, othrwise skip them. Mit freundlichen Grüßen SEEBURGER AG Bernd Eckenfels -- SEEBURGER AG, Edisonstrasse 1, D-75015 Bretten, Germany Fax:+49(0)7252 96-2222 Fon:+49(0)7252 96-1256 http://www.seeburger.de/ We integrate B2B Solutions
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Powered by eList eXpress LLC