Re: [ebxml-bp] 12/7/2003: [Fwd: "Late Binding" of TimeToPerform]

["David RR Webber" <david@drrw.info>]

Lars,

Thanks for clarifying the use case.

I believe that the draft context header mechanism
I circulated, linked to the CPA - with the rules
specified in that header - allows us to do
what you are asking.

Particularly the business partner can see clearly
in that context header the exact range of responses
that are involved and why contextually.

There are some subtle nuances as I found out once
I studied this and looked at the generalized case of
being able to set any particular characteristics of the
BPSS / ebMS standard header - based upon rules
that relate to the particular service or product being
transacted.

That is why I made three namespaces - context header,
BPSS, and Document - so we could construct XPath
expressions appropriately.

Thanks, DW.

----- Original Message ----- 
From: <Lars.Abrell@teliasonera.com>
To: <ebxml-bp@lists.oasis-open.org>
Sent: Friday, January 09, 2004 12:15 PM
Subject: RE: [ebxml-bp] 12/7/2003: [Fwd: "Late Binding" of TimeToPerform]


Monica Martin wrote:

>As for the needs you cite, Anders made an important distinction if you are
>redirecting the business process or changing the product delivery date
>that is specified in a business document.  The next question is if that
>delivery date changes does it redirect/augment the business process, and
>does the business agreement allow for that.  Perhaps we should address
>this more fully in Monday's meeting.

Just to clarify the use case.
The "change" in the product delivery data is no actual *change* of the
delivery date for the same instance of a product.
It is more like different delivery dates for different products or services
ordered in different conversations using the same Binary Collaboration.

For example in a conversation #1 we order a product A from supplier X with a
delivery date 6 weeks from now. It's OK for us if the order response telling
us that the supplier can execute this order is sent to us after, let say,
one week. I.e. the value in timeToPerform for this simple BC should be P7D.

In another conversation #2 (instantiated from the same BC) we order a
service B that needs to be carried out by the same supplier X within 4
hours. In this conversation we need to receive an order response telling us
if the supplier is able to carry out the order or not within, let say, 10
minutes. Otherwise we might need to engage another supplier. I.e this time
the value in timeToPerform for this same simple BC should be PT10M.

What I am looking for is the best way of informing our trading partner
(supplier X) and his BSI of the value that should be used in the
timeToPerform attribute for each conversation. One way might be to use the
SBDH. Another way could have been using the ebMS message header if the ebMS
spec had provided us with for example a "timeout" attribute in the
ConversationID element.

It would be nice if one were able to specify a time range for the
timeToPerform attribute in the CPPA spec that the trading partners could
agree upon should be explicitly specified in runtime. I also think that if a
trading partner is disagreeing with the value given in runtime to the
timeToPerform attribute and not able not give a substantive response before
the expiration of the time specified it would be appropriate for him to send
an exception signal instead.

* Lars.Abrell@TeliaSonera.com * +46 (0) 705 619080
* Kilsgatan 4, Box 299, SE-401 24 Gothenburg, Sweden



-----Original Message-----
From: Monica J. Martin [mailto:Monica.Martin@Sun.COM]
Sent: Thursday, January 08, 2004 11:22 PM
To: Abrell, Lars E. /TSS - Networks and Production /031-770 90 80,
0705-61 90 80
Cc: ebxml-bp@lists.oasis-open.org
Subject: Re: [ebxml-bp] 12/7/2003: [Fwd: "Late Binding" of
TimeToPerform]


See comments between ****....****.

>Abrell2: It is of course possible to create multiple separate Binary
Collaborations that differs only in the timeToPerform value and where the
process flow it self is exactly the same. First, IMHO I don't think the
timeToPerform value is always an equally "important" business rule to be
agreed to and pre-known and determined in the same way as the process flow
it self is. I also believe that this approach could end of with to many
Binary Collaborations to handle when you actually only need one.
>
>Second, during the Christmas holidays I have read the "UN/CEFACT - Standard
Business Document Header Technical Specification Revision 2.1"
(http://webster.disa.org/cefact-groups/atg/downloads/index.cfm/). It seems
to me as if this specification specifies the things that I sometimes think
needs to be communicated between trading partners at runtime. I believe this
Standard Business Document Header to some degree overlaps the ebMS spec and
especially the ConversationId, Service, Action and CPAId elements, but at
the same time provide me with most (if not all) of the additional
information elements I have been missing.
>
>The Standard Business Document Header has a "block" called "BusinessScope"
that in turn contains a "Service Information" block and a
"CorralationInformation" block. This "CorralationInformation" block contains
an element called "ExpectedResponseDateTime" that could be communicated at
runtime between the trading partners and seems to give me what I have been
looking for. See below.
>
><BusinessScope>
>  <Scope>
>    <Type>BusinessProcess</Type>
>       ...
>       <CorrelationInformation>
>         ...
>
<ExpectedResponseDateTime>2003-05-10T00:31:52Z</ExpectedResponseDateTime>
>         ...
>       </CorrelationInformation>
>  </Scope>
></BusinessScope>
>
>The idea and theory around BusinessScope (see appendix 3) look also very
interesting.
>
>But I'm somewhat confused about when it's best to communicate the other
BusinessScope kind of information in a Standard Business Document Header and
when to use the elements provided by ebMS.
>
>
****mm2: The SBDH is primarily used to enable enterprise applications
and legacy systems to be able to access metadata to compile the business
document.  That is why the BusinessScope is important.  There has been
quite a bit of discussion about how this works or overlaps with ebMS,
and if the SBDH is actually sent with the business document in a payload
over the wire to enable parser and translator processing.  As for the
needs you cite, Anders made an important distinction if you are
redirecting the business process or changing the product delivery date
that is specified in a business document.  The next question is if that
delivery date changes does it redirect/augment the business process, and
does the business agreement allow for that.  Perhaps we should address
this more fully in Monday's meeting.

Dale, could you officiate? Thanks.****

>Kulvantunyou: I am not sure if I understand the whole complexity of the
issue or not.
>I preferred the BPSS being declarative at design time of business rule b/c
>1) business rules are captured declaratively and 2) it supports the advance
>dynamic binding b/w partners. So my thought for the use case described by
>Lars is to create multiple Binary Collaborations with different
>timeToPerform's for both BC and BTA for different classes of
>products/services rather than overriding at the runtime.
>
>I'm not sure whether the BPSS (I mean specifically BC) is supposed to be
>context specific or context indenpendent according to the Architecture. My
>thought is that it is context specific.
>
>Abrell: I have compiled a short summary proposal for Work Item 55 about the
late
>binding of the timeToPerform attribute. This is my very first proposal and
>I'm not sure if this is sufficient, but I hope so. There are still several
>implications and other things that needs to be worked out, and it would be
>very interesting also to listen to other opinions.
>
>mm1: Lars, after reviewing the slides, the note and thinking about the
learning
>session and teleconference discussions, there are several items that we
>should perhaps address, and talk about in the context of BPSS:
>
>    * Late binding on the BPSS
>          o Specifying the timeToPerform to the greatest time possible in
>the context of this type of collaboration [1]
>                + Allowing late binding to accommodate logical business
>document requirements
>                + Specifying when the late binding may occ
>    * Potential to override BPSS attributes in the CPP/A between the
parties
>[2] [3]
>    * Breaking the response into multiple responses: where more timing
>applies to specific requirements (such as delivery)
>or either from the perspective of the requestor or the responder [4]
>
>I encourage the team to think about these items in the context of
>roles/partners, timeToPerform, and the dynamic aspects of late binding
>we discussed last week. [5]
>
>[1] May likely not be as tight you would like given the interactions
defined
>by internal systems to support. Which brings up an interesting point on
>being able to keep these systems loosely coupled. For example, we have one
>case where the order may go by alternate means and not logged in the B2B
>system, and this really complicates this company's processing. We have to
>account for that at the process level.
>[1] Determine if this could meet your needs: See Section 9 of CPP/A
>document.
>[3] May complement any late binding functionality if adopted.
>[4] This may be driven either by business or service level agreements, or
on
>a case by case basis. For the latter, I would think we will have to discuss
>the per-instance impact and its appropriateness.
>[5] Related Work Items 23, 25, 28, 46 and 55.
>
>-------- Original Message --------
>Subject: "Late Binding" of TimeToPerform
>Date: Wed, 03 Dec 2003 03:21:23 +0100
>From: Lars.Abrell@teliasonera.com
>To: Monica.Martin@Sun.COM
>
>
>
>Monica,
>During our use of BPSS we have in different situations noticed a need for
>different values in the TimeToPerform attribute for the same Binary
>Collaboration. This specific use case is based on the BC:NegotiateOrder in
>the NeBI specification (www.nebi.biz). Please see the attached slides. This
>very simple use case for ordering different "Field Service Products" in the
>telecom industry is based on using only the BTA:OrderRequestByBuyer and the
>BTA:OrderAcknowledgmentBySupplier in the BC:NegotiateOrder. This simple
>version of the BC:NegotiateOrder is used for ordering different products
>using the same generic BusinessDocument (BD:Order). In the product specific
>part of this BD:Order (i.e, in the order rows) the different "Field Service
>Products" are specified. Depending on the different "Field Service
Products"
>(or combinations of  "Field Service Products") in a BD:Order or different
>"time to delivery" of the specified product, there is a need also to have
>different values in the TimeToPerform attribute of the BC:NegotiateOrder.
>
>Feel free to use this in any way you think is appropriate
>
> <<TimeToPerform.ppt>>
>Regards //Lars
>Lars Abrell, TeliaSonera
>
>
>
>
>
>