[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [ebxml-bp] Re: ebBP 2/7/2005: Updated Combined Summary of BSI-MSI
Duane, I'm not so sure you do follow what Sacha is driving at! Even within a single server - if I have a BPM engine, and ebMS, a Registry and some CPAs and CAM templates how do I make this all run smoothly together? If I am talking from my ebMS to a partners different ebMS, will their backend BPM engine manage state in the same way that mine is - given the signals and transactions we are using? Can he use rule templates that I have? If I replace the ebMS with another ebMS - what configuration is needed? We answered a lot of these Q's in the XML2004 interop for the six tools we configured - and created integration components based on particularly the features in CPA. But we did not link in a BPM engine - and that is another whole level of detail. Right now with BPSS V2 I think we can do some interesting and useful things - but its going to take BPSS V3 before a much richer environment can be specified here. 2005 is beginning with a lot of cool stuff happening! Cheers, DW ----- Original Message ----- From: "Duane Nickull" <dnickull@adobe.com> To: "Sacha Schlegel" <sschlegel@cyclonecommerce.com> Cc: "ebXML BP" <ebxml-bp@lists.oasis-open.org> Sent: Tuesday, February 08, 2005 11:44 AM Subject: Re: [ebxml-bp] Re: ebBP 2/7/2005: Updated Combined Summary of BSI-MSI > Sacha: > > ebXML is a component based architecture, we tried not to enforce > specific programming styles when building the specs. > > I do not see a high value in a concrete interface between messaging and > BPM. Since this is behind a corporate firewall, it is unlikely someone > would incur a SOAP + XML hit to have those components talk to each other. > > What is far likelier is that it will be in some specific language such > as Java and use RMI. Nothing in either inteface should be dependent > directly upon the data model in the BPSS schema or eb MS XML. > > I think I now understand what it is you are trying to do. An > implementation guide may be a good way to do this. Probably a good idea > for all instances of MSI and BSI to be dropped from the spec since they > have caused far too much confusion over the years. > > Duane > > Sacha Schlegel wrote: > > >ebBP team, Duane, > > > >comments inline ... > > > >Am Montag, den 07.02.2005, 13:26 -0800 schrieb Duane Nickull: > > > > > >>Monica J. Martin wrote: > >> > >> > >> > >>>Section 4: Language Overview > >>>================ > >>>FROM: > >>>The BSI is completely separate from the Message Service Interface > >>>(MSI). In particular an MSI MAY be used without a BSI. A CPA, which > >>>contains a reference to a ebBP definition serves as the basis for the > >>>configuration of the BSI to enforce the protocol and semantics of the > >>>ebBP definition, as depicted in Figure 1. > >>> > >>> > >>I would caution you that there are companies who have prior use of the > >>term "MSI" in the context of a programmers interface to ebXML MS > >>software. Accordingly, it may be subject to trademark and/or > >>copyright. I cannot verify this any further at this time. > >> > >>Discussing concrete interfaces also violates one of the core principles > >>of ebXML in the architecture which was not to impose any specific > >>implementation style *where unnecessary* upon those building software. > >>Prescribing interface based design is one constraint that is not in > >>alignment with either the core requirements or architecture of ebXML. > >> > >> > > > >I cannot argue against the "principles of ebXML" as I have not been > >there at that time the pricnciples were put into stone. > > > >But I can ask why? What was the rationale to not have components based > >ebXML building software? > > > >Why are is a Message Service Interface mentioned. This interface is that > >abstract that not even the interface is described. Guess it is not even > >an interface if it is not defined. > > > >With everything being so abstract everyone tends to do whatever he/she > >thinks is right. And because it is so abstract, virtual, logical, > >everyone is having a valid solution. > > > >Please excuse me but I see value for ebXML end users to be able to pick > >their building ebXML components. > > > >I feel silly because it is currently just one interface the one between > >the MSH and an ebXML business process component. The one between an > >ebXML business process component can follow thereafter. > > > > > > > >>Accordingly, it may be a better idea to discuss this as the association > >>with the messaging component of the architecture, not the interface. > >>Align the functionality abstract of the implementation. Interface > >>descriptions are probably better handled in a JSR type group than the > >>core standard. > >> > >> > > > >OK so there is a place for interface descriptions. The thing to me is if > >they are not part of the core standard they are simply of less > >importance, in particular to the Independet Software Vendors (ISV's). > > > >JSR as in Java Specification Request is not what I am looking for as > >that is language specific. > > > >But I recognise you saying that there is room for interface > >descriptions. > > > >Regards > > > >Sacha > > > > > > > >>Duane > >> > >> > >> > > > > > > > > -- > *********** > Senior Standards Strategist - Adobe Systems, Inc. - http://www.adobe.com > Vice Chair - UN/CEFACT Bureau Plenary - http://www.unece.org/cefact/ > Adobe Enterprise Developer Resources - http://www.adobe.com/enterprise/developer/main.html > *********** > > >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]