[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: Hardin (ebBP) 3/7/2005: Conversations in the context of ebBP-CPA-ebMS
The UDEF link was protected. Here is the public link: http://www.opengroup.org/projects/udef/ Registration is free. ~~~~~~~~~ john c hardin CIO - crossconnections.ws 313.930.5323 cell mailto:john@crossconnections.ws "The new electronic interdependence recreates the world in the image of a global village." Marshall McLuhan, "Gutenberg Galaxy", 1962 john c hardin wrote: > Thanks for the feedback, Monica - I finally joined the ebBP group, and > will be on the call today. > > What I was referring to was a wider issue that the overall architecture > of an SOA-like, cross-industry framework needs to address. That being, > the data element semantics in the actual business documents, which is > what I believe you are referring to by illustrating the "references to > semantic information that provides (loosely) a context to the business > information". > > There should a pattern in the architecture for patterns that allows for > 'semantic resolution' between two formats. Mechanisms providing semantic > analysis, mapping code generation and finally, transformation between > two or more formats that were mapped, needs to be much more automated. > Contivo does an excellent job of this, with a drawback: the enterprise > vocabularies are described in - you guessed it - english. This works > well for enterprise semantic metadata repositories. However, a numeric > format (CCTS) or alphanumeric format (UDEF) needs to be used in the > global framework. > > CPA should reference the semantic metadata entry (maybe Schema, could be > other however) in the registry that contains the semantic identification > information for each element, and the BP execution should link to the > transformation mechanism. Without human intervention. In english. ; > > > By the way, the UDEF has been picked up by the Open Group as the > potential host for the data element trees: > > http://www.opengroup.org/projects/udef/protected/ > > Let's remove the human mapping activity, eventually. Comments? > > > > > English please...let's look at this simply. The key is to provide > > a mapping that begins with semantic understanding. For example, ebBP > > allows references to semantic information that provides (loosely) a > > context to the business information provided. The semantic (and > > syntactic) understanding between the specifications is that they can > > effectively use (at their level of intelligence) basic constructs and > > patterns between them. For example, ebBP provides a process > > specification, roles and relationships and activities associated with > > both. CPA and ebMS takes that 'context' and uses them at the level it > > operates. This is semantic understanding. Of course there is more work > > to go (and we all acknowledge it). Your turn. Thanks. > > > > > > Monica J. Martin wrote: > >> john c hardin wrote: >> >>> Yes, absolutely. Actually today and all week, I am making notes for >>> submission to the group re: the next steps to take with the Technical >>> Architecture document. >>> >>> The link between the Reg-MS-CAP-BP execution layers is very crucial >>> for the next generation of completely distributed, loosely-coupled >>> apps. Public / Private processes are targeted for my personal goals. >>> I want to enable the entire network, with interlocking, but >>> independent processes. >>> >>> The two missing items, from my understanding, are the ubiquitous >>> semantic identifiers in all formats, and the event causality / event >>> monitoring handles capable of deciphering the undoubtedly complex >>> layers of events, firing events, firing OTHER events, which in turn >>> are used as decision points that fire yet more events, ad nauseum... >>> >>> Monica, what are your thoughts on this last part? >>> >> >> >> >> >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]