RE: web services uber alles? [was] Re: [ebxml-dev] RE: ebXML deliveries

["Welsh, David" <David.Welsh@nordstrom.com>]

> James Bryce Clark wrote:
> 
> > I think of a 'service level agreement' as an agreed set of 
> boilerplate 
> > economic deal terms beyond the semantic scope of a CPA.
  
Absolutely true in the business world.



> When we bury 
> > economic terms in the CPA (or MSG) layer, they are unexposed to the 
> > semantic tools at the BP collaborative layer, so they can't 
> be strung 
> > together with conditionality to create robust series of 
> transactions.
> >>>
> 
> 
> Duane N. wrote:
> 
> The Business Proces Schema is bound to two or more CPP's in the CPA 
> creation process (or should be at least but that process is not 
> defined).  Since one can reference the business information 
> via the BPS, 
> the economic terms could be inherently available to one who reads the 
> CPA and can follow the links.

There's more business service level details typically involved with a business collaboration that what's going to be captured in any CPP or agreed to CPA. The CP? is a technical layer convention, where as you mention later it's business driving details are declared later. 
Web Services here is perhaps really just let one CPP specify an easy to call business service 'application capability'; and what's behind the Web Service technically is a totally other matter.



> 
> I may concede that SLA may be a stratch of the term when 
> referring to CPA.
> 
>  
> > In the larger scheme, I think of the CPA as a binding agreement 
> > constrained to cover roughly three things:  the parties' 
> manifestation 
> > of identity, their modem handshake (so to speak) of 
> communication loci 
> > and protocols, and their mutual designation of one or more business 
> > processes to be shared.  Those business processes are 
> defined not in the 
> > CPA, but in a BP schema to which the CPA points.  
> >>>>

Actually the BP schema is only a very very simple manifestation of the true business collaboration as being defined by the bounds of the UMM; and as we discovered this past week @ eBTWG the BPSS v1.0 is kinda 'in need of immediate first-aid repair'. Still, even with a repaired BPSS  I'd caution any
software maker who is *only* look at the BPSS as a runtime instance, that they're not looking at capturing the entire BP model and where ebXML/eBTWG is heading off to.

-Dave