RE: [mdt-sbvr.dev] Re: The kernel of SBVR without programmingconsiderati

["Dave Carlson" <dcarlson@xxxxxxxxxxxxxxx>]

Mark,

I had also noticed that "vocabulary namespace" (in MRV) is defined in terms
of "vocabulary" (in VDBV) and had planned to submit that to the sbvr-rtf
issues.  We should create a combined list of RTF issues for submission.

The SBVR definition of "thing" states in a Note that: "Every other concept
implicitly specializes the concept 'thing'."  The SBVR 1.0 metamodel
represents this such that every metaclass is derived from "thing", either
directly or indirectly via generalization. A top-level container must be
allowed to include any "thing".

In addition to your list below, where neither "conceptual schema" or "fact
model" may include representations and expressions, other concepts are
excluded because they are direct specializations of "thing":

* namespace
* reference scheme
* language
* set (if used to represent the extension of a concept)
* And other concepts that are not defined within MRV (such as "community").

-- Dave

> 
> Stan and Sjir and I have been discussing how to model MRV in 
> terms of itself.  Here's my summary of the limitations of 
> "fact model" as a "top level container" for MRV:
> 
> * no provision for a URI
> * no ability to recursively embed a fact model within a fact model
> * does not include representations or expressions
> 
> And there are a couple of related issues:
> 
> * Arguably, MRV is deficient since it does not include the 
> concept "vocabulary", which (a) is needed to describe MRV 
> itself; (b) is referenced by the definition of "vocabulary namespace"
> * SBVR does not address the relationship between "conceptual 
> schema" or "fact model" and "body of shared meanings".
> 
> If we think this list is complete, then I think we should 
> submit a formal "issue" to the SBVR RTF with the goal of 
> coming to a consensus on how to proceed.